Self-Concept, Self-Monitoring and the Relationship of Human Social Interaction

Self-monitoring research has developed into a significant literature that spans across disciplines. However, little research has examined the influence of selfmonitoring on the development of the self-concept, which sets the stage for the current research endeavor. In the first study, it was hypothesized that directional questions "Are you extraverted?" would influence the self-concept of self-monitors to a greater extent, as they would be more likely to generate supporting thoughts which would mediate this effect. Findings generally supported the end effect of the hypothesis, but not the mediating process with high self-monitors reported a greater change in self-ratings after receiving a directional question. While low self-monitors were influenced more by thoughts that they retrieved in response to directional questions. The second study tested the assumptions of a proposed model of self-concept development that implicated self-monitoring as playing a central role. Selfmonitoring correlated significantly and positively with similar magnitude with both communication and psychological measures. These results point to self-monitoring being both a communication and psychology measure. Participants' self-ratings changed on most domains of the Big Five traits from time 1 to time 2 indicating that appraisal of both communication acts and psychological traits influenced their selfconcept as the findings of these two studies support


iv Preface
The focus of this research project was to examine how individuals construct their self-concept through interaction style, and how their pre-existing interaction style affects how people integrate social cues, and encode and decode messages. Currently no research exists regarding the role or relationship of ongoing interaction to the dynamic nature of social cognition, especially how communication and one' s self monitoring function. It was hypothesized that high self-monitors will be more likely to exhibit a change in self-concept between the first and second session on their self-rated trait that was assessed with the directional question; and that high self-monitors will be more likely to exhibit a change in self-concept based on the quantity of thoughts generated.
There is a great justification for and significance of the problem being addressed. Communication research has recently started to integrate concepts of cognition as explanations of communication behavior. This includes emphasis on structures, input and the process of the creation of communication. These are important aspects of social cognition, yet lack the "social" facet aspect of social cognition. The aspect of individuals' motivation and goals are neglected and focus is placed on the generic mapping of the cognitive mechanisms that affect communication. Self-monitoring is an important construct in both disciplines as it serves as a theoretical link between psychology and communication in regards to both interaction style and self-concept and may help bridge the disciplines together.
Self-monitoring is a construct that has been implicated as an important moderator in classical areas of psychology such as attitude/behavior congruence, v differential bases of attraction and friendship selection and cross-situational consistency. It has also become an important construct in the field of communication studies. As both disciplines converge to the similar theoretical basis of social cognition, both fields would benefit from research that is interdisciplinary in nature and uses constructs that are germane to both perspectives, such as self-monitoring.
However, very little research in either discipline has examined the impact of selfmonitoring on the malleability of self-concept using a social cognitive methodology.
A series of studies were conducted to further stimulate interdisciplinary research in these two fields. Study 1 examined the influence of directional questions on self-concept and the influence of self-monitoring. Study 2 tentatively tested a proposed model of the development of self-concept and its subsequent assumptions in which self-monitoring plays a central role in integrating both communication and psychological perspectives. The two studies support the model. Table   Study 1 Tables  Introduction   Scholars and pedagogues have attended to social cognition, or how one perceives the self, as central or integral to explaining and understanding human interaction. Sociologists Mead ( 1934) and Cooley ( 1902) developed symbolic interactionism and it's associated processes of co construction of juxtaposed social reality, meanings, and the self. A body ofresearch exists (Cappella, 1981) that firmly establishes the influence of input on social cognition from significant others and social comparison. However, currently no research exists regarding the role or relationship of ongoing interaction to the dynamic nature of social cognition, especially how communication and one' s self monitoring function. The focus of this research project is to examine how individuals construct their self-concept through interaction style, and how their pre-existing interaction style affects how people integrate social cues, and encode and decode messages.

LIST OF TABLES
Concomitant to the joint construction of selves through interaction with others is the social cognition process of creating, noticing, verifying, and recreating one's self-concept. Affirmation or disconfirmation of one's self concept by another is also coupled with the simultaneous reciprocity of acceptance or rejection of the other's self concept as interaction proceeds. Negotiating the self-concept with another requires an individual to engage at some level in what Snyder (1974) identified as self-monitoring. Self-monitoring has been defined as the amount of expressive control than an individual has across various situations (Snyder, 1974). To further explicate the relational importance of self-monitoring and communication, it can be assumed that self-monitoring, a fundamental interaction style (i.e. one is either low or more accessible. The process of integration is the same in both theories. They both have a structure that an individual changes based on interaction, while self-schema or the working self-concept is more encompassing in that other factors influence change.
Since social cognition is a relatively new approach in communication research, the cognitive communication foundation is still being established.
Communication researchers typically focus on the interpretative and choice-making powers while explaining interaction and explain their processes in a cognitively grounded manner (McPhee, 1995). This includes emphasis on structures, input, and the process of the creation of communication. These are important aspects of social cognition, yet lack the "social" aspect of social cognition. People are not solely information-processing computers. This is an issue that psychology addressed as social cognition research grew. The aspect of individuals' motivation and goals are neglected and focus is placed on the generic mapping of the cognitive mechanisms that affect communication.
Since research on the cognitive aspects of communication is still developing, interdisciplinary collaboration may contribute information that will provide a more complete picture of the cognitive aspects of communication. Motivation and goals are issues that McPhee ( 1995) has discussed as being less emphasized in social cognitive communication research; they can be introduced by studying selfmonitoring, a psychological construct that focuses on an individual's motivation and goals with respect to her or his interaction with the other. High self-monitors are motivated to be socially appropriate; thus they communicate with others in this manner. This desire and how successfully appropriate they are in communicating then in turn will influence their self-concept.

4
Self-monitoring also ties in the reciprocal influence of self-concept and communication. Goss (1989) stresses since "we build our self-concepts with interactions with others, self-monitoring influences our self-concept, as self-monitors examine their self-concept 'in action" '(p. 88). Therefore, self-monitoring is an important construct in both disciplines as it serves as a theoretical link between psychology and communication in regards to both interaction style and self-concept.

Self-Concept
The self-concept, or how the individual sees the self, has been an area of research interest for quite some time. Certain perspectives have found that the selfconcept is stable; individuals actively try to reinforce their established self-concept, while other perspectives have found that the self-concept is malleable. A more contemporary perspective is that the self-concept is both malleable and stable, depending upon what structure of the self-concept is being discussed or theorized.
This thesis discusses both sides of the argument surrounding the nature of selfconcept. Following the discussion of the various perspectives and their corresponding research, is a more in-depth focus on the malleable (impressionable, and changeable) nature of the self-concept. This thesis also addresses the processes that have been theorized as responsible for the change in self-concept. This section will conclude with research that frames the current empirical question that will be posed. One thrust derived from this literature review is an assessment of the role that directional questions, and consequent retrieved memories associated with them, play in changing one's working self-concept.

5
The second section examines the construct of self-monitoring; a construct that may provide qualifications of when directional questions elicit biased memory searches with the resulting change in self-concept. Scant research has examined the self-concept of low and high self-monitors. A cogent approach to examining the selfmonitoring literature would be to review the literature that is the most significant to the current research question. The researcher claims that the variability of a high selfmonitor's self-concept is more loosely defined than a low self-monitor's self-concept, which will then result in a high self-monitor's self-concept being more amenable to change when presented with a directional question. Thus various relevant areas of research were unpacked and are presented: stability of the self-concept; malleability of the self-concept; motivational bases of the self-concept etc .. The first area that will be discussed involves differences in cross-situational variability of the low and high self-monitor. The author assumes that a more situationally variable person, such as a high self-monitor, will have a less consistent self-concept than a low self-monitor. The second area of inquiry involves research on differences between the attitude/behavior relationship of the low and high selfmonitor. This research provides insight into the dynamic self-concept of the high self-monitor. The final area of research reviewed discusses the different types of social information to which a low self-monitor attends compared to a high selfmonitor. Since high self-monitors look to the situation for behavioral guidance, they should have a self-concept that is more inconsistent, as they have information about how to act in various situations, yet have no consistent self-guide. Stability of  The previously discussed notion of the self-concept can be defined as the constellation of categories that an individual uses to describe self. These categories span many levels of detail. These categories include country of origin, the region of the United States where an individual resides, level of self-monitoring, down to the traits that an individual uses to describe their personality, among others. These categories also consist of the types of roles that an individual uses when in the presence of others, how an individual interacts with others, and how typically other people have treated them in the past. Self-concept also has an evaluative component in that it is the repository of what type of regard an individual has for one's self. The self-concept is molded through social comparisons with other people and allows us to evaluate our standing among others on any issue that we choose to acknowledge. Now that self-concept has been conceptualized, one of the fundamental questions addressed in self-concept research is: Will I be tomorrow who I am today and who I was yesterday? Some researchers on self-concept claim that an individual's self-concept is relatively stable across the whole lifespan, or for a majority of the life span. For example, Mortimer and Lorence (1981) found that five dimensions of self-concept "including well being, self-doubts, sociability, competence, and unconventionality'' were stable from late adolescence to early adulthood. These results are significant, in that stability of the self-concept is held during a life period usually marked by instability and change.
Certain inherent characteristics of individuals affect their self-concepts. An individual may be white, female, with an average IQ. Self-perception of these qualities comprise the basis of the individual's self-concept. Individuals enact these qualities in differing degrees throughout their social interactions with others. Besides stable demographic features, individuals may have certain personality characteristics that are a predominant part of their self-concept. They may be extraverted or introverted, and this, in tum, affects their behaviors and attitudes toward others.
These individuals may place themselves in certain situations that are characteristic of the self-concept. For example, Snyder and Gangestad (1982) reported that low selfmonitors, who have a self-concept that is internally constructed from their moods and beliefs, were more willing to enter situations that reflected their personal dispositions.
High self-monitors, with a more situationally-based self-concept that is more situationally-based, were most willing to enter situations in which extraverted roles were clearly defined, such as being a meeting facilitator.
Self-verification. Some researchers state that individuals actually engage in self-verification; that is, they actively try to reinforce their established self-concept.
In interaction with others, individuals who want to verify their self-concept may try to draw the attention of others to certain traits or attitudes that they possess. Swann and Read ( 1980) discuss a variety of these self-verification techniques that individuals use, which include social, behavioral and cognitive means.
One technique used for self-verification is to seek out social feedback that readily confirms an individual's self-concept. Swann and Read (1980) found that individuals who perceived themselves as being '<unlikable" would spend more time reading negative statements that their partner produced about them, while "likable" individuals would spend more time reading positive statements that their partners produced.

8
Individuals also engage in self-verification by behaviorally eliciting responses that confirm their self-concept. Swann and Read (1980) found that individuals that perceived themselves as being likable elicited more favorable reactions from partners.
The amount ofliking was statistically mediated by individuals' increased praises to their partner, and this effect was stronger when they believed that the partner had a negative impression of them.
Individuals also confirm their self-concept through their cognitive processes. Swann and Read ( 1980) found that individuals that perceived themselves as being likable recalled more positive than negative evaluative statements that their partner made about them. Similarly, Markus (1977) found that schematic individuals (or individuals that have a self schema that is strongly characterized by a particular trait) would be less willing to accept information that was incongruent with their schema.
Thus, on the basis of the research reviewed to this point, it appears that the self does seem to remain stable, and that individuals actively try to maintain this conception of self through self-verification.
The Self-Concept as Malleable Markus and Kunda ( 1986), and numerous other psychology researchers, propose that the self-concept may be both stable and malleable. Inside the more stable, general self-concept, Markus and Kunda (1986) believe that there is a working self-concept, or a malleable self-concept that individuals have for the current moment in time. Markus and Kunda (1986), drawing from Bern's (1972) self-perception theory, posit that the working self-concept is affected by our current motivations and goals and the norms of social situations. These working self-concepts guide behavior in how individuals should act in the current situation. Once individuals leave a social situation, they may employ a different working self-concept, although in contrast, they still maintain a stable, general self-concept that was intact across these social situations. They should still recognize the self as a member of their respective gender category, yet their perception of their level of a more abstract concept, such as a personality trait, may be higher or lower.
Self-concept can be less rigid than previously thought, based on the context or the phenomenological experience associated with social stimuli (Markus, 1977).
Aschematic individuals (or individuals who did not attribute a trait in question as being characteristic of their self-concept) were more inconsistent in their responses when asked to indicate whether the trait was indicative of self or not. This result is significant on numerous fronts. First, it supports the idea that the self is not solely stable, underscoring the capacity of the self-concept to be both stable and malleable.
Another important consideration from Markus's (1977) study provides guidance to what aspects of the self-concept are malleable and which processes germane to this plasticity. Perhaps individuals have a self-concept that is rigidly stable on some characteristics of their personality, while other aspects are less rigid, and may possibly change with situations or goals.
Motivational Bases. Much of the literature on the malleability of self-concept emphasizes the self-esteem needs of individuals. When an individual feels too unique lO or similar to everyone else, they may be motivated to change their self-concept to establish a social comparative equilibrium. Markus and Kunda (1986) capitalized on this prospect, and investigated the malleability of self-concept through use of a traditional research paradigm in which individuals experienced one of these two socially undesirable outcomes. Markus and Kunda (1986) posited that these differing states are aversive states to the self-concept, and will have a significant impact on the working selfconcept. To reduce threat to the working self-concept, the researchers theorized that individuals would attempt to conscript memories that stress the opposite state. In other words, if people are made to feel overly different from others, they will retrieve memories or behaviors that implicate their similarity with others. If people feel overly similar with others, they retrieve memories in which their behaviors were different from other individuals. The cognitive act of retrieving either type of memory has processing implications, which reduces response latencies when deciding whether a word (reflective of the retrieved memory) is self-descriptive or not. Markus and Kunda ( 1986) found that participants who were told that they were overly unique or similar from others did not differ on their choices of either unique or similar items when describing themselves, providing evidence that the general self-concept is stable. However, participants made to feel unique had shorter response latency times when deciding if conforming words were descriptive of them, indicating that their working self-concept may have activated memories or concepts in which the participant was similar to others. This pattern extended to participants who were made to feel similar to others. In a word association task, participants made to feel unique later associated uniqueness with negative terms, and similarity with positive terms. Thus, a valence is attached to certain memories or structures associated with one's self-concept. Markus and Kunda (1986) indicate that the individual's working self-concept is malleable, and can change ifthere is a motivational basis behind the change. Individuals theoretically conduct a memory search through their past experiences, trying to find those particular events that fulfill the current need. The individual may define their current self-concept based on their most recent communication style or active memory structure. This theory captures the malleability and fleeting characteristics that represent the self-concept, as evidenced by Markus ( 1977) and Markus and Kunda ( 1986). Kunda and Sanitioso (1988) believed that individuals recruit one-sided memories to support the hypothesis or theory and they neglect to look for memories that do not support the initial hypothesis. The researchers set up a paradigm in which they told students that a certain personality characteristic (either introversion or extraversion) was related to being successful in graduate school, something that is extremely desirable and self-esteem enhancing for most students. Kunda and Sanitioso ( 1988) then had students theorize why either introversion or extraversion would have a significant relationship with academic success. Kunda and Sanitioso (1988) found that students who developed theories of why one trait was important in being successful in graduate school, perceived themselves as having the trait to a greater extent. Sanitioso, Kunda, and Fong (1990) examined the process that Kunda and Sanitioso ( 1988) had proposed, and found evidence that supported their proposal of a biased memory search when participants answered questions that assessed an individual's self-concept. Congruent self-concept memories that supported a supposedly desirable trait were more accessible to individuals. That is, when participants were told that a certain trait was more desirable, the participants conducted a biased memory search. This biased memory search activates these characteristics and they became more accessible than incongruent memories in the working self-concept. This was reflected through shorter response periods and onesided answers on memory-listing tasks reflecting the desirable trait.

12
Self-esteem, or how positive or negatively one evaluates self, which is also an evaluative component of self-concept, is also amenable to change when there is a motivational basis. Jones, Rhodewalt, Berglas, and Skelton (1981) examined the effects of strategic self-enhancement and self-deprecation on self-esteem. In this study, participants were involved in a mock interview where they had previously observed a confederate and were provided with incentives to use either a selfdeprecating or self-enhancing strategy. The strategy that they had used had a corresponding effect on their self-esteem. Participants had a higher self-esteem when they engaged in a self-enhancing strategy and a low self-esteem when they engaged in a self-deprecating strategy. This effect was replicated even when participants were explicitly told by the experimenters to act a certain way. This study shows that communication behaviors that are motivationally based can change self-esteem.
Also, the process of pretending to act a certain way can affect self-esteem. Direction of Self-Concept Change. When individuals are presented with different social stimuli, either self-or other-initiated, individuals either accept or reject integrating the information into their self-concept. When they do accept the information, it can either increase or decrease their perceived level of a trait in their self-concept. Rhodewalt and Agutsdottir ( 1986) discuss this concept when they describe the phenomenal self, a concept similar to Markus's and Kunda's (1986) working self-concept. Phenomenal self is a subset or collection of self-relevant information that is currently activated. Rhodewalt and Agutsdottir ( 1986) believe that individuals have certain latitudes of rejection or acceptance. That is, if a person engages in certain behaviors, acceptable to the self, the behavior falls within the latitude of acceptance. If a person engages in behaviors that fall outside the range of acceptable behaviors to the self, the behavior is in the latitude of rejection.
If a behavior falls in the latitude of acceptance, individuals will engage in selfperception processes whereby their self-concept will change, reflecting their attention to their current behavior. If the behavior falls in the latitude of rejection, the behavior will elicit cognitive dissonance. Cognitive dissonance is an aroused state in which the individual's behaviors do not match their attitudes, and they try to rectify this discrepancy through various processes.
For example, Rhodewalt and Agutsdottir (1986) believed that participants would reduce cognitive dissonance by reporting that they were less like the selfdiscrepant behavior elicited. The process of using either self-perception processes or experiencing cognitive dissonance is mediated by the individuals' self-knowledge of their behaviors. If someone has a developed self-knowledge structure and they engage in a self-discrepant behavior, he or she will experience cognitive dissonance, while a less developed structure will result in self-perception processes. Rhodewalt and Agutsdottir ( 1986) found that depressives (depressed people) who had to act in a self-enhancing manner, and nondepressives (people who are not depressed) who had to act in a self-deprecating manner, felt cognitive dissonance, which had an opposing effect on their self-esteem. Depressives who acted in a selfdeprecating manner and nondepressives, acting in a self-enhancing manner, emphasized a change in self-esteem that reflected their behavior. These areas of research point to the role of motivational bases in changing self-esteem. Several studies assessed what effect, if any, certain types of questions would have.
Can Self-Concept Change Without a Motivational Basis?
One issue raised bt the preceding iswhether self-concept can change. Fazio, Effrein, and F al ender ( 1981) examined the effect that certain types of questions would have on an individual's self-concept. Fazio et al. (1981) proposed that asking certain types of self-directed questions without a motivational basis may prime or activate a different self-concept. In the study, individuals silently read questions to themselves and then answered them. The questions were either indicative of introversion or extraversion. Based on the type of questions asked, individuals actually answered in a manner similar to the type of questions. Individuals who asked themselves introverted questions responded in a less extraverted manner than individuals who asked themselves more extraverted questions.
These questions influence participants' behavior Fazio et al. (1981) Individuals acted in either a more introverted or extraverted manner towards a confederate. These self-directed questions also affected the participant's selfperception of their own introversion or extraversion. Participants who asked themselves introverted questions thought that they were more introverted than participants who asked themselves extraverted questions. Sherman, Skov, Hervitz, and Stock (1981) reported that participants who were asked to explain either failure or success on an upcoming anagram task resulted in the participants having either higher or lower expectations of behavior, which then translated into actual performance.

15
As previously discussed, the rationale for a third study was that a directional question would provoke individuals to think of memories that were one-sided in favor of the question. Kunda, Fong, Sanitioso, and Reber (1993) conducted a study assessing whether asking an individual a directional question about one's introversion or extraversion influences their self-concept. Individuals will engage in a "positivetest strategy" (Klayman & Ha, 1987), in which information will be found (in the form of memories) which supports the question and there will be a general lack of memories retrieved that do not support the question being asked. Kunda et al. (1993) constructed a research study in which participants answered a question about themselves, such as "Are you happy/unhappy with your social life?' Participants then listed any thoughts that answered the question. After listing their thoughts, participants then completed a self-rating scale that asked them to rate themselves on how happy or unhappy they were with their social life. Kunda et al. ( 1993) found that the directional question facilitated the retrieval of memories that answered the question in a one-sided manner, providing support for a biased memory search.
Participants also rated themselves as being happy or unhappy with their social life (above or below neutral) based on this memory search.
In subsequent studies, Kunda et al. (1993) explored some of the parameters or the boundary conditions in which a change in self-concept was less likely to occur. Kunda et al. (1993) found that a directional question no longer had an effect on selfconcept if participants' perceived themselves as being more variable on the particular trait. Kunda et al. ( 1993) also found that participants' listed more two-sided thoughts when they described themselves as being more variable on the trait in question.
Regarding Kunda et al.'s (1993) findings that an individuals' perceived variability on a trait has a significant relationship on two-sided thoughts listed and also subsequent self-ratings; and arose the usage of a personal characteristic to account for these differences. The personal characteristic that can most likely account for these differences is an individual's differences in self-monitoring (Snyder, 1974). According to Snyder (1974), self-monitoring can be described as the degree to which an individual can "monitor their self-presentation or communication style, expressive behavior, and verbal affective display'' (p. 527). In other words, high self-monitors are able to put on the right "social mask" when it is called upon, while low self-monitors tend to follow their internal dispositions and present themselves in a way that is a reflective of how they feel or think during a given moment rather than what is perceived to be "socially correct" or "expected" in that situation. The Kunda et. al (1993) work allows the bridge between self-concept and self-monitoring.

Self-Monitoring as a Construct
Self-monitoring has been defined as the amount of expressive control than an individual has across various situations (Snyder, 1974). Low self-monitors are typically thought of as individuals who let their moods or general dispositions on various issues guide their behaviors or outward expressions when they are interacting with those around them. If a low self-monitor is interacting within a group, they typically reveal their true self through what they say, or how they behave with outward expressions of emotion. The low self-monitors' behavior is consistent across situations. High self-monitors are typically thought of as individuals who have no consistent base of behaviors, as they are typically change their own behaviors to fit the social situation. They mask or hide their "true self' constantly changing their expressions based upon the external, or rather the social stimuli of which they find themselves a part. The high self-monitors' behavior is inconsistent across situations, because it adapts to changing situations.

Self-Monitoring Scale
One way to better discern who a self-monitor is involves examining the factors that compose the original Self-Monitoring Scale (Snyder, 1974). Snyder (1974 stated that self-monitoring consists of five aspects: (a) concern with being socially appropriate in one's presentation; (b) attention to social comparison information as cues to appropriate expression; (c) the ability to control and modify one's self-presentation; (d) being able to use this ability in particular situations; and (e) the consistency of behavior across situations (Snyder, 1974).
The Self-Monitoring Scale (Snyder, 1974) was validated through various procedures and appears to be both a psychologically meaningful and unified construct. However, other researchers have argued that Snyder's (1974) Self-Monitoring Scale consists of more than one latent variable. Briggs, Cheek, and Buss (1980) found that the Self-Monitoring Scale (Snyder, 1974) consisted of three distinct factors: acting, extraversion, and other-directedness, which all lacked significant intercorrelations with regards to the independent factors. Various other factor-analytic studies support the claim that the Self-Monitoring Scale consists of multiple independent factors (Gabrenya & Arkin, 1980;Lennox & Wolfe, 1984;Nowack & Kammer, 1987;Snyder & Gangestad, 1986). Some studies have found as many as four factors (Gabrenya & Arkin, 1980) while others have found two general factors (Briggs & Cheek, 1988;John, Cheek, & Klohnen, 1996) in the Self-Monitoring Scale.
Based on their data, Briggs et al. (1980) warrant caution in conceptualizing self-monitoring as a unitary construct. Briggs et al. ( 1980) found that otherdirectedness, the willingness to change one's behavior to suit others, was positively correlated with neuroticism and shyness, and negatively correlated with self-esteem.
While extraversion, was negatively correlated with neuroticism and shyness, and positively correlated with a different measure of extraversion.
Another concern is the construct of self-monitoring is not truly operationalized in the Self-Monitoring Scale. Snyder and Gangstad (2000) state that the self-monitoring scale is a valid measure of self-monitoring because of its intrinsic validity. Hence, hundreds of studies have shown a significant difference when people were assessed on their self-monitoring.
Mi ell and Le Voi (1985) found that using the three subscales of the Self-Monitoring Scale (Snyder, 1974), (derived from Briggs et al., 1980) in analysis were more useful predictors than employing the scale as a whole. Mi ell and Le Voi ( 1985) replicated the same type of patterns as Briggs et al. (1980). Some subscales of the Self-Monitoring Scale went in the opposite direction of the other subscales when they were used in analysis. Therefore, the predictive utility of the scale was lowered when subscales that are moving in opposite directions were combined. Dividing selfmonitoring into three constructs when conducting analyses has also been utilized in other studies (Carver, 1989;Hull, Lehn, & Tedlie, 1991;Richmond, Craig, & Ruzicka, 1991;Snyder and Gangestad, 1986;Sullivan & Hamish, 1990).
Results like Miell and Le Voi's (1985), and Briggs et al.'s (1980) classic factor analytic study started to create doubt among researchers about the meaning of self-monitoring as developed by Snyder (1974) and measured by the Self-Monitoring Scale (Snyder, 1974). Not taking these criticisms lightly, Snyder and Gangestad ( 1986) wrote a rebuttal directly to Briggs et al. (1980) and Miell and Le Voi ( 1985) in which they claim that self-monitoring is a psychologically meaningful construct and does measure one main underlying factor. Snyder and Gangestad (1986) provide support for self-monitoring from various sources. One critical source of support is strong empirical findings when selfmonitoring has been used as a way to differentiate individuals. This point serves as an indirect statement of the reliability of a construct. According to Snyder and Gangestad (1986), simply relying on factor analysis to determine a scale's internal structure may ignore a more complex phenomenon, which has been found to be significant in numerous studies. Gangestad and Snyder ( 1985) also cite direct support for the construct by reanalyzing studies of self-monitoring and conducting analyses of the internal structure of the Self-Monitoring Scale (Snyder, 1974). Through the use of a taxometric analysis, Gangestad and Snyder ( 1985) found that their measure of selfmonitoring does tap an underlying causal variable.

20
In a later analysis, Briggs and Cheek (1988) found that there were two factors that composed the Self-Monitoring Scale, not one, and that the first factor correlated with extraversion, exhibitionism, and social potency. Briggs and Cheek ( 1988) also discuss some of the problems with the general theoretical approach taken in the development of the self-monitoring concept from inception to the current time (e.g. Gangestad Snyder, 1974;Snyder & Gangestad, 1988). One problem that Briggs and Cheek ( 1988) identified that it is assumed that a person is either a high or low self-monitor, instead of being placed on a continuum of self-monitoring. Miller and Thayer ( 1989) also support the claim that self-monitoring may not be a discrete class variable. Miller and Thayer ( 1989) conducted analyses with both discrete and continuous measurements of the Self-Monitoring Scale and found better empirical evidence in treating self-monitoring as a continuous variable, evidenced through higher factor loadings.
However, Gangestad and Snyder (1991) believed that Miller and Thayer's ( 1989) analysis consisted of false psychometric premises and a general misunderstanding of Gangestad and Snyder's (1985) claims. According to Gangestad and Snyder ( 1991 ), Miller and Thayer ( 1989) implicitly proposed that if a continuously measured scale correlates with a dichotomous variable, then a dichotomized version of the measure would measure the variables better. Hence, when Miller and Thayer ( 1989) found that the continuous scaled measures better fit the variable, they believed that self-monitoring was a continuous variable. Gangestad and Snyder (1989) stated that a continuously scored indicator would actually correlate higher with a dichotomous variable, than with a dichotomized version of the variable.
The second concern that Briggs and Cheek ( 1988) presentis similar to past arguments about the specious nature of the self-monitoring construct. Basically, one factor of the scale, termed other-directedness (Briggs, Cheek, & Buss, 1980;Miell & Le Voi 1985) is indicative of a person who "may have a neurotic maladjustment" (p. Briggs and Cheek (1988) lay with the general use of the term self-presentation, which is found within the self-monitoring individual. Briggs and Cheek (1988) argue that high self-monitors engage in self-presentation, while low self-monitors never engage in self-presentation. This contradicts Snyder's (1974) conception of the self-monitor in which low self-monitors do engage in selfpresentation. That is, a low self-monitor's style of self-presentation is consistent across situations, and is indicative of their internal states. However, Briggs and Cheek (1988) believe that self-presentation means something different. Selfpresentation is conceptualized as the individual communicating in ways that reflect themselves to others in a positive light, no matter in which situation they find themselves. Briggs and Cheek (1988) believe that this is something that is not characteristic of a low self-monitor. Thus, Erving Goffinan ( 1959) states that we are actors following scripts which are determined by the situation. Extending this to the current research paper, it is believed that high self-monitors are better adept at understanding and following the prescribed scripts.

674). A final concern of
Several parallel theoretical lines appear to lay similar support for the communicator reaction to the environment stimulus. First, social psychologist Kurt Lewin ( 1936) devised the formula B=f(P ,E) to represent the notion that behavior is a function of the person and the environment, interactionists have been interested in the relationship between personality characteristics and the situation as well as the resulting behaviors. Second, Erving Goffman ( 1967), a symbolic interactionist, developed a dramaturgical approach, building upon G.H. Mead's work. This theory explained communication behavior on the basis of a particular situation or "frame" thereby giving situation the defining influence in all interaction. Using Goffman's explanation: "the emphasis is placed upon how selves and identities emerge out of situational constraints that shape the 'performance' of the role players" (Layder, 1981). Goffman believes that our interaction is guided by following particular rules, "scripts" of a situation.
People act particular ways in different situations because they are following the rules and expectations of their interpretation and perception of the situation. An individual may perceive a situation to have certain rules because a situation resembles another, the rules of which they are familiar (Lord, 1982). Lord ( 1982) states, "the notion that behavior is a function of the situation as perceived or interpreted by the individual ... has been a basic premise of social psychology" (p. 1084). He, like Cody and McLaughlin ( 1985) developed a method of classifying situations based on knowledge and perceptions. This paper will not fully delve into individuals' perceptions, because although that is an important area of study.

Cross-Situational Variability
Similarly, psychologist's theoretical concepts give validity to the focus of interaction responses shaping self-monitoring that has seldom been discussed by either critics (Briggs and Cheek, 1988;Briggs, Cheek, & Buss, 1980) or by supporters (Gangestad & Snyder, 1985;Snyder, 1974;Snyder &Gangestad, 1988) is the issue of context effects. Nesler, Tedeschi and Storr (1995) describe context effects as a phenomenon in which responses to one personality inventory influences responses to other scales or behaviors measured within the same testing session. Not surprisingly, Nesler et al. (1995) found that when participants read instructions that indicated that the Self-Monitoring Scale measured one's ability to act, with females' scores on the scale being higher while males' scores on the scale were lower than the standard testing condition. In a second experiment, Nesler et al. (1995) had participants respond to the Self-Monitoring Scale after responding to questions about sexual deception. Nesler et al. (1988) found that in this case, scores on the Self-Monitoring Scale were lower than the standard testing condition. Lippa (1976) found that when participants were asked to role-play either introverted or extraverted teachers over subsequent trials, high self-monitors were better able to act out the various performances better than low self-monitors, based on naive judge's ratings. High self-monitors' expressive behaviors changed more across trials to fit the requirements of the situation compared to low self-monitors (Lippa, 1976). However, there was one facet of personality that high self-monitors had trouble changing across trials. Lippa (1976) found that high self-monitors were consistently rated by judges as being more extraverted across situations; their extraversion "leaked out" even when they were required to role play an anxious teacher.
24 Snyder and Monson (1975) examined the differences that self-monitoring has on an individual's sensitivity to situational cues to act in a socially appropriate manner, and the behavioral consistency of self-monitors. Snyder and Monson (1975) found that high self-monitors would reliably conform to group discussions that differed over time. In a second study, Snyder and Monson (1975) found that high self-monitors reported more situational variability than low self-monitors on the traits of generosity, honesty, and hostility. Lippa and Donaldson (1990) conducted a study assessing the cross-situational consistency of self-monitors' behaviors across interpersonal relationships. In this study, participants filled out diaries in which they indicated who they were interacting with, their current inhabited setting, and the behaviors that they engaged in.
Participants filled out a computer assessment indicating people that they interacted with, the role or relationship with that person, behaviors that are typically engaged in, in these situations, and the settings typically inhabited with these people. Lippa and Donaldson (1990) found that the various measures were highly intercorrelated, and that consistency measures were negatively correlated with self-monitoring. That is, high self-monitors show less cross-situational consistency in relationships High self-monitors' tendency to be more cross-situationally variable in expressive behavior implies that these individuals that are more adaptable to differing situations. Caldwell and O'Reilly (1982) examined the role of self-monitoring and performance in boundary spanning jobs. Boundary spanning jobs are characterized as sending and receiving information both inside and outside the organization's boundaries. Individuals who have boundary spanning jobs must present organizational information to other constituencies, outside of their organization, and thus, create variable situation and interaction demands. This type of job brings an individual into contact with many different types of people and social settings, as varied organizations conduct business with his/her home organization. A boundaryspanning job would probably be best performed by high self-monitors, according to Caldwell and O'Reilly (1982), as they can alter their self-presentations in the various situations that they find themselves in. Caldwell and O'Reilly (1982) found that high self-monitors outperformed low self-monitors in boundary spanning jobs.

Attitude/Behavior Congruence
In an early study of attitude and subsequent behaviors of self-monitors, Snyder and Tanke (1975) found that low self-monitors have a higher correspondence between behaviors and attitudes then high self-monitors. Zanna, Olson, and Fazio (1980) extended the research on the attitude behavior relationship among selfmonitors and found that only low self-monitors who held previously unchanging views about a topic held consistent behaviors and attitudes. Further evidence by Paulhus (1982) showed that low self-monitors' are more apt to experience cognitive dissonance (a type of cognitive conflict) when they are asked to engage in behaviors that are against their attitudes, or that they would view with a negative valence. To rationalize this attitude behavior discrepancy, low self-monitors change their attitudes to reflect the recent behavior, and actually perceived the behavior to be more enjoyable. However, since high self-monitors have lower attitude/behavior congruence, they did not experience cognitive dissonance, or all of its subsequent effects. 26 Snyder and Kendzierski ( 1982) examined determinants of the correspondence between attitudes and behavior concerning affirmative action. Snyder and Kendzierski (1982) initially found that correspondence between attitudes and behavior was low for both low and high self-monitors. However, when increasing the availability of attitudes, by allowing individuals to think about their attitudes, created a high attitude/behavior correspondence for low self-monitors, but not for high selfmonitors. In another condition of the study, Snyder and Kendzierski (1982) found that increasing relevance, or the importance of attitudes, created an attitude/behavior correspondence between low and high self-monitors. The similarity of responses between low and high self-monitors can probably be justified by interpreting their actions through different motivational biases. Relevance, or importance of decision possibly affected low self-monitors on an internal level. Low self-monitors may have taken personal responsibility for the outcome of their decision. High self-monitors may have interpreted the relevance information on a self-presentation level, and felt compelled to be socially desirable. Mellema and Bassili ( 1995) examined the relationship between values and attitudes and the possible moderating role of self-monitoring. Mellema and Bassili (1995) predicted that there would be a stronger relationship between the values and attitudes of low self-monitors compared to high self-monitors. Mellema and Bassili (1995) drew these predictions from the past research findings of Snyder and DeBono {1987), who found that attitudes for the low self-monitor were more indicative of their actual values, while a high self-monitor's outward displays of attitudes were more based on socially appropriateness or what attitude best fits the "attitude of the social situation." The "attitude of the social situation" is manifested through a groups' decision, or through a description in a vignette. Mellema and Basilli ( 1995) found that values for low self-monitors were more predictive of attitudes towards issues.

Attention To Differing Sources of Social Information
Besides "attitudes'', another dimension of the situation is how the communicator perceives her/his partner, especially with respect to social knowledge about her/him. Elliot ( 1979) examined how differing levels of self-monitoring affect individuals' preparation for a future interaction, and the impact that presenting an accurate or fabricated impression has on this preparation. Elliot ( 1979) predicted that since high self-monitors are eager to acquire information about a target, they will spend more of their time trying to learn information about the target, while a low selfmonitor will let his or her internal states guide the future interaction. Elliot ( 1979) found that when high self-monitors planned a fabricated interaction, they "bought" more information about the partner, while low self-monitors relied more on internal information in both types of interactions.
When we typically choose our partners for social activities, we may choose them based on usually one of two sets of considerations (Snyder et al., 1983). Snyder, Gangestad, and Simpson (1983) conducted a study examining the different considerations that high and low self-monitors take into account when they are choosing with whom they want to interact. Snyder et al. ( 1983) were mainly concerned with the systematic differences in the nature of friends who comprise the social networks of low and high self-monitors.

28
One set of considerations involves choosing friends that are experts, or are knowledgeable about a task. These types of friends are typically chosen for a particular task or activity. Snyder et al. (1983) believe that high self-monitors choose activity partners that are specialists, making it easier for them to display their competence in various domains. As a result of "expert picking," high self-monitors should tend to have a social world, where they pick different friends for different situations. Conversely, low self-monitors choose activity partners that they like, or share similar views with, resulting in low self-monitors having a consistent set of friends that they interact with in various situations. Snyder et al. (1983) found that their predictions were true. High self-monitors preferred partitioned social worlds in which they used specialists for activities and low self-monitors chose friends that were similar to them and did all activities with them. These results show that high self-monitors prefer multiple friends for varying situations, while low self-monitors prefer friends who are similar to them and with whom they interact across various situations.
Does this type of result also apply to dating relationships? These types of patterns also prevail in romantic relationships oflow and high self-monitors (Snyder of a link between length of relationship and level of intimacy as opposed to low selfmonitors. Again, Snyder, Berscheid, and Glick ( 1985) dichotomize as Snyder, Gangestad, and Simpson ( 1983) had done with the two sets of considerations that are taken into account when choosing an activity partner. However,  dichotomized the sets of consideration that individuals take when initiating dating relationships and serving as factors related to attraction.  propose that high self-monitors want attractive partners as it serves as a boost to their public appearance, while low self-monitors are more invested in internal attributes of the other person. Individuals either attend to physical appearance or to personality attributes when making relationship choices.
These patterns of attraction, physical appearance or personality, which also served as indicators of initiation of relationships among low and high self-monitors . High self-monitors held physical appearance as the most relevant determiner of and most relevant on whether they would chose someone for a date. Low self-monitors chose personality characteristics as their main determiner when choosing a date. Supporting Snyder et al. ( 1983 ), Jamieson, Lydon, and Zanna (1987) reported similar results. Initial attraction towards an individual was based on attitude similarity for low self-monitors, and activity preference similarity had a stronger influence on high self-monitors' attraction.
Besides attending to either physical or internal characteristics of potential dating partners, self-monitoring also serves as a moderator for the self-attribution of emotions. Graziano and Bryant (1998) demonstrated that low self-monitors used the self as a guide to evaluate and regulate emotions, although high self-monitors would look to situational cues to understand their emotions. High self-monitors believed that certain stimulus objects were more attractive when they heard a measure of their heartbeat increase, when in actuality the supposed measure of their heartbeat was manipulated by the experimenters. This effect did not replicate among low selfmonitors. When they heard a fake measure of their heartbeat increase when viewing certain objects, there was no difference in perceived attraction. Graziano and Bryant ( 1998) also found that high self-monitors found comedy skits to be funnier when a laugh track was included, while low self-monitors did not report a difference in perceived jocularity of the comedy skits that included a laugh track.
Other research examined the role that self-monitoring plays in individuals' usage of consensus information to predict future behavior in a new situation (Krosnick & Sedikides, 1990). Consensus information is "data regarding how most people would act in the situation" (p. 718). Their predictions were based on earlier findings by Kulik and Taylor ( 1981) who found that high self-monitors used consensus information more than low self-monitors.
This previous research on consensus information typically yielded two findings: l )People either use the consensus information (or the base rate information of how most people would act); or, 2) they may use their past behaviors in similar situations to predict how they would act in the new situation. Krosnick and Sedikides (1990) predicted that certain personality characteristics, such as self-monitoring, potentially would moderate individuals use of consensus information, or information that is more idiosyncratic in nature. Krosnick and Sedikides ( 1990) predicted that high self-monitors, being overly concerned with projecting a socially positive image, would use consensus information that is related to engaging in good deeds and avoiding socially undesirable actions, compared to low self-monitors. However, Krosnick and Sedikides (1990) were also examining the types of consensus information as previously indicated. Through a series of studies, Krosnick and Sedikides (1990) found that high self-monitors were more responsive to complimentary or socially desirable consensus information than low self-monitors; while low self-monitors were more responsive to threatening consensus information.
In other words, high self-monitors were more likely to engage in socially desirable behaviors when others engaged in the socially desirable behavior; while low selfmonitors would help engage in socially desirable behavior when no one else would.
Thus, self-monitoring is used for predicting future actions based on differing types of consensus information.
Self-monitoring has been used to further understand the different social information processing goals of low and high self-monitors. Chen, Shechter and Chaiken ( 1996) discussed two different types of processing goals. One type of motivation in processing information is to attain accuracy, or to "be right." The second goal in processing is to "be friendly" or to put across a friendly, socially desirable image to others. Chen et al. ( 1996) also discuss these goals and their implications on the depth of processing. They state that accuracy-motivated goals are "indicative of both heuristic and systematic processing in which the individual processes information in a relatively impartial, open-minded manner'' (p.263) The researchers mention that impression-motivated goals are indicative of the "selective use of heuristic processing, set at achieving the immediate social goal at hand" (p.263). Chen et al. (1996) predicted that high self-monitors were more likely to use impression-motivated processing, and low self-monitors would use more accuracybased processing. In their first study, Chen et al. ( 1996) discovered that high selfmonitors were more impression motivated, and tended to reflect attitudes of the same valence as their partner, when compared to low self-monitors. In their second study, Chen et al. ( 1996) found that accuracy-motivated individuals (low self-monitors) displayed their accuracy "bias" in both initial attitudes, reflecting heuristic processing and then this bias continued with later attitudes in which low self-monitors engaged in more systematic processing. Snyder and Cantor ( 1980) also examined different types of social knowledge that low and high self-monitors attend to in their social interactions. Based on the self-monitoring typology, low self-monitors tended to reflect their internal dispositions when interacting with others; while high self-monitors behaviors tended to fit in with the current social interaction. Since both have divergent styles of interaction, and sources of social knowledge in how to interact, it should also relate to their self-knowledge. Their self-knowledge reflects their social knowledge or behaviors and their social behaviors are indirect ways of measuring their selfknowledge. Taking this into consideration, Snyder and Cantor ( 1980) predicted that low self-monitors would have a better internally based, cognitive representation of their characteristic selves compared to high self-monitors. On the other hand, high self-monitors would have more knowledge or a "richer" knowledge base of externally based cognitive representations of prototypes, or the typical, ideal way of acting across situations. These predictions were supported when Snyder and Cantor ( 1980) found that low self-monitors provided more informative descriptions of their own characteristic selves. High self-monitors were better able to provide "richer" descriptions of prototypes or of individuals that had different traits.
Low and high self-monitors show an increased attention to sources of information that were converse to each other. As a result, the feedback from these sources of information can have differing effects on low and high self-monitors' selfesteem after self-presentation. An interesting study by Jones, Brenner, and Knight (1990) found that when low self-monitors were asked to play self-serving, unethical roles that are typically unrepresentative of self, they had an increase in self-esteem when audience feedback indicated that they failed to act in a convincing manner. Jones, Brenner, and Knight (1990) also found that high self-monitors reported a higher self-esteem if they received audience feedback indicating that they were successful in playing the same role.
Earlier research by Ickes, Layden, and Barnes, ( 1978) examined the greater emphasis low self-monitors' place on self as the informational base for their selfconcept. Ickes et al. (1978) created an experimental scenario in which objective selfawareness was either present or not present for low and high self-monitors and found that low self-monitors were affected more than high self-monitors in the objective self-awareness condition. Low self-monitors' responses shifted in the objective selfawareness condition in that they described their self-concept as being more unique.
However, high self-monitors showed no change across conditions. In general then, results show the greater emphasis low self-monitors place on stable self as the key to understanding how they see themselves.
These results indicate that it was reaffirming and self-esteem enhancing for low self-monitors to have difficulty acting like someone else and convincing an audience. High self-monitors' self-esteem was increased when they were able to convince a crowd of being someone different. This places the emphasis of self as being the main source of self-esteem for the low self-monitor, while success at "playing the role" was more important to the high self-monitor's self-esteem. An interesting modification of the self-monitoring literature was that playing the role of someone socially undesirable did not reduce the effect of self-esteem enhancement in the high self-monitor, rather successful acting for an audience was more self-esteem enhancing.
High self-monitors, who develop a variable self-concept through various social situations, most likely did not change their description of their self-concept as the experimental situation did not "necessitate" a change in their socially constructed self-concept. However, high self-monitors do not look to the self to assess their selfconcept. The self-awareness condition creates an emphasis on the self, however high self-monitors do not use self for describing self-concept.

Review of the Theoretical Integration of Self-Concept and Self-Monitoring
This review of the self-monitoring literature brings up significant points that bridge self-monitoring as a possible moderator of the relationship between directional questions and self-concept. Referring back to Kunda et al. (1993) the investigators assessed how individuals perceive their variability on either extraversion or introversion across situations. They determined that the perceived variability consequent to cueing was related to whether retrieval of thoughts in response to a directional question had effected a subsequent self-rating on that trait. This indicator seems rather similar to the aspect of variability as evidenced by the research on crosssituational consistency, attitude/behaviors and attention to different sources of information in the self-monitoring literature. To establish self-monitoring as bridging construct before communication and perspectives, the researcher designed and executed two studies. Therefore: H 1 : It is hypothesized that high self-monitors will be more likely to exhibit a change in self-concept between the first and second session on their self rated trait that was assessed with the directional question.
H 2 : High self-monitors will be more likely to exhibit a change in self-concept based on the quantity of thoughts generated.

Overview of Study One
The first study was designed to test to see the reaction of one's selfconcept after the presence of a strong directional question. The second study was designed to test how significantly related/ similar certain measurements between the disciplines of psychology and communication studies are with regards to the areas of self-monitoring and self-concept. Therefore, in the first study, the researcher adapted Kunda et al. 's ( 1993) methodology. Similar results were expected, i.e. high selfmonitors, or individuals who have a variable self-concept including either extraversion or introversion, will perceive themselves as possessing differing levels of that trait after recalling thoughts that support a directional question ("are you introverted?", "are you extraverted?".
The hypothesis for Study One may be explained by two differing accounts.

36
First, high self-monitors are more situationally responsive and they will perceive themselves differently based on the thoughts that they recalled. In this case, the thought process and the research paradigm (presentation of a directional question) itself will serve as the situation. Second, since low self-monitors perceive themselves as being consistent and behaving on internal dispositions, they will disregard the thought process as solely being a situation that has no true bearing on how they perceive themselves. Furthermore, thought process (recalling and writing down thoughts to the proposed directional question) and the research paradigm will have no effect on self-concept, as evidenced through self-ratings.

Method
Participants Participants consisted of 53 (21 males and 32 females) undergraduate students from University of Rhode Island enrolled and listed on an instructor's roster from either a COM 101 (Public speaking) or COM 206 (Research methods) introductory communication studies course. Of the 53 participants, 52 indicated their race/ethnicity, with 45 (86.5%) being European-American, 2 (3.8%) being Asian-Americans, 2 (3.8%) being Hispanic-Americans, and 3 (5 .8%) indicated "other" as their ethnicity. The age range of the participants was rather homogenous ( 18 to 25) with a mean of 19.7, and a standard deviation of 1.55.

Materials 37
Participants completed two counterbalanced forms during the first session of the experiment and four forms in random order (with certain constraints) during the second session. The first session took approximately fifteen minutes. Participants were given a packet consisting of a title page, a personality scale and a self-rating form. The revised Self-Monitoring Scale (Snyder & Gangestad, 1986) consists of 18 items that will be changed from a True-False format to a likert-type scale with a 1 to 5 response format, with 1 indicating "strongly disagree", to 5 indicating "strongly agree". A higher score on the scale indicated having a higher degree of selfmonitoring. The revised Self-Monitoring Scale (Snyder & Gangestad, 1986) was used as it has been reported to have a a of +.70, which is higher than the internal consistency of the original Self-Monitoring Scale (Snyder, 1974). The revised scale was also used as it has been reported to be a more distilled measure of selfmonitoring (Snyder & Gangestad, 1986).
Another form that participants filled out was a self-rating form based on the Big Five construct of personality. The Big Five construct was used, as it has become a prominent personality structure in both personality and social psychological research. The Big Five include extraversion, conscientiousness, emotional stability, culture, and agreeableness. The self-rating form consisted of 25 items (five indicators for each construct of the Big Five) on a 10-point scale with a negative trait anchored at the left and a positive trait anchored on the right. Accordingly, higher scores indicate more positive judgments of self. Some indicators were randomly reversed to prevent response set in participants.
The second session of the experiment also took about fifteen minutes.
Materials and procedures were modified from Kunda et al. ( 1993 ), in which participants were given packets containing various forms in a random order under certain constraints. One form that participants completed was a form with the instructions "Please list below some examples of your past behaviors, thoughts, or feelings that come to your mind as you try to answer this question" (p.67). Half of the participants were either presented with one of two questions, "Are you introverted?" or "Are you extraverted?". These two trait items will be used as they were previously replicated in Kunda et al's work (1993). In addition, past research on the impact of extraversion on communication has been very strong. One result has been that extroverts talk more then introverts (Goss, 1989). To apply this, we assume that there will be a subsequent effect in the quantity of words used in descriptions to answer the two proposed questions. On another form they completed a self-rating form consisting of the Big Five, the same as the self-rating form in the first session.
The third form that participants filled out was assessing how variable they see themselves on the particular trait on a scale from 1 to 10. On the fourth form, demographic information was measured by asking open-ended questions assessing gender, age, and race/ethnicity. These four forms were placed in random order with one constraint. The stimulus form (ie "Are you introverted/extraverted?") always came before the selfrating form. This constraint on random order was impinged for two reasons. First, the causal variable always preceded the dependent variable, allowing the researchers to make a stronger claim for causation. Second, because instructors and professors were unwilling (after initial assent) to allow data collection, the researcher did not have enough participants to allocate for another independent variable assessing changes in self-rating in the second session before receiving the directional question and after receiving the question. The self-rating form and the Revised Self-Monitoring Scale (Snyder and Gangestad, 1986) were counterbalanced to prevent presentation effects.

Procedure
Overview. The experiment was conducted in two sessions, each session taking approximately fifteen minutes. In the first session, the experimenters indicated the purpose of the experiment, the requirements of the study, and that participants will complete the informed consent document, a self-monitoring scale and a selfrating. Experimenters told the participants that the second session would consist of answering questions and then completing a self-rating, a measure of variability and demographic information. There was at least a one-week lag time between sessions so that participants were less likely to remember their self-ratings in the first session.
First Session. In the first session, an experimenter came to the class, introduced herself and told participants that they are conducting an experiment that will be assessing the interaction of personality and situational factors regarding interpersonal interactions. The experimenters indicated that the study would be conducted over two sessions. The experimenters mentioned that for the first session of the study, participants would fill out various scales. For the second session of the experiment, the experimenters indicated to the participants that they will be asked a few questions and then they will fill out various forms. Possible participants were notified that they can withdraw at any time and will not be penalized for sole completion of the first stage. The class was told that anyone who does not want to participate can leave with their professor and he/she will have an alternative exercise provided for them. Both failure to participate (or participation) did not affect the student's grade in any way.
After covering the guidelines of the study, the researchers then indicated that anyone who would like to participate in the study should raise her/his hand, and an informed consent document was distributed to those individuals. Those participants willing to be a part of the study were handed a packet to complete. The packet consisted of a title page, the Revised Self-Monitoring Scale (Snyder and Gangestad, 1986), and a self-rating form. Participants were then instructed to proceed through the packet until they were finished.
After the participants have completed the packet, the participants submitted their forms to the experimenter. The participants were then instructed to write a word and one number that they could easily recall as a password for both the first and second sessions on their data packets. This word and number were used as a means to track and match participants across sessions, and to keep them anonymous; not for any other means. This step was taken to insure both anonymity and confidentiality for the participants.
Second Session. Experimenters returned to the class after one week, or after 19 days had expired. Upon entrance, the experimenters reminded the participants that they were here to proceed with the second session of the experiment. The researchers then reiterated that the participants who previously agreed to participate in the second session should raise their hands and the experimenter gave them the packet.
The packet that participants filled out consisted of five pages. The packet consisted of a title page, a page with a question asking participants to indicate the thoughts or feelings that they think of when answering whether they are (a) introverted or (b) extraverted. The next page of the packet consisted of a 25-item self-rating scale followed by a page assessing the participant's perceived variability on the indicated trait. A fifth page assessed demographic information.

Descriptive Statistics
Self-monitoring.
Self-monitoring was relatively equal across gender and ethnicity. An independent-samples t-test revealed that there was no significant gender difference in relation to self-monitoring t 50 = .503, p = .617; with means and SDs of 55.81 and 7 .51 for males, and 54.71 and 7.88 for females. A one-way AN OVA revealed a lack of a significant ethnicity effect for self-monitoring F 3 , 47 = .169, p = .917. The means and SDs for ethnic groups were 55.09 and 8.15 for European-Americans, 55.5 and 3.54 for Asians-Americans, and 53.00 and 4.24 for Hispanics. However, these results should be viewed as tentative since the sample was 86.5 percent European-American. Extraversion.
Extraversion had the same pattern of results as self-monitoring regarding gender and ethnicity. Self-rated extraversion was relatively similar for males and females, t(49) = -.731, p = .468, two-tailed. Means and SDs were 5.5 and 1.73 for males and 5.97 and 2.50 for females. Across ethnicity, a one-way ANOV A revealed no significant effects, F(J, 46 ) = .755, p = .525. The means and SDs based on ethnic groups were 5.93 and 2.27 for European-Americans, 3.50 and .71 for Asians-Americans, and 6.00 and 1.41 for Hispanics. Highly unequal Ns for ethnic groups was also a concern regarding this analysis. Kunda et al's (1993) measure of variability was correlated with selfmonitoring, to assess whether this initial predictor and self-monitoring tap the same construct. There was no significant relationship between self-monitoring and selfreported variability of introversion/extraversion, r = -.068, n = 45, p = .659.

Diagnostics/Manipulation Checks
Since the forms were counterbalanced in the first session, an independent samples t-test was conducted to determine whether there was an order effect on responses to forms. Ratings of extraversion and self-monitoring scores were analyzed for differences depending on the order of presentation. Presenting the self-monitoring scale and then the self-rating scale resulted in means and SDs of 5.68 and 2.56 for self-rated extraversion (n =25); 54.00 and 8.22 for self-monitoring (n = 24) vs.
presenting the self-rating form and then the self-monitoring scale with means and sd's of 5.89 and 1.88 for extraversion (n = 26), and 56.15 and 7.18 for self-monitoring (n = 28). There was no significant difference between extraversion and self-monitoring based on order of presentation, t(49) = -.326, p = .746, and t(50) = -1.003, p = .321 both two-tailed, respectively.

43
A 2x2x2x2 mixed-model ANOV A (Study 1, Table 1 ), consisting of one within-subjects variable and three between-subjects variables, was used to analyze under what conditions; personal characteristics, contextual conditions, and the interaction of personal characteristics with contextual conditions affect change in selfconcept. The within-subjects variable, change in extraversion, consisted of two levels; pre and post directional question rating of extraversion. One between-subjects variable was level of self-monitoring, as either low or high based on scores on the self-monitoring scale. Participants who scored less than or equal to the theoretical median of 54 were coded as low self-monitors, while those who scored 55 and above were coded as high self-monitors.
Another independent variable was the type of question (either introverted/extraverted) that participants were asked. A third between-subjects variable was congruence/incongruence of question with self-rated extraversion. A discussion of coding for the third between-subjects variable is in order. First, the participants were coded as either introverted or extraverted based on self-ratings in the first session. A rating of 1 through 5 indicated self-rated introversion and 6 through 10 indicated self-rated extraversion. This information was then compared to the type of question that they were asked. Congruency of question occurred when participants' self-rating matched the type of question that they were asked.
There was no significant change in extraversion, F( 1,42) = 1.468, p = .232, from time l to time 2. Type of question did not have a significant main effect on change in extraversion, F(l ,42) = .050, p = .825. Congruency of directional question with an individual ' s self-rating did not have a significant main effect on change in extraversion, F( 1,42) = 1.028, p = .317. However, there was a significant main effect for self-monitoring on change in extraversion, F(l ,42) = 5.992, p = .019. Means for extraversion for low self-monitors decreased from session l to session 2, from 5.367 and 4.133; and means increased for high self-monitors from 5.619 to 6.036. There was also a two-way interaction between type of question and congruence on change in extraversion, F(l,42) = 12.154, p = .001. Participants given a congruent extraverted question exhibited a decrease in extraversion, from 7.50 to 6.33; participants given an incongruent extraverted question exhibited an increase in extraversion from 4.167 to 4.667. Participants that were given a congruent introverted question, rated themselves higher on extraversion, from 3.567 to 4.6; while participants who were given an introverted question that was not congruent, rated themselves more introverted, from 6.738 to 4.738.
There were no significant two-way interactions between type of question and self-monitoring or congruence of question and self-monitoring, F(l ,42) = .001, p = .98, and F(l ,42) = .050, p = .825. There was no significant three way interaction between type of question, congruence of question, and self-monitoring on change in extraversion, F( 1,42) = .587, p = .448.

Listed Thoughts
Two raters coded thoughts in response to the directional question asked during the second session. The raters were two undergraduate students, who were trained by the researcher to code the listed thoughts into various fixed categories. Thoughts were either coded as supporting the question, refuting the question, irrelevant to the question, and total thoughts overall. Inter-rater reliability was sufficiently high, Cronbach's a's of .93, .97, .89, and .95 for agreeing, disagreeing, irrelevant, and total thoughts, respectively. The results confirmed that the raters were judging coded responses reliably with each other.
There were no significant gender differences on any of the types of thoughts Also, at-test revealed no significant difference between high and low self-monitors on types of thoughts, t(so) = -1.224, p = .227; 1(so) = .77, p = .441; t(sO) = 1.132, p = .263; t(so)= .011, p = .991 for agreeing, disagreeing, other, and total thoughts. However, correlations showed some significant relationships between self-monitoring and types of thoughts listed. There was a significant relationship between selfmonitoring and agreeing thoughts, r = .251, n =50, p = .040; and total thoughts, r = .251, n = 50, p = .039, both one-tailed, when self-monitoring was not median split into two groups, which is necessary for at-test.
Means and SDs for agreeing thoughts across racial groups were 1.42 and 1.53 for European-Americans, 0 and 0 for Asian-Americans, 2.83 and 2.00 for Hispanics.
Means and sd's for disagreeing thoughts were .67 and 1.35 for European-Americans, 1 and 1.41 for Asian-Americans, 0 and 0 for Hispanics. Means and SDs for other or irrelevant thoughts were .09 and .35 for European-Americans, 2.5 and 3.54 for Asian-Americans, 1.5 and 2.12 for Hispanics. Means for sd's for total thoughts were 2.18 and 1.87 for European-Americans, 3.5 and 2.12 for Asian-Americans, 3.5 and 4.95 for Hispanics. There was a marginally significant interaction between type of question and self-monitoring on disagreeing thoughts, F(l ,42) = 3.283, p = .07. Low self-monitors generated an average of 1.33 disagreeing thoughts compared to high self-monitors who generated .411 disagreeing thoughts when asked an extraverted question. Low self-monitors generated an average of .567 disagreeing thoughts when asked the introverted question, compared to high self-monitors who generated an average of 1.119 disagreeing thoughts. For the other types of thoughts, there was no interaction between type of question and self-monitoring, F's (1,42) of .002, .365, 2.393, p's > .05. There was a significant interaction between type of question and congruence, F(l ,42) = 6.783, p = .013, for disagreeing thoughts, and for total thoughts, F(l ,42) = 6.790, p = .013. There was a significant interaction between self-monitoring and congruence on total thoughts listed, F(l,42) = 5.728, p = .021. Low self-monitors generated an average of 1.867 total thoughts when presented with a question that matched their self-concept, while high self-monitors generated an average of 3.155 total thoughts when asked a question that matched their self-concept. Low self-monitors asked a question that did not match their initial self concept generated an average of 3.00 total thoughts, while high self-monitors who were asked a question that did not match their initial self-concept generated an average of 1.393 total thoughts. No significant twoway interaction between self-monitoring and congruence on agreeing thoughts, F( 1,42) = 2.831, p = .10. Low self-monitors that were presented with a question that matched their self-concept generated an average of 1.00 agreeing thoughts, while high self-monitors presented with a question that matched their self-concept generated an average of 2.036 agreeing thoughts. Low self-monitors presented with a question that did not match their self-concept generated an average of 1.33 agreeing thoughts, while high self-monitors presented with a question that did not match their self-concept generated an average of .661 agreeing thoughts. There was no significant interaction between self-monitoring and congruence on disagreeing or other thoughts, F(l,42) = 1.434, p = .238; F(l,42) = .130, p = .720. There was no three-way interaction between type of question, self-monitoring, and congruence on any of the four types of thoughts. F's (1,42) = 1.20, .181, .601, 2.504, p's > .05.
Mediation. (Table 3)  However, correlation showed that there was a significant relationship between self-monitoring and change in extraversion, (Study 1, Table 3): r = .252, n = 50, p = .038, one-tailed. This datum indicate that the degree of self-monitoring is related to the type of change in extraversion ratings in the second session. Regarding types of thoughts, the number of disagreeing thoughts and total thoughts were related to change in extraversion: r = .30, n = 50, p = .017; r = .369, n = 50, p = .004, respectively both one-tailed. There was no significant relationship between the number of agreeing thoughts and change in extraversion, and other thoughts and change in extraversion, r = .17 4, n = 50, p = .114; r = .065, n = 50, p = .328, respectively both one-tailed.
There was a significant relationship between the number of disagreeing thoughts, total thoughts and change in extraversion: r = .519, n = 27, p = .003; r = .358, n = 27, p = .034 when the participants were presented with the introverted question. When participants were presented with the extraverted question, there was a significant relationship between the number of agreeing thoughts, total thoughts and change in extraversion, r = .484, n = 24, p = .009; r = .414, n = 24, p = .022.
Regarding self-monitoring, there was a marginally significant relationship of r = .378, n = 19, p = .055 between the number of other thoughts and change in extraversion for low self-monitors when presented with either type of question.
There was a significant relationship between number of disagreeing thoughts and change in extraversion, r = .522, n = 31, p = .002, when high self-monitors were presented with either type of question. There was also a significant relationship among high self-monitors between the number of total thoughts and change in extraversion, r = .417, n = 31, p = .009.

50
Regarding congruence of question with initial self-concept change in extraversion and the number of disagreeing thoughts, which correlated significantly at r = .376, n= 33, p = .015, (one-tailed) when participants were presented with congruent questions. When presented with congruent questions, there was also a significant correlation between change in extraversion and number of total thoughts, r = .427, n = 33, .007, (one-tailed). There was no significant relationship between change in extraversion and the number of any type of thoughts when the question was incongruent with initial self-concept.

DISCUSSION
The present findings suggest that changes in extraversion occur only under certain conditions and when certain personality characteristics are present. Kunda et al's (1993) mediating variables (thoughts) are only used (explicitly through writing) in certain conditions or combinations of conditions. Sometimes these thoughts influence change in self-concept, and then sometimes it is only certain types of thoughts that have an influence on change in the self-concept.
Overall, there was no significant change in extraversion from session 1 to session 2 (5.78 to 5.68). However, this does not take into account the various moderators that may have increased or decreased extraversion, resulting in scores "washing out" variables that may influence extraversion by decreasing it and other variables acting to increase it, would negate each other out when this data is aggregated.

Malleability of Self Concept
Of the three main variables used, self-monitoring was the only construct that had a significant main effect on change in extraversion, with low self-monitors rating themselves more introverted, and high self-monitors rating themselves more extraverted in the second session. A tentative interpretation of this result is that indeed each group's (low and high self-monitors') self-concept are malleable, but the change that occurs is self-verifying. Low self-monitors had a lower mean self-rating of extraversion (5.36), displayed a lower extraversion rating (4.13); high selfmonitors had a higher mean self-rating of extraversion (5.619), which increased during the second session to 6.036.
Malleability of self-concept does occur in both groups, but it occurs in a direction that reinforces who they already are. This advocates that "change" is actually fortifying the group's initial self-concept, similar to the argument that Swann and Read (1980) advocate. This was a surprising result as it was intuitively believed that low self-monitors would be less likely to change their self-concept at all based on situational constraints, while high self-monitors would have more self-concept change potential, either increasing or decreasing their self-rated level of extraversion based on the situation at hand, as high self-monitors' self-concept is based on prototypical situations, rather than stable attributes that low self-monitors use to describe their characteristic self (Snyder & Cantor, 1980). When a participant was asked a directional question congruent or incongruent with her/his initial self-concept, there was an interaction. However, interaction occurred in an unexpected way. When an extraverted person was asked if they were extraverted, they rated themselves more introverted, and an introverted person rated self more extraverted. Introverted individuals asked if they were introverted rated themselves more extraverted. Extraverted people asked if they were introverted rated themselves more introverted. These results are unexpected.
The types of thoughts that were produced in this two-way interaction may provide some insight into these unusual results. Taking these results as a whole it seems that social desirability may play a part in explaining these findings. When introverted individuals are asked if they are introverted or extraverted, they used disagreeing thoughts to guide their judgment.
Since introversion is not as socially desirable as extraversion, introverts when asked if they are introverted might say something like "I am ... , but I do like to do this ..... ", using a more defensive style when answering the congruent question. Therefore, they retrieve thoughts that support the question because they are able to accept who they are and retrieve thoughts that do not support the question as they are trying to maintain that they are not total introverted "geeks" and promote their self-esteem though listing socially desirable behaviors or traits. This is further supported by the amount of total thoughts (3.25 for incongruent question, 3.033 for congruent question) introverted individuals used compared to extraverted individuals ( 1.988 congruent, 1.143 incongruent). In either case, they rate self as higher on extraversion, as dissonance may have been created in which their introversion was made salient through either their the question presented or the thoughts that they retrieved, resulting in them rating themselves higher on extraversion.
Extraverts used a less active approach towards congruent and incongruent questions. When asked if they were introverted, many extraverted participants simply wrote "no" and did not bother to think about any possibilities when answering the question. They seemed to be less actively processing the information. Further evidence to support this is that they rated themselves as more introverted, perhaps being biased by the presented question, and less by cognitive processes. However, extraverted participants presented with a congruent question may have exhibited a decrease in extraversion due to chance, as they displayed very little disagreeing thoughts.
Regarding thoughts listed, there was a two-way interaction between selfmonitoring and congruence of question. High self-monitors generated more total thoughts when they were presented with a congruent question compared to low selfmonitors. Low self-monitors generated more total thoughts than high self-monitors when presented with an incongruent question. This could mean that the type of situations in which low and high self-monitors find themselves elicit different cognitive strategies. Low self-monitors engage in more processing when they are in a situation that disputes their self-concept, while high self-monitors engage in more processing or active thinking when they are in situations that support their selfconcept. This makes sense, because when the low self-monitor is in a situation that is supportive of self there is no need to monitor the attributes of the self. Conversely, in disparate situations they need to maintain a sense of who they are. High selfmonitors withdraw from situations incongruent with their self-concept, and become more cognitively active when they are in situations that support their self-concept.

Influences on Extraversion Change
Several predictors and mediators are related to change in extraversion.
Overall, self-monitoring was the only marginally significant predictor (p = .07) of change in extraversion, but congruence and type of question were not significant.
According to Kunda et al ( 1993 ), retrieval of thoughts in support of the question are mediators that influence ratings of self-concept. However, it seems that the number of disagreeing thoughts and total thoughts were related to a change in extraversion.
When participants were presented with the introverted question, there was a significant positive relationship between disagreeing thoughts, total thoughts and change in extraversion. This study supports Kunda et al. (1993). When participants were presented with the extraverted question, there was a positive relationship between agreeing thoughts and change in extraversion.
Regarding self-monitoring and congruence, there was a significant positive correlation between other thoughts listed and change in extraversion for low selfmonitors. It was interesting that low self-monitors' change in extraversion was not significantly related to agreeing or disagreeing thoughts, perhaps providing support that low self-monitors are less malleable in self-concept than high self-monitors when retrieving agreeing and disagreeing thoughts to a directional question. For high selfmonitors, there was a positive relationship between disagreeing thoughts listed, total thoughts and change in extraversion. This relationship implicates that the process of disagreeing for high self-monitors has more influence on change in extraversion.
Interestingly, there was a significant positive correlation between change in extraversion and number of disagreeing thoughts and total thoughts in response to a congruent question, and there was no relationship between types of thoughts retrieved and change in extraversion when being presented with an incongruent question.

Limitations
Even though these are interesting results, we need to view this empirical inquiry as tentative in nature. One limitation of this study was choosing introversion and extraversion as the directional questions. Introversion tends to be viewed as socially undesirable compared to extraversion. This social desirability was implicated as why introverted individuals rated themselves higher on extraversion in the second session, and could also explain extraverted individuals' withdrawal response to incongruent questions. Another issue with the use of introversion and extraversion is that they are similarly related to low and high self-monitors. There may be a possible confound between extraversion and high self-monitors that needs to be addressed statistically. However, it was decided that introversion and extraversion should be used, since they were used in Kunda et al' s (1993) study and introversion/extraversion are probably the most salient way to categorize people. But this salience also carries with it controversy.
Another limitation, purely statistical, is that the sample size was not large enough to compute a 2x2x2x2 (with one within-subjects variable) ANOV A without producing some biased results. It would have been optimal to have at least 80 participants to complete the analyses. This problem will be fixed in the future, by including more participants.

Implications for Future Research
Some areas that could be addressed in the future are whether these effects replicate when the traits are switched from introversion/extraversion to something less controversial. Perhaps hardworking or something related to conscientiousness could be used. However, different traits may not evoke any type of thinking of whether it pertains to the individual, as it may be innocuous and not pique the attention of the participants.

56
Another area that could be examined is whether directional questions and retrieved thoughts translate into actual behavior in a group setting or in a dyad. If participants are asked if they are extraverted, do they act more extraverted when they are conversing with others in a group? Also, the different social cognitive mechanisms of low and high self-monitors need to be examined, as there seems to be a general lack of research on how each group attends differently to information and makes decisions.

Overview of Study 2
The initial analysis in Study 1 piqued the development of a model of selfconcept that seeks to align two diverging perspectives of self-concept development (see Figure 1). In communication research, the self-concept is theorized to develop through communication acts while in psychology, self-concept is believed to be derived from appraisal of behaviors or perceived personality traits. Therefore, each discipline has a different focus of what "material" accounts for the development of the self-concept.

57
To solve this interdisciplinary schism, and to try to initiate an interdisciplinary approach to understanding self-concept, this researcher proposes that self-monitoring influences both the development of personality traits and communication acts (see Figure 1 and Figure 2). In tum, these personality traits and communication acts directly and indirectly influence each other, which help shape the self-concept.
Among various communication and psychology constructs, the author believes that self-monitoring is the most amenable to mediating the influence of both communication and psychological aspects of the developing self-concept. This is because self-monitoring is a construct that is based on both psychological and communication aspects of interacting with others. With this in mind, two analyses were conducted to try to test the assumption of this model regarding self-concept, psychological communication perspectives of self-concept development and the role of self-monitoring. Therefore, it is hypothesized that communication and psychological measures of similar constructs (such as semantic differential extroversion and Big Five extraversion) will be highly correlated. Results supporting this hypothesis would indicate that the self-concept is comprised of direct appraisals of internal dispositions, or personality traits.

Procedures and Material
Participants primarily filled out various psychological surveys and communication-based surveys. Psychological forms consisted of the Revised Self-Monitoring scale (Snyder & Gangestad, 1986) measures that were similar to the first study which assessed whether the participant perceived him or herself as introverted or extraverted, their thoughts that supported or refuted this question, how variable they were on the personality trait in question, and a self-rating that measured the Big Five. The self-monitoring scale and the Big Five self-rating were the only psychological measures that will be examined in these studies, and the other measures will be used in future studies. Consequentially, the self-rating form was assessed over 2 time periods.
Participants also filled out communication forms that consisted of a measure of Affinity-Seeking Instrument (Bell, Tremblay, Buerkel-Rothfuss, 1987); Communication Adaptability (Duran, 1983); Communication Style (Norton, 1983); Interaction Involvement (Cegala, 1981 ); and a 15-item Semantic Differential Scale (Mccroskey, Hamilton, Weiner, 1974). The Affinity-Seeking Instrument, ASI, (Bell et al., 1987) measures to what extent an individual is able to get others to like him or her. It consists of 13 items scored on a seven point Likert-type scale, ranging from strongly disagree to strongly agree. The ASI consist of two factors: !)Affinity Seeking Competence, which measures the ability to say what is necessary to be seen as interpersonally attractive; and 2) Strategic Performance, which measures the ability to play roles in order to be liked. The ASI has both acceptable reliability (ASC alpha of .80 to .87 and SP alpha of .83) and validity (Bell et al., 1987).
Communication Adaptability (Duran, 1983) consists of the ability to adapt one's interaction goals and behaviors to match what is expected in an interaction. It consists of 30 5-point Likert scale items, with responses that range from "always true of me" to "never true of me". The scale consists of six subscales: 1) social composure, or feeling relaxed in social situations; 2) social experience, or enjoying and participating socially; 3) social confirmation, or maintaining the other's social image; 4) appropriate disclosure, which is adapting one's disclosures to another's intimacy level; 5) articulation, the use of appropriate syntax and grammar; and wit, using humor to diffuse social tension. Reliability of the subscales is in the . 70s and .80s (Duran, 1992) and concurrent validity is strong (Duran, 1983).
The Communicator Style Measure (Norton, 1978) examines communicator style, or the way that one interacts to signal how literal meaning should be taken. The measure consists of 51 items on a five point Likert scale that ranges from "NO!" to "YES!". It consists of 11 different types of communicator style, including friendly, impression leaving, relaxed, contentious, attentive, precise, animated, precise, animated, dramatic, open, dominant, and communicator image. Reliability ranges from .37 for friendly to .82 for dominant (Norton, 1978). The CSM has shown both content validity and criterion-validity (Norton, 1983).
The Interaction Involvement Scale (Cegala, 1981) measures the degree to which people are engaged in their conversations with others. The measure consists of 18 items scored on a 7-point Likert-type scale with responses ranging from "not at all like me" to "very much like me". The measure consists of three subscales, including perceptiveness, attentiveness, and responsiveness. Perceptiveness measures being aware of message meanings, attentiveness means hearing and observing, and responsiveness relates to a person's certainty about how to respond to others during a conversation. Reliability ranges from .86 for responsiveness to .90 for the overall scale. Concurrent validity for the measure is rather strong (Cegala, 1981 ).
The 15-Item Semantic Differential (Mccroskey, Hamilton, Weiner, 1974) consists of 15 items that measure five dimensions of perceived speaker credibility.
These dimensions include sociability, extroversion, competence, composure, and character. Mccroskey et al. (1974) report that the scales have adequate reliability and validity. Participants in the current study answered these questions in self-report format, describing how they perceive themselves on these traits.
Demographics and general characteristics of the participants were assessed, including such things as gender, ethnicity, and age. The procedure was relatively straight forward, there were two sets of forms, either psychological or communication-based, that were presented in a counterbalanced order so that Analyses Overall, there were seven different measures of psychological and communication self-concept were assessed in the study. One measure, the psychological self-rating of the Big Five was distributed twice, raising the total to 8 different measures. Within these measures, there were a variety of subscales that were used for analysis. One exception was the Revised Self-Monitoring Scale, in which it has been argued that the scale should not be divided into the subscales that other researchers have developed through factor-analytic methods. The challenge is raised because as this has not been the original conception of the construct, and the scale as a whole has produced significant intrinsic validity (Snyder & Gangestad, 1986). Accounting for all measures and subscales, 38 different subscales were available for analyses.
The 27 communication scales and subscales (Affinity Seeking has 2 subscales, Communication Adaptability has 6 subscales, Communicator Style had 11 subscales, Interaction Involvement had 3 subscales, and Source Credibility 15-item Semantic Differential had 5 subscales) were examined for content emphasis and then categorized into one of the Big Five domains when applicable. One subscale of Communication Style, communicator image, was dropped from the first analysis as it is theorized to be used as a dependent variable (Norton, 1978), and this author did not think that it would be useful for this analysis.

67
The Affinity Seeking instrument (Table 4)  Semantic differential (Table 4)  Participants' average self-monitoring score (Table 4)  Factor 2 of the Big Five, agreeableness was correlated with another psychological measure, self-monitoring, (Table 5) r = -.07 p = .14. Agreeableness was correlated with communication measures that were theorized to be related to agreeableness, they included Communicator Style subscale of friendly, r = .28 p = .000, contentious, r = -.30 p = .000 and the Affinity Seeking subscale of affinity seeking competence, r = .19 p =.000.
Second Analysis: Change in Self-Concept Change in self-rating or self-concept was analyzed in three ways. In the first approach, there was a comparison of self-ratings within groups across the two testing conditions. In the second approach, there was a "cross-comparison" made comparing the 1st session of one group versus the 2nd session of other group, this analysis would provide insight into whether assessing information about the self in either a psychological or communication-based manner would influence self-rating. In the final analysis, there was a comparison of time 1 and time 2 self-ratings within group based on level of self-monitoring to determine whether high self-monitors showed a greater change in self-rating after completing various psychological or communication measures.
Multiple paired-samples one-tailed t-tests were conducted comparing selfratings across sessions for the group of participants that received communication measures in the first session and psychology measures in the second session ("communication-first") and the group that received psychology measures in the first session and communication measures in the second session ("psych-first") ( Table 7).
Ratings on conscientiousness significantly increased from session 1 to session 2, means of were 35.08 to 37.88 with standard deviations of 6.91 and 7.27, t(105) = 5.46 p = .000. Ratings on emotional stability significantly increased from session 1 to session 2, means of 34.88 to 36.23 with standard deviations of7.32 and 7.44, t(105)= 2.04 p =.022. Ratings on culture significantly increased from session 1 to session 2, means of 38.18 to 39.26 with standard deviations of 6.50 and 5.45, t(105) = 1.85 p =.034.
Cross comparisons were conducted, comparing session 1 self-ratings of the com-first group and session 2 self-ratings of the psych-first group using independentsamples t-tests (Table 8). There was a marginal difference between self-ratings of extraversion for these groups, means of 35.61 and 37.11, standard deviations of7.50 and 7 .32, t(252) = -1.59 p = .057, thus self-ratings were higher in the second session of the psych-first group. There was a marginal difference between self-ratings of agreeableness for these groups, with means of 38.16 and 39.28, and standard deviations of 5.92 and 6.21, t(251) = -1.45 p = .074. There was a significant difference between mean self-ratings of conscientiousness for these groups, means of 34.09 and 37.88, and standard deviations of 5.93 and 7.27, t(196.93) = -4.41 p = .000.
There was a marginal difference between self-ratings of emotional stability for these groups, means of 34. 72 and 36.23, standard deviations of 7.42 and 7.44, t(252) = -1.59 p = .057. There was significant difference between self-ratings of culture for these groups, means were 35.33 and 39.26, with standard deviations of 5.67 and 5.45, t(252) = -5.54 p = .000.
Further cross comparisons were conducted, comparing session 1 self-ratings of the psych-first group and session 2 self-ratings of the com-first group employing the same statistical procedures (Table 8). There was not a significant difference between self-ratings of extraversion for these groups, means of 36.31 and 37 .53 and standard deviations of 7.23 and 7.86, t(254) = -1.29 p = .094. There was not a significant difference between self-ratings of agreeableness for these groups, means of 38.88 and 38.92 and standard deviations of 6.06 and 5.33, t(254) = -.06 p = .951.
There was a significant difference between self-ratings of conscientiousness for these groups, means of38.10 and 35.15 and standard deviations of7.28 and 6.85, t(254) = 3.34 p = .001, indicating that com-first session 2 self-ratings were higher then psychfirst session 1 self-ratings. There was not a significant difference between selfratings of emotional stability for these groups, means of 35.25 and 35.01 and standard deviations of 7.91 and 6.93, t(254) = .26 p = .40. There was not a significant difference between self-ratings of culture for these groups, means of 38.34 and 38.22 and standard deviations of 5.57 and 6.26, t(254) = .16 p = .44.
For the third analysis, participants were divided into groups based on their self-monitoring scores (Table 9). Participants who scored 54 and above were assigned to a group as high self-monitors and groups that scored below 54 were assigned to a group as low self-monitors. These groups were then compared with paired-samples tests to determine whether low or high self-monitors rated themselves differently across sessions. Low self-monitors reported no significant difference between ratings of extraversion in session 1 to session 2 ( Table 9)  High self-monitors reported no significant difference between ratings of extraversion in session 1 to session 2 ( Table 9)  For the second set of analyses, there was mixed support for the hypotheses.
Participants who received communication measures before the second self-rating reported significant increases in their self-ratings in all factors of the Big Five.
Participants who received psychological measures before their second self-rating reported significant increases on factors 3 through 5 of the Big Five. However, the crossed comparisons that were conducted provide greater insight into these findings.
Crossed comparisons indicate that participants who received psychological measures in the first session had marginally higher self-ratings of extraversion at session 2 than participants in the first session before receiving any communication measures. Similar results were found when participants had marginally or significantly higher self-ratings on the other factors of the Big Five compared to participants who had not yet received any measures. Regarding the other crossed comparison, participants that received communication measures had similar selfratings at time 2, compared to session 1 self-ratings of participants who had not received psychological measures. The exception was for conscientiousness, which participants reported a higher level at time 2. 78 Finally, low and high self-monitors reported being similarly influenced by the psychological and communication measures. Both groups reported significant increases in Big Five factors 3 through 5. Thus, there was little support that change in self-rating would be moderated by self-monitoring.
Although these results are interesting and tentatively assess the assumptions of the proposed model, there are some drawbacks that need to be discussed. Like all correlational research, it is nearly impossible to assess the influence that various factors have on other factors. Relational data provide information on the common variance that each shares, but this common variance may be influenced by a third factor (or more) that is not directly observable. Another drawback of this research is the manner in which forms were presented to participants. Participants received each set of forms in a particular order. Theoretically, forms that came later in the set were influenced by the responses on the earlier forms. Forms that came later are also more likely to be filled out differently due to response burden, or participant fatigue. The former issue is typically solved with placing forms in random order, but the number of forms that were used and the design of the study made it logistically impossible to order the forms randomly. As Study 2 evidences, self-monitoring can be conceptualized as both a communication and a psychological construct, as it significantly correlated with measures that were both psychological and communication-based. Also importantly, self-ratings change after appraising both communication acts and psychological traits.
These changes mark as self-concept is influenced and subsequently adjusted by these two types of appraisals. This indicates that both the psychological perspective of the development of the self-concept and communication perspective are both "right". A construct that is germane to both perspectives such as self-monitoring can be used to integrate the two perspectives into one coherent model.
Results from Study 1 remind us that self-monitors are influenced differently by social information. High self-monitors seem to engage in more heuristic processing as evidenced by their lack of disagreeing thoughts, and low self-monitors systematically process information about self and make a decision regarding self (or self adjustment) as evidenced by their disagreeing thoughts. High self-monitors may not negotiate their self-concept in the traditional sense of self-reflected appraisal, but rely more on interactions with others in their social settings. In effect, they develop their self-concept through others. Then, this raises the issue of whether they have a self-concept or a meaningful conceptualization of who they are.
It is crucial to remember that the results implicating self-monitoring as both a communication and a psychological construct are both correlational and tentative in nature and require further analysis, such as structural equation modeling, to determine if both of these variables load on a latent factor of self-concept. Again, with correlational research, it is impossible to assume causality, which is implicated in the proposed model. The current analysis did not determine directionality of causes, but it did provide support for some underlying assumptions regarding the model. This is an important first step towards advocating and outlining a path for further empirical investigation.
Future studies that could be conducted would use actual communication acts to further examine the influence that communication has on the self-concept. It was not feasible, economically or time-wise (as Professors would not allow more then fifteen minutes per session), to conduct a laboratory experiment that would allow the author to examine this influence. Therefore, r esults from the two experiments were extrapolated from a rather artificial environment through artificial means. On the other hand, the surprising patterns and significant results strengthen the posited interdisciplinary joining. This limitation reduces the generalizability of the current results as it now fundamentally cogent to consider the self-concept as malleable and fluid in nature. These artificial limitations reduce the rich amount of social information that occurs when dyads or groups interact with each other. This is a significant problem that has held back the potential prominence of the social cognitive perspective in general and has been extensively discussed in both the disciplines of psychology and communication. This investigation is a major step in the multiple and previously independent perspectives of the communication and psychology fields to yield a stronger explanatory model for self-monitoring. You have been asked to take part in a research project described below. The researcher will explain the project to you in detail. You should feel free to ask questions. If you have more questions later, call Dr. Geoff Leatham, the person mainly responsible for this study, at (401) 874-4735, and he will discuss them with you. You must be at least 18 years old to be in this research project.

82
You have been asked to take part in a study that is concerned with the interaction of personality and situational factors regarding interpersonal interaction for the purpose of furthering information regarding self-concept and self-monitoring.

What will be done:
If you decide to take part in this study here is what will happen: You are asked to participate in two sessions. In the first session you will complete two scales. After finishing the first session, you will then have a one-week absence and you will meet with the researchers again to complete the second session in which you will complete four separate forms. The first form will be a general question, followed by three separate forms. Both sessions will take approximately fifteen minutes each, for a total participation time of approximately thirty minutes. The research data will be kept for at least three years following the completion of the study.

Risks or discomfort:
You will not experience any risks of discomfort as a result of your participation in this study. Anyone that does not want to participate in the two sessions of the study can leave with their professor and he/she will have an alternative exercise provided for you. Both failure to participate (or participation) will not affect the student's grade in any way.

Benefits of this study:
Although there will be no direct benefit to you for taking part in this study, the researcher may learn more about the variables under consideration. You will have the experience of being a participant in a research study.
You have been asked to take part in the research project described below. If you have any questions, please feel free to call Crystal Fonseca at (401)  874-4735 the people mainly responsible for this study.
The purpose of this study is to furthering information regarding self-concept and selfmonitoring. Responses to these items will consist of circling numbers of the various forms that are distributed to you. Confidentiality will be maintained as none of the information will identify you by name. Your password that you create will ensure participant's confidentiality, while also ensuring that the participant's responses can be linked from session 1 and session 2. The participant will write a word and one number that they will easily be able to recall as a password for both the first and the second sessions on their data packets, again ensuring that the participants name is held confidential.
YOU MUST BE AT LEAST 18 YEARS OLD to be in this research project.
If you decide to take part in this study, your participation will involve filling out a total of six questionnaires pertaining to self-monitoring and self-rating.
The possible risks or discomforts of the study are minimal, although some of you may feel some embarrassment answering questions about private matters, but remember that you can refuse to answer any question.
Although there are no direct benefits of the study, your answers will help increase the knowledge regarding the variables of interest (self-monitoring and self-ratings).
Your part in this study is anonymous. That means that your answers to all questions are private. No one else can know if you participated in this study and no one else can find out what your answers were. Scientific reports will be based on group data and will not identify you or any individual as being in this project. The research data will be kept for at least three years following the completion of the study.
The decision to participate in this research project is up to you. You do not have to participate and you can refuse to answer any question. Whatever you decide will in no way penalize you, affect your grade, or status as a student. Instructions: Please read each statement below and then circle a number that best describes how much you agree with each statement in regards to how you see yourself, ranging from "strongly disagree" to "strongly agree". Instructions: Please rate yourself on these 25 traits by circling one number between each pair of adjectives that best reflects how you see yourself.

STUDY 2 FORMS
Informed Consent Documents

Informed Consent Document
This study is concerned with the interaction of personality and situational factors regarding interpersonal interaction. You are asked to participate in two sessions. In the first session you will complete two scales. After finishing the first session, you will then have a two-week absence and you will meet with the researchers again to complete the second session in which you will complete four separate forms. The first form will be a general question, followed by three separate forms. Both sessions will take approximately ten to fifteen minutes each.
Should you wish to discuss any aspect of this study please feel free to contact Dr. Sandra Ketrow in the Communications Studies Department at the University of Rhode Island (401-874-4733). We have no reason to believe that you will experience any discomfort as a result of participation in this experiment. If you do experience any discomfort, please contact Dr. Ketrow.
Your responses will be anonymous so please do not write your name on any of the data collection forms. However, in order to establish what forms participants filled out across the two sessions, you are requested to legibly print your name on a separate sheet.
Instructors of the class will not see the sheet of those who have participated in the study. A research assistant, independent of the class and the instructor, will be the only individual that sees the forms and has access to these forms. After the study is completed, all forms that have been signed will be destroyed. You may withdraw from this study at any time, during or between sessions, with no negative consequences. Simply inform the experimenters that you wish to end participation.
After you have signed the consent form, it will be placed in a file with the other consent forms in no particular order, separate from your data packets. This is done to ultimately protect your anonymity. Your informed consent to participate in this study will end on May 18 You have been asked to take part in a research project entitled "The Malleability of Self-Concept and the Moderating Role of Self-Monitoring" described below. The researcher will explain the project to you in detail. You should feel free to ask questions. If you have more questions later, call Dr. SandraKetrow, the person mainly responsible for this study, at (401) 874-4733, and he will discuss them with you. You must be at least 18 years old to be in this research project.
Description of the project:

97
You have been asked to take part in a study that is concerned with the interaction of personality and situational factors regarding interpersonal interaction for the purpose of furthering information regarding self-concept and self-monitoring.

What will be done:
If you decide to take part in this study here is what will happen: You are asked to participate in two sessions. In the first session you will complete five scales. After finishing the first session, you will then have a one-week respite and you will meet with the researcher again to complete the second session in which you will complete seven separate forms. The first form will be a general question, followed by seven separate forms. Both sessions will take approximately fifteen minutes each, for a total participation time of approximately thirty minutes. The research data will be kept for at least three years following the completion of the study.

Risks or discomfort:
You will not experience any risks of discomfort as a result of your participation in this study. If you do riot want to participate in the two sessions of the study, you can leave with your professor and an alternative exercise will be provided. Neither your failure to participate nor participation will affect the student's grade in any way.

Benefits of this study:
Although there will be no direct benefit to you for taking part in this study, the researcher may learn more about the variables under consideration. You will have the experience of being a participant in a research study.

Confidentiality:
Your part in this study is confidential. None of the information will identify you by name. All records will ensure your confidentiality, while also ensuring that your responses can be linked from session 1 and session 2. You will write a word and one number that you can easily recall as a password for both the first and the second sessions on your data packets, again ensuring that your name is held confidential. The researcher will collect all consent forms without perusing any of them. All forms will be stored in the faculty supervisor's filing cabinet. The purpose of this study is to further knowledge about self-concept and selfmonitoring. Your responses to these items will consist of circling numbers on the various forms that are distributed. Confidentiality will be maintained as none of the information will identify you by name. Your password that you create will ensure your confidentiality, while also ensuring that your responses can be linked from session 1 and session 2. You will write a word and one number that they will easily be able to recall as a password for both the first and the second sessions on your data packet, again ensuring that your name is held confidential.
YOU MUST BE AT LEAST 18 YEARS OLD to be in this research project.
If you decide to take part in this study, your participation will involve filling out a total of six questionnaires pertaining to self-monitoring and self-rating.
The possible risks or discomforts of the study are minimal, although some of you may feel some embarrassment answering questions about private matters, but remember that you can refuse to answer any question.
Although there are no direct benefits of the study, your answers will help increase the knowledge regarding the variables of interest (self-monitoring and self-ratings).
Your part in this study is anonymous. That means that your answers to all questions are private. No one else can know if you participated in this study and no one else can find out what your answers were. Scientific reports will be based on group data and will not identify you or any individual as being in this project. The research data will be kept for at least three years following the completion of the study.
The decision to participate in this research project is up to you. You do not have to participate and you can refuse to answer any question. Whatever you decide will in no way penalize you, affect your grade, or status as a student. Because there is no such thing as a "correct" style of communication, none of the following items has a right or wrong answer. Please do not spend too much time on the items. Let your first inclination be your guide. Try to answer as honestly as possible. All responses will be strictly confidential.
Some questions will be difficult to answer because you honestly do not know. For these questions, however, please try to determine which way you are leaning and answer in the appropriate direction.
The following scale is used for each item: Feeling NO! no ? yes YES! Definition of feeling strong disagreement with the statement disagreement with the statement neither agreement nor disagreement with the statement agreement with the statement strong agreement with the statement For example, if you agree with the following statement, "I dislike the coldness of winter," then you would circle ("yes") on your scale. Some of the items will be similarly stated. But each item has a slightly different orientation. Try to answer each question as though it were the only question being asked. Finally, answer each item as it relates to a general face-to-face communication situation-namely, the type of communicator you are most often. Instructions: Please read each statement below and then circle a number that best describes how much you agree with each statement in regards to how you see yourself, ranging from "strongly disagree" to "strongly agree".
1. I find it hard to imitate the behavior of other people.

Role of Self-Monitoring
Possible construct that may integrate both perspectives