DIFFERENCES BETWEEN MEN IN FACTORS PREDICTING NEGATIVE ATTITUDES TOWARD EFFEMINACY

“Dude, you’re so gay.” “Man up.” “Stop crying like a girl.” Sayings like these reflect an immense pressure on men in American society to conform to a rigid set of masculine ideals (Kimmel & Mahler, 2003; Pascoe, 2007; Porter, 2010). Adherence to these strict masculine guidelines enables men to avoid being perceived as the worst possible thing a man can be perceived as...feminine (Brannon, 1976; Porter, 2010). The present study sought to better understand the specific characteristics of men that predict negative attitudes toward male effeminacy. One hundred and twenty-eight self-identified male participants were measured on their endorsement of traditional masculinity, masculine attribute importance, and selfdiscrepancy along traditional masculine attributes. They were then asked to read one of four male target descriptions that differed according to gender expression (masculine, feminine) and sexual orientation (straight, gay). Participants’ negative attitudes toward the male target were evaluated. Results indicated that hostility and discomfort toward effeminacy were predicted by endorsement of traditional masculinity and masculine attribute importance. Results also indicated that endorsement of traditional masculinity predicted general negative attitudes, hostility, and discomfort toward homosexuality. Lastly, the results showed that endorsement of traditional masculinity and masculine attribute importance predicted negative attitudes toward male targets who embodied traditional masculine qualities.


LIST OF TABLES
Women, has appealed to American society to challenge traditional masculine norms.
He argues that these stringent masculine gender ideals oppress men and as a natural consequence, oppress women as well. In a recent TED talk, Porter (2010) referred to American society's "collective socialization of men" as the "man box," containing the "ingredients of how we define what it means to be a man." The reality of the man box, he argued, is such that the fear men have of the consequences they face in breaking free from the man box is what keeps many from challenging its oppressive nature.
In their recent article, Dragowski and Scharrón-del Río (2014) reviewed much of the extant masculinity literature and outlined several ways in which adherence to traditional masculinity is especially harmful for boys and men as they try to prove their masculinity on a daily basis. The authors discussed empirical studies that provide support for the idea that these rigid behavioral guidelines are detrimental to men's emotional, social, and physical lives. For instance, they summarized studies that found that boys and men avoid displays of emotions unless related to anger or aggression (e.g., Oransky & Marecek, 2009), tend to lack meaningful friendships with other males despite an inner craving for this level of intimacy (e.g., Way 2004Way , 2011Way , 2013 and engage in high risk behaviors (e.g., Way 2011). Perhaps even most disheartening, this immense pressure to act masculine even manifests itself in the nature of the way men attempt suicide, as studies have found that while men are less likely than women to attempt suicide, they are more likely to use fatal methods and therefore complete a suicide (Canetto, 1997;Stack & Wasserman, 2009).
The present study was an effort to help answer Porter's (2010) call to better understand traditional masculinity so it can ultimately be challenged and reconstructed. Specifically, the present study sought to more fully understand the specific characteristics of men that are related to negative attitudes toward male effeminacy, a quality that is particularly threatening to men and will be explored more in depth in the literature review.

LITERATURE REVIEW
While there is no denying that sexual minority youth are targets of homophobic bullying, this type of harassment is also directed at heterosexual youth, particularly males who challenge traditional notions of masculinity (Kimmel & Mahler, 2003;Poteat, Mereish, DiGiovanni, & Koenig, 2011). A recent study by Poteat, DiGiovanni and Scheer (2013) found that boys engaged in homophobic behavior to a greater extent than girls and reported to be more negatively affected by homophobic bullying than girls. In a larger scale nationwide study, 1,965 students in grades 7 through 12 reported that being called "gay" or other homophobic epithets was rated by male students -both heterosexual and homosexual -as their most negative experience of harassment (Hill & Kearl, 2011). Specifically, heterosexual boys reported greater distress when they were victimized in ways that included homophobic epithets than heterosexual boys who had been victimized in ways not including the use of these terms (Swearer, Turner, Givens, & Pollack, 2008). In her observations of high school students, Pascoe (2007) observed this unique connection between homophobia and masculinity as reflected in her statement that "the sort of gendered homophobia that constituted adolescent masculinity did not constitute adolescent femininity" (p. 56). In other words, Pascoe (2007) observed the fundamental connection of homophobia to masculinity, but not to femininity.
Emerging evidence that males are affected by homophobic bullying in a manner distinctive from females supports the idea of a gendered homophobia; one that has less to do with direct animosity toward homosexual individuals and more to do with the nature and construction of masculinity (Pascoe, 2007). To better understand this relationship, we must first examine the construction of American masculinity.

Conceptualization of American Masculinity
Traditional masculinity. Brannon (1976) parsed what he believed to be the essence of traditional American masculinity into four tenets: "No Sissy Stuff," reflecting the idea that masculine men must avoid associating with anything feminine; "The Big Wheel," characterizing the notion that masculine men should seek to obtain wealth, power, and status; "The Sturdy Oak," illustrating the idea that truly masculine men are reliable and avoid showing emotions in a feminine way (for example, by crying); and lastly, "Give 'em Hell," encapsulating the idea that masculine men should value aggression and risk-taking.
Critical to the traditional masculine identity is its placement in direct opposition to femininity (Brannon, 1976;Herek, 1986;O'Neil, 2008). Brannon (1976) argued that at its core, masculinity is the "relentless repudiation of the feminine" (Kimmel, 2003, p. 123) and men should actively and ardently pursue the avoidance of activities or behaviors that would be considered feminine. Traditional gender norms construct masculinity and femininity as distinct categorical entities in a way that subordinates and devalues femininity. In a recent TED Talk, Porter (2010) shared the following poignant anecdote of how femininity is denigrated by traditional  (Brannon, 1976;Connell, 1995). In addition to its separation from femininity, American masculinity differentiates among men: between those who "measure up" and those who do not.
The masculine hierarchy. In conceptualizing American masculinity, gender theorist R. W. Connell (1995) proposed the idea of a masculine hierarchy. At the top of the hierarchy is hegemonic masculinity, consisting of ideal men who fully embody traditional masculine characteristics (Connell & Messerschmidt, 2005;Pascoe, 2007).
Within the body of the hierarchy are masculinities that contain characteristics of hegemonic masculinity to varying extents but are considered less acceptable forms of masculinity (Connell & Messerschmidt, 2005). Complicit masculinity is a category describing men who do not necessarily embody traditional masculine qualities but who reap the benefits of masculine hegemony and are therefore reluctant to challenge this patriarchal system (Dragowski & Scharrón-del Río, 2014). Subordinated masculinity is comprised of men who occupy the lowest level of the hierarchy who are oppressed as a result of lacking important characteristics of traditional masculinity 1 , 1 Note that many theorists and researchers argue that all men within the hierarchy are oppressed by traditional and hegemonic masculinity (e.g., Connell & Messerschmidt, 2005;Dragowski & Scharrón-del Río, 2014) for example homosexual men. Connell (1995) also recognized the complex interactions between masculinity and culture in including marginalized masculinities within the hierarchy, describing men who may be considered powerful in terms of their masculine characteristics but not necessarily in terms of their class or race.
Hegemonic masculinity. The idea that hegemonic masculinity, a masculinity that fully embodies all four tenets described by Brannon (1976), is a standard that American men are held to and told to aspire to is idealistic and unreasonable (Pleck, 1981). Take for example the following "hypothetical man" described by Brannon (1976) reflecting an Americanized hegemonically masculine man: [He] never feels anxious, depressed, or vulnerable, has never known the taste of tears, is devoid of any trace or hint of femininity. He is looked up to by all who know him, is a tower of strength both physically and emotionally, and exudes an unshakable confidence and determination that sets him apart from lesser beings. He is aggressive, forceful, and daring, a barely controlled volcano of primal force. (p. 36) Pleck (1981) argued that hegemonic masculinity creates a crisis for men by holding them to unrealistic and unattainable standards, which are unachievable for many American men (Brannon, 1976;Connell, 1995). As a result, Pleck (1981) reasoned that many men are left feeling inadequate in terms of their own masculinity.

Masculinity: The Precarious Identity
The social pressure placed on men to behave in a way that mirrors an unrealistic male archetype has led researchers to examine how this has impacted the nature of masculinity as a gender identity. Specifically, how it is tenuous and fragile.
This instability arises as masculinity or manhood, unlike femininity or womanhood, is not an identity secured at birth (Vandello, Bosson, Cohen, Burnaford, & Weaver, 2008). Instead, it must be achieved through what Pleck (1981) describes as a "risky, failure-prone process" (p. 20). Vandello and colleagues (2008) conducted a series of studies that provided empirical support for this notion. For example, they found that participants, both male and female, found the statement "no longer being a man" more logical and conceivable than the statement "no longer being a woman," suggesting that compared to femininity, masculinity is rather unstable. As such, "precarious," which is defined by Merriam-Webster as something "characterized by a lack of security or stability that threatens with danger" ("Precarious," n.d.) was the term that Vandello and colleagues (2008) felt most appropriately described masculinity as a gender identity.
This masculine hierarchy that subordinates men who do not meet its rigid standards (Connell, 1995;Connell & Messerschmidt, 2005;Laberge & Albert, 1999) acts like a ubiquitous, looming, emasculating cloud where men who possess few traditional masculine characteristics are at risk for being viewed as less manly than those who embody it more fully (Vandello et al., 2008). As a result, men are highly motivated to seek ways in which to maintain and bolster their position within the hierarchy (Pascoe, 2007). One pervasive and pernicious way that men accomplish this is by threatening the masculinity of other men by labeling them with, in the most emasculating fashion, the term "fag" (Kimmel & Mahler, 2003;Pascoe, 2007).
"Fag discourse." "Walk down any hallway in any middle school or high school in America and the single most common put-down that is heard is 'That's so gay.' It is deployed constantly, casually, unconsciously" (p. 1453). This observation made by Kimmel and Mahler (2003) is followed by their argument that boys experience this discourse "not as an assessment of their present or future sexual orientation but as a commentary on their masculinity" (p. 1453). Similar observations are made in Pascoe's (2007) ethnographic field study, "Dude, You're a Fag," in which part of her work explored the use of homophobic epithets by male students in a typical American high school; a phenomenon she refers to as "fag discourse" (p. 54). She observed that male students who exhibited behaviors considered to be feminine or related to other nonmasculine attributes such as "being stupid or incompetent" (p. 57) were often the recipients of this type of harassment. Pascoe (2007) delineates between the use of the terms "gay" and "fag" in that both males and females used gay to describe anything as stupid, whereas fag was almost exclusively used by boys directed at other boys to emphasize that individual's lack of masculinity.
These observations are reflected in the idea inherent in the findings of Glick and colleagues (2007) that an important distinction exists between what it means for a male to be homosexual versus what it means to not be masculine; mainly that the two concepts are not inextricably linked. One male student interviewed by Pascoe (2007) expressed this idea when he said, "Being gay is just a lifestyle. It's someone you choose to sleep with. You can still throw around a football and be gay" (p. 58). This sentiment reflects the idea that embodying traditional masculine traits, even if only a few, positions men higher on the masculine totem pole than those who display feminine traits, regardless of that individual's sexual orientation.
Examining fag discourse has helped to inform an understanding as to the purpose it serves men, mainly to help bolster their position within the masculine hierarchy. Pascoe (2007) proposed the "hot potato hypothesis" (p. 61) by observing that threatened men could restore their masculinity by emasculating other men and therefore shift their position within the hierarchy. She explained that, "In this way the fag became a hot potato that no boy wanted to be left holding" (p. 61). Given that traditional masculine ideology appears to be at the center of fag discourse between men (Pascoe, 2007), it seems important to include it as a characteristic that might predict negative attitudes toward effeminacy.

Negative Attitudes Toward Effeminacy as a Reaction to Masculinity Threats
Research in masculinity is beginning to investigate the notion that males are particularly vulnerable to threats against their masculinity (Bosson & Vandello, 2011;Glick et al., 2007;Vandello et al., 2008). Glick, Gangl, Gibb, Klumpner, and Weinberg (2007) examined this notion by measuring negative responses toward effeminate and masculine gay males after experiencing a threat to their masculinity.
In their study, male participants completed a gendered personality test and were subsequently given feedback that their results fell in either the average masculine or feminine score range. After receiving the feedback, participants completed a survey of attitudes toward a masculine or effeminate homosexual male. The results indicated that males who experienced a threat to their masculinity by being told their score fell in the average feminine range exhibited more negative affect toward the effeminate gay male target than the masculine one. While this study is important because it suggested that when males experienced a threat to their masculinity they tended to express greater negative attitudes toward effeminacy rather than homosexuality, it did not examine if men differed in their reaction toward other men as a function of certain personal characteristics (e.g., if they differed in how much they endorsed traditional masculinity) that are related to masculinity. Understanding these differences therefore seems important in understanding men's attitudes toward one another.
A study by Theodore and Basow (2000) attempted to explore some of these individual differences in relation to attitudes of homophobia. In their study, male participants were asked to complete the Attitudes Toward Homosexuality Scale (ATH) (Millham, San Miguel, & Kellogg, 1976) as well as measures designed to discern the extent of gender self-discrepancy and gender attribute importance. To measure gender self-discrepancy, Theodore and Basow (2000) used the Ought Self Questionnaire (OSQ) that was adapted from the Self-Attribute Questionnaire (SAQ) (Pelham & Swann, 1989). They measured gender attribute importance using the Attribute Importance Questionnaire (AIQ; also modified from Pelham and Swann's (1989) original SAQ measure). They found that gender self-discrepancy and gender attribute importance as well as the interaction between these variables predicted homophobia in their sample.
Gender self-discrepancy. Self-discrepancy theory posits that multiple perspectives of the self exist and that these different perspectives are often incongruent with one another. Higgins (1987) proposed that inconsistencies exist among three distinct "selves" including the "actual" self (which is the self as it is realistically perceived), the "ideal" self (which is the self in a desired form), and the "ought" self (which is the way the self should be based on a set of standards). Additionally, Higgins (1987) distinguished between two perspectives that can be taken of the self.
These are the views of the "own," which is the perspective that an individual takes of his or her own self and that of the "other," which is an outsider perspective for example, a friend, parent, or society.
Higgins ' (1987) theory contributed to the notion that experiencing contradictory beliefs about the self can lead to feelings of discomfort and other negative emotions. While each type of discrepancy has the potential to negatively affect males and females, Higgins (1987) argued that the discrepancy between the "actual/own" and "ought/other" self is characteristic of the conflict many men experience with traditional masculinity. An actual/own versus ought/other conflict is one in which the characteristics a man feels he actually possesses are not consistent with those he feels others around him believe he should posses. In this case, the discrepancy is between the view men have of themselves as masculine and the normative standard that society proscribes for masculinity. Higgins (1987) argued that specific self-discrepancy types are associated with specific negative emotions. If men are more likely to experience an actual/own versus ought/other self-discrepancy conflict, then according to Higgins' (1987) theory they are also more likely to experience the specific kinds of negative emotions found to be associated with this type of discrepancy, including feelings of vulnerability, agitation, and fear. To support this hypothesis, Higgins (1987) surveyed male and female undergraduates on their perceived actual/own versus ought/other discrepancy and found that this particular discrepancy was significantly associated with emotions of fear and threat, including feeling ashamed, not proud, and experiencing a constant preoccupation with the discrepant state.
To illustrate how the ought/other masculine self has the potential to create salient discrepancies for American men, consider Goffman's (1963) description of an "ideal" American male: "In an important sense, there is only one complete unblushing male in America: A young, married, white, urban, northern, heterosexual Protestant father of college education, fully employed, of good complexion, weight, and height, and a recent record in sports" (p. 128). By its very definition, this conceptualization excludes certain American men. While some men may be able to make changes to more fully embody this ideal, the reality is that many of these characteristics cannot be altered (e.g., race, sexual orientation, certain physical characteristics) further perpetuating a state in which many men experience a continual masculine selfdiscrepancy. Related to gender self-discrepancy is gender attribute importance.
Gender attribute importance. Gender attribute importance is a concept extended from the idea of attribute importance, which posits that characteristics of great personal value to a person's self-identity will have a significant impact on that person's sense of self-worth (Pelham & Swann, 1989). Specifically related to masculinity, gender attribute importance is the extent to which traditional masculine characteristics are important to a man's own masculine identity. For example, while traditional American masculinity is characterized by aggression, some men may value aggression in terms of their masculinity to a greater extent than other men. Theodore and Basow (2000) found that the impact of experiencing an actual/own versus ought/other discrepancy depends on the extent to which the ought/other traits are important to an individual's identity. In other words, they found that if traditional masculine characteristics were important to man's masculine identity and they experienced an actual/own versus ought/other discrepancy, then they were more likely to exhibit attitudes of homophobia. These findings suggested that masculine attribute importance is an essential variable that needs to be considered alongside masculine self-discrepancy and endorsement of traditional masculinity in predicting negative attitudes toward effeminate males. To date, no study has examined how masculine attribute importance moderates the relationship between masculine self-discrepancy and negative attitudes toward effeminacy.

Use of Vignettes
The use of analogue methods has been found to be an effective alternative to non-analogue forms of research particularly when the hypotheses are investigating sensitive topics, such as gender and sexuality, and there is a concern that demand characteristics might bias participant responses (Huebner, 1991). Furthermore, vignettes facilitate the creation of situations that emulate reality where judgments are made (Alexander & Becker, 1978).

Purpose of Study
A review of the literature has revealed that to date, no study has examined how endorsement of traditional masculinity, masculine attribute importance, and the interaction between self-discrepancy and masculine attribute importance predicts negative attitudes toward effeminacy in other men. The aim of the present study was to extend previous research by Glick and colleagues (2007) and Theodore and Basow (2000) in an attempt to understand how these specific characteristics differentiated between men in predicting negative attitudes toward effeminacy.
Participants were asked to rate their negative responses toward one of the four male target descriptions in the vignettes similar to those used in Glick et al.'s study (2007). The male character in the vignette, "Michael," was portrayed as a college student whose gender expression and sexual orientation was manipulated resulting in four separate profiles. These included a masculine gay Michael, a masculine straight Michael, an effeminate gay Michael, and an effeminate straight Michael. The masculine types were described as preferring stereotypical masculine activities, organizations, career goals, and possessed traditional masculine traits. Likewise, the effeminate types were described as preferring stereotypical feminine activities, organizations, and career aspirations, as well as possessed female-specific traits.
Michael's sexual orientation was indicated by reference to his girlfriend or boyfriend (see Appendix A).
There were five independent variables in this study. Three were continuous selection variables including (a) the extent to which male participants endorsed traditional masculine characteristics, (b) participants' perception of their masculine self-discrepancy, and (c) the importance of traditional masculine attributes to participants' masculine identity. Additionally, the four target descriptions contained two randomized independent variables including (a) gender expression, either masculine or effeminate; and (b) sexual orientation, either straight or gay. There were four continuous dependent variables that measured four dimensions of participants' negative attitudes (general, fear, hostility, and discomfort) towards "Michael" (the male target description).

Research Hypotheses
This study sought to investigate how endorsement of traditional masculinity, masculine attribute importance, and the interaction between masculine selfdiscrepancy and the masculine attributes predicted negative attitudes toward effeminacy in other males. It also examined how these variables predicted subsets of negative attitudes including fear, hostility, and discomfort.
Hypothesis 1. The first hypothesis focused on endorsement of traditional masculinity's ability to predict general negative attitudes, as well as more specific types of negative attitudes including fear, hostility, and discomfort. Hypothesis 1 stated that there would be an interaction effect of the male target's gender expression across levels of endorsement of traditional masculinity for (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort, such that participants who highly endorsed traditional masculinity were expected to express greater (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort toward the effeminate male targets than the masculine targets, regardless of the target's sexual orientation. Support for this hypothesis was expected to be indicated by an overall significant regression model as well as a significant Endorsement of Traditional Masculinity (ENDMASC) x Gender Expression (GENEXP) interaction term across the four regression models predicting each outcome of (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort.

Hypothesis 2. The second hypothesis focused on masculine attribute
importance and its ability to predict general negative attitudes, as well as more specific types of negative attitudes including fear, hostility, and discomfort. Hypothesis 2 stated there would be an interaction effect of male target's gender expression across levels of masculine attribute importance for (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort, such that participants who expressed greater levels of masculine attribute importance (i.e., feeling that traditional masculine characteristics are highly important to their sense of masculinity) were expected to express greater (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort toward the effeminate male targets than the masculine targets, regardless of the target's sexual orientation. Support for this hypothesis was expected to be indicated by an overall significant regression model as well as a significant Masculine Attribute Importance (MAI) x Gender Expression (GENEXP) interaction across the four regression models predicting each outcome of (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort.

Hypothesis 3.
The third hypothesis focused on the interaction of masculine attribute importance and self-discrepancy on predicting general negative attitudes, as well as more specific types of negative attitudes including fear, hostility, and discomfort. Hypothesis 3 posited that participants with a higher sense of selfdiscrepancy between their actual/own versus ought/other masculine identity were expected to express (a) general negative attitudes, (b) fear, (c) hostility, and (d) discomfort towards the effeminate male targets than the masculine male targets, only if they also expressed a high level of masculine attribute importance. Support for this hypothesis was expected to be indicated by a significant regression model as well as a significant Masculine Attribute Importance (MAI) x Masculine Self-Discrepancy (MSD) x Gender Expression (GENEXP) interaction term that when probed would indicate the hypothesized interaction effect.

Participants
The sample for this study was comprised of 128 self-identified male undergraduates from the University of Rhode Island ( 2.3%). One participant did not indicate his sexual orientation.

Measures
Participants completed five demographic questions and four measures.

Demographic questions. Participants completed a brief demographic
questionnaire where they were asked to indicate their age, gender, class year, ethnicity, and sexual orientation (see Appendix B).
Endorsement of traditional masculinity. The extent to which participants endorsed traditional masculine ideology was measured using the Male Role Norms Inventory -Short Form (MRNI-SF) (see Appendix C) (Levant, Hall, & Rankin, 2013).
The MRNI-SF is comprised of 21 items representing the seven subscales in the original measure. The subscales included "Avoidance of Femininity," "Negativity toward Sexual Minorities," "Self-reliance through Mechanical Skills," "Toughness," "Dominance," "Importance of Sex," and "Restrictive Emotionality." Participants responded to items such as "A man should never admit when others hurt his feelings" on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Mean scores of all the items were calculated with higher scores indicative of greater levels of endorsement of traditional masculine ideology (possible score range of 21-147). In the present study, the coefficient alpha was .92.
Masculine Self-Discrepancy. To measure participants' perceptions of their self-discrepancy along the masculine attributes, a modified version of The Ought Self Questionnaire (OSQ) (Theodore & Basow, 2000) was used (see Appendix D). This modified version was called the Masculine Self-Discrepancy Scale (MSDS). The original OSQ scale was adapted from the Self-Attribute Questionnaire originally developed by Pelham and Swann (1989) and was divided into three subscales, (a) masculine, (b) feminine, and (c) bipolar (masculine-feminine). In previous studies (e.g., Theodore & Basow, 2000), the bipolar scale yielded low reliability (Cronbach alpha of .56). Before deciding to discard this subscale from the current study, the original factor analytic structure of the items were reviewed. Six of the eight items were shown not to load onto either the masculine or feminine factor for college males and as such the subscale was not used in this study. Furthermore, the following two items, "aggressive" and "dominant," loaded highly on the masculine scale (loadings .42 to .54) but were not included as part of the masculine subscale in the original study. They were however, included on the masculine scale for the present study.
The resulting 10-item measure used a 5-point Likert scale (1 (0%) = not at all alike, 5 (100%) = completely alike) and asked participants to consider the discrepancy between their actual/own and ought/other perspectives. Items were reverse coded, with higher scores indicative of greater discrepancies between how masculine participants perceived themselves to be versus how mainstream American society believes masculine men ought to be (possible score range of 21 to 105). Only the items from the masculine subscale were used in the final analyses. In the present study, the coefficient alpha was .83.

Masculine attribute importance.
To determine how important masculine attributes were to participants' perceptions of their masculine identity, a modified version of The Attribute Importance Questionnaire (AIQ) (Theodore & Basow, 2000) was used (see Appendix E). The modified scale, called the Masculine Attribute Importance Scale (MAIS), contained the same masculine attributes as the OSQ except that the items were reordered and slightly reworded so as to minimize the probability that participants would selectively rate the importance of attributes based on how they previously rated themselves on the OSQ (Theodore & Basow, 2000). Participants were asked to rate the importance of each attribute to their masculine identity on a 9- Each of the students found the directions to be somewhat cumbersome and difficult to understand. Most of the confusion had to do with the concept of the "ought self." As a result, the directions were modified to reflect the idea that participants should think about the "ought self" as mainstream American society. Secondly, adjustments were made to the Likert scale in terms of its visual presentation. The initial scale was a 9-point scale with three descriptions below that served as a guide.
After pilot testing, the scale was reduced to a 5-point scale that included percentages instead of numbers to represent the extent of the participant's self-discrepancy.
Additionally, visual representations of the percentages reflecting the actual/ought discrepancy were added for clarity.
Lastly, modifications were made to the way in which each item was presented to participants. In the initial scale, participants read the directions and then were presented with each individual item (e.g., "Aggressive"). As a result of the pilot testing feedback that the directions (even after revised) were still somewhat lengthy, the items were presented to participants as a question that would remind participants what to consider without having to refer back to the directions. Instead of being shown just the item word, the participants would now see a question such as, "How aggressive YOU think you are in comparison to how aggressive SOCIETY thinks you, as a man, should be?"

Descriptive Statistics and Assumptions
Descriptive analyses indicated that the data met the required assumptions of normality, linearity, and homoscedasticity. The data did not violate the assumption of normality as the skewness and kurtosis values indicated acceptable skewness values of -1 to 1 and kurtosis values of -1.5 to 2. Also, histograms and P-P Plots of regression standardized residuals for each predictor indicated that the data met the assumption of normality. A review of bivariate scatter plots indicated that the assumptions of linearity and homoscedasticity were met for all predictors. Lastly, correlations between predictor variables in each analysis indicated that multicollienarity was not an issue as there were no correlations greater than r =.70.
The outlier labeling method was used to assess the data for outliers. Based on this technique, values that exceeded two standard deviations were not considered outliers if their mean was less than the value achieved by multiplying the Interquartile Range by a constant value of 2.2 (Hoaglin & Iglewicz 1987;Tukey 1977). Using this method, three outliers were identified on the Masculine Attribute Importance scale.
The process of winsorizing was selected as the most appropriate method to deal with these values. The outliers were modified to the next highest or lowest value in the distribution that were not considered outliers. This process is recommended when there are small amounts of legitimate outliers primarily resulting from measurement error (Hawkins, 1980).
The three selection variables were centered prior to running the regression analyses. This is advised when regression analyses include interaction terms, as centered variables improve the interpretability of the results and enhance the comparability of scales. Mean centering involves subtracting a variable's raw score from its mean (Aiken & West, 1991).

Missing Values
A missing values analysis was run in SPSS indicating that no variable had more than 5% missing values. As a result, listwise deletion technique was utilized in dealing with any missing data. that ENDMASC x GENEXP would be a significant predictor of general negative attitudes, ENDMASC x SEXORT was the only predictor to add significantly to the prediction of general negative attitudes (β = .32, p < .000). Regression coefficients and standard errors can be found in Table 2.
Regression coefficients and standard errors can be found in Table 4. When examining the effeminate targets, there was no interaction between gender expression and sexual orientation in predicting hostile attitudes. As can be seen in Figure 3, participants who did not highly endorse traditional masculinity expressed less hostile attitudes toward both straight and gay effeminate targets in comparison to participants who highly endorsed traditional masculinity and who exhibited much greater attitudes of hostility toward all effeminate targets. For all masculine targets, there was an interaction between gender expression and sexual orientation of the target Figure 2. ENDMASC x SEXORT interaction predicting hostility.  Figure 3. ENDMASC x GENEXP x SEXORT interaction in predicting hostility towards masculine targets.  GENEXP significantly predicted attitudes of discomfort (β = .24, p = .005) as did ENDMASC x SEXORT (β = .38, p < .000). Regression coefficients and standard errors can be found in Table 5. Note. ENDMASC = Male Role Norms Inventory, GENEXP = Gender Expression, SEXORT = Sexual Orientation, B = unstandardized regression coefficient, SEB = Standard error of the coefficient, β = standardized coefficient. * p < .05. ** p <.01. Figure 5 illustrates the interaction between gender expression for levels of endorsement of traditional masculinity such that participants who highly endorsed traditional masculinity expressed greater attitudes of discomfort towards effeminate male targets than masculine male targets, regardless of sexual orientation whereas males who did not highly endorse traditional masculinity expressed greater attitudes of discomfort towards masculine male targets. Similarly, Figure 6 illustrates the interaction between sexual orientation for levels of endorsement of traditional masculinity such that participants who more highly endorsed traditional masculinity expressed more attitudes of discomfort toward gay male targets than straight targets, regardless of gender expression, whereas males who did not highly endorse traditional masculinity expressed greater attitudes of discomfort towards straight male targets.

Endorsement of Traditional Masculinity
Masculine Effeminate Figure 6. ENDMASC x SEXORT interaction in predicting discomfort attitudes.

Hypothesis 2
General negative attitudes. A multiple regression was run to predict general negative attitudes from masculine attribute importance (MAI), gender expression, sexual orientation, and the following interactions: MAI x GENEXP, MAI x SEXORT, and MAI x GENEXP x SEXORT. The correlations of the variables are shown in Table 6. This linear combination of predictors did not significantly predict general negative attitudes, R 2 = .08, adjusted R 2 = .04, F(6,121) = 1.82, p = .101, however MAI x GENEXP, which was predicted a priori to be a significant predictor of general negative attitudes, was significant (β = .22, p = .028). Regression coefficients and standard errors can be found in Table 7. A new analysis was performed with only the lower order terms included in that interaction term. The new model, while closer to an acceptable p-value remained non significant, R 2 = .05, adjusted R 2 = .03,  Note. MAI = Masculine Attribute Importance, GENEXP = Gender Expression, SEXORT = Sexual Orientation, B = unstandardized regression coefficient; SEB = Standard error of the coefficient; β = standardized coefficient. * p < .05. ** p <.01.

Fear.
A second multiple regression was performed to predict fear attitudes from masculine attribute importance, gender expression, sexual orientation, and the following interactions: MAI x GENEXP, MAI x SEXORT, and MAI x GENEXP x SEXORT. For correlations of the variables refer to Table 6. Contrary to the stated hypothesis, the overall model did not significantly predict fear attitudes, R 2 = .06, adjusted R 2 = .01, F(6,121) = 1.23, p = .296, nor did MAI x GENEXP significantly add to the prediction of fear attitudes (β = -.05, p = .650). Regression coefficients and standard errors can be found in Table 8.

Hostility.
A third multiple regression was performed to predict hostility attitudes from masculine attribute importance, gender expression, sexual orientation, and the following interactions: MAI x GENEXP, MAI x SEXORT, and MAI x GENEXP x SEXORT. Refer to Table 6 for the correlations of the variables. The linear combination of predictors did not significantly predict hostility attitudes, R 2 = .06, adjusted R 2 = .01, (6,121) = 1.30, p = .265). Contrary to the hypothesis, MAI x GENEXP did not significantly predict hostile attitudes (β = .05, p = .641). Instead, MAI x SEXORT was a significant predictor of hostile attitudes (β = .21, p = .041), but was not probed further as this relationship was not predicted a priori. Regression coefficients and standard errors can be found in Table 9.

MAI x GENEXP x SEXORT. Refer to
Regression coefficients and standard errors can be found in Table 10. Figure 7 illustrates the interaction effect of gender expression for levels of masculine attribute importance such that participants who felt that traditional masculine characteristics were highly important to their own sense of masculinity expressed greater attitudes of discomfort toward effeminate male targets than males with a lower level of masculine attribute importance, regardless of the sexual orientation of the male target description.
In other words, males who did not feel that traditional masculine characteristics were very important to their sense of masculinity expressed greater attitudes of discomfort toward the masculine male targets, regardless of whether the target was straight or gay.   Table 11. Support was not found for this part of the hypothesis, as the linear combination of predictors was not significant, R 2 = .07, adjusted R 2 = .01, F(8,119) = 1.12, p = .357, nor was MAI x MSD x GENEXP a significant predictor of general negative attitudes (β = .17, p = .322). Regression coefficients and standard errors can be found in Table 12.  The overall combination of linear predictors was not significant, R 2 = .09, adjusted R 2 = .02, F(8,119) = 1.37, p = .215, nor was the a priori interaction between MAI x MSD x GENEXP a significant predictor of fear attitudes (β = .02, p = .906). Regression coefficients and standard errors can be found in Table 13.
x MSD x GENEXP, MAI x MSD x SEXORT, and MAI x MSD x GENEXP x SEXORT. Refer to Table 11 for correlations of the variables. Similarly, no support was indicated for the hypothesized moderation effect on attitudes of hostility toward effeminate targets. The linear combination of predictors was not significant, R 2 = .05, adjusted R 2 = -.01, F(8,119) = .805, p = .599, nor was the a priori interaction MAI x MSD x GENEXP a significant predictor of hostility attitudes (β = .18, p = .281).
Regression coefficients and standard errors can be found in Table 14.

Discomfort.
A fourth multiple regression was performed to predict discomfort attitudes from masculine attribute importance, masculine self-discrepancy, sexual orientation, gender expression, and the following interaction terms: MAI x MSD, MAI x MSD x GENEXP, MAI x MSD x SEXORT, and MAI x MSD x GENEXP x SEXORT. Refer to Table 11 for correlations of the variables. Lastly, no support was indicated for the hypothesized moderation effect on discomfort attitudes towards effeminate male target descriptions as the linear combination of predictors was not significant, R 2 = .07, adjusted R 2 = .01, F(8,119) = 1.14, p = .343, nor was the a priori interaction between masculine attribute importance and gender expression in predicting discomfort attitudes (β = .09, p = .598). Regression coefficients and standard errors can be found in Table 15. Note. MAI = Masculine Attribute Importance, MSD = Masculine Self-Discrepancy, GENEXP = Gender Expression, SEXORT = Sexual Orientation, B = unstandardized regression coefficient; SEB = Standard error of the coefficient; β = standardized coefficient. * p < .05. ** p <.01. Note. MAI = Masculine Attribute Importance, MSD = Masculine Self-Discrepancy, GENEXP = Gender Expression, SEXORT = Sexual Orientation, B = unstandardized regression coefficient; SEB = Standard error of the coefficient; β = standardized coefficient. * p < .05. ** p <.01.

DISCUSSION
The purpose of this study was to examine if characteristics relating to traditional masculinity would differentiate between males regarding negative attitudes toward an effeminate male peer. Specifically, this study aimed to understand how endorsement of traditional masculinity, masculine attribute importance, and the interaction of masculine self-discrepancy and masculine attribute importance predicted negative attitudes including fear, hostility, and discomfort toward a male vignette character, "Michael" who had one of four character profiles: masculine/straight, masculine/gay, effeminate/straight, or effeminate/gay. Three significant main findings emerged from this study including support for the main hypotheses regarding negative attitudes toward effeminacy, as well as results not initially predicted (including support for the ability to predict negative attitudes toward homosexuality as well as toward traditional masculinity).

Negative Attitudes Toward Effeminacy
Research supporting the idea that negative attitudes toward homosexuality actually reflect negative attitudes toward effeminacy in other males (Glick et al., 2007;Kimmel & Mahler, 2003;Pascoe, 2007), lent support for the hypothesis that negative attitudes of certain males would be greater toward effeminate "Michael" regardless of "Michael's" sexual orientation. Results of the present study suggested that the degree to which males expressed hostility and discomfort toward an effeminate male peer was related to the extent to which they endorsed principles of traditional masculinity.
Male participants who agreed with the notions of traditional masculine ideology as measured by Levant, Hall, and Rankin's (2013) Male Role Norms Inventory Short-Form, expressed feelings of hostility and discomfort after reading a description of a male college student who embodied stereotypical feminine characteristics (such as pursuing a college degree in the arts, not being emotionally stoic, and who engaged in activities traditionally considered feminine, such as shopping).
In addition to the role of endorsement of traditional masculinity in predicting hostility and discomfort toward effeminacy in other males, masculine attribute importance was also found to predict attitudes of discomfort toward effeminate males.
For participants who felt that traditional masculine characteristics such as aggression, competitiveness, or courage were important to their sense of masculinity, reading a description of an effeminate male peer evoked attitudes of discomfort, regardless of that peer's sexual orientation.
The definition of manhood as summarized by Brannon (1976) includes the tenet "no sissy stuff," underlying the importance of avoiding any feminine behavior or activity as one of the fundamental components of being a man. In considering this definition, effeminate "Michael" violates this condition of manhood. When asked about how participants felt after reading about effeminate (gay or straight) "Michael," those who highly endorsed traditional masculine ideology and characteristics expressed feelings of hostility and discomfort toward him. It seems reasonable to assume that men might feel a sense of hostility and discomfort toward males who violate traditional notions of masculinity given the fact that men often face harsh consequences for nonmasculine behavior or appearance. This finding elucidates some of the nuances that may underlie the emotional component of gender policing that occurs between men; other men's gender non-conforming behavior may elicit attitudes of hostility and discomfort just by being the "wrong sort of boys" (Kenway, Willis, Blackmore, & Rennie, 1997, p. 103).
Relation to previous research findings. Findings in the present study diverge somewhat from Glick and colleagues (2007) examination of negative attitudes toward effeminacy in a few important ways. Firstly, Glick et al. (2007) found that only males who experienced a threat to their masculinity (e.g., were told their score fell in the "feminine" personality range on a gendered personality test) exhibited an increase in negative attitudes toward effeminate but not masculine gay targets. The researchers noted that males who experienced this type of threat exhibited an increase in fear, hostility, and discomfort in response to reading a vignette about an effeminate gay male.
Secondly, it is also important to note that Glick et al.'s (2007) study featured male vignettes that differed in gender expression (masculine or effeminate) but not in sexual orientation (only gay). While their results support the notion that when sexual orientation is held constant greater negative attitudes are evoked by effeminacy, the study does not account for the negative stereotype of effeminate gay males as the antithesis of what it means to be masculine in the traditional sense. In the present study however, vignettes featured effeminate and masculine targets that were both gay and straight to better determine if negative attitudes toward effeminacy also exist toward straight males. Indeed, results suggested that males who endorsed traditional masculinity and believed that masculine attributes were important to their sense of masculinity exhibited hostility and discomfort toward effeminate males, regardless of their peer's sexual orientation.
Thirdly, in the present study, fear toward effeminacy was not predicted by any of the independent variables. While Glick et al. (2007) found that participants expressed greater attitudes of fear toward an effeminate gay male target than toward a masculine gay male target after experiencing a threat to their masculinity, my study did not include a threat component and may explain in part why fear was not predicted by any of the variables. It is important to note however that in the present study, participants who highly endorsed traditional masculinity and masculine attribute importance did express feelings of hostility and discomfort toward an effeminate male peer in the absence of a threat condition. This might suggest that attitudes of fear are more reactionary in nature and are likely to be elicited only when an explicit masculinity threat is present, as opposed to feelings of hostility and discomfort, which may be elicited by more indirect means.
Lastly, the finding that hostility and discomfort may be elicited toward an effeminate male in the absence of a threat condition is a departure from the findings in Glick et al. (2007). It is important to note that Glick et al. (2007) did not examine how male participants might differ in their negative attitudes toward their male target descriptions based on the variables examined in this study. Therefore, it may be inaccurate to conclude that a threat to one's masculinity must be present in order to elicit negative attitudes toward an effeminate male. In the present study, when these characteristics were examined, negative feelings were evoked even in the absence of an explicit threat to the participants' masculinity. Perhaps this suggests that males with these characteristics harbor moderate feelings of hostility and discomfort toward effeminate males in general and/or that for these males, simply reading a vignette about an effeminate male posed a cogent enough threat able to evoke these feelings.
In their series of studies, Vandello and colleagues (2008) provided support for the idea that every day men have to prove that they are in fact "manly" and live under a constant threat of having their masculinity challenged, the result of which is an overwhelming sense of vulnerability. This study provides support for the notion that males who highly endorse traditional masculinity or who feel strongly that traditional masculine attributes are important to their own sense of masculinity, feel a sense of hostility and/or discomfort toward another same-age peer who exhibits effeminate qualities. Even though effeminate "Michael" poses no direct threat to participants' masculinity, he is a reminder of what a man, according to American society's ideal, should not be. One can argue that by the time men enter high school, college, or the work force, they are conditioned to react negatively toward other men who do not abide by traditional masculinity's strict code of conduct.
In terms of findings relating to self-discrepancy, Higgins' (1987) research found support for the notion that ought/other self-discrepancies were correlated with feelings of fear, hostility, and threat. These "agitation-related emotions" (p. 323) are a product of the consequences that one faces when they do not live up to what they "ought" to be doing or how they "ought" to be behaving, according to an "other"-in this case American society. Unfortunately, the present study did not find support that masculine self-discrepancy along traditional masculine attributes predicts negative attitudes toward effeminacy in other males. This may have to do with the complexity of the Masculine Self-Discrepancy Scale itself and should be examined in future research.

Negative Attitudes Toward Homosexuality
One major finding not predicted by the hypotheses was the ability of endorsement of traditional masculinity and masculine attribute importance to predict negative attitudes toward homosexuality. Even though it was initially hypothesized that negative attitudes would be directed toward "Michael's" effeminacy rather than toward his homosexuality, results suggested that homosexuality continues to evoke negative attitudes even when the target male's gender expression is taken into consideration.
Results suggested that males who highly endorsed traditional masculinity exhibited general negative attitudes, hostility, and discomfort toward gay "Michael" descriptions, regardless of whether he was characterized as masculine or effeminate, with the exception of hostility, which will be discussed below. While attitudes of hostility and discomfort were expressed toward both effeminacy and homosexuality, the findings suggested that homosexuality predicted slightly better than the target's effeminacy for males who highly endorsed traditional masculinity. The same was also true in terms of discomfort for males who highly endorsed traditional masculinity, except that high masculine attribute importance only predicted discomfort toward effeminacy but not toward sexual orientation. In this case, masculine attribute importance appeared to be an important variable when considering negative attitudes toward effeminacy, regardless of that target's sexual orientation. In trying to understand why homosexuality evoked hostility and discomfort for certain males over and above effeminacy, consider that traditional masculinity explicitly endorses heterosexual relationships and categorizes sexual relationships with men as a "feminine desire" (Kimmel, 2003, p. 126). As such, males who agree with American society's masculine ideal may exhibit negative attitudes toward gay men solely because their homosexuality defies traditional statutes of masculinity, particularly the "no sissy stuff" clause.
Interestingly, for males who highly endorsed traditional masculinity, "Michael's" gender expression appeared to matter in terms of hostility. These participants expressed greater attitudes of hostility toward masculine gay "Michael" than participants who did not highly endorse traditional masculinity. However, for effeminate "Michael" there was no interaction between level of endorsement and sexual orientation; both groups expressed greater hostile attitudes toward gay "Michael" than straight "Michael," although the level of hostility was significantly greater for males expressing high levels of endorsement of traditional masculinity.
These findings are interesting for several reasons. Firstly, they suggest that despite the ostensibly protective nature of appearing traditionally masculine, a man's sexual orientation may still play a role in how other men react to them. Therefore, feelings of hostility may be evoked not because of how drastically different those males are to masculine gay "Michael," but because of how similar they are. In other words, the hostility does not stem from the idea that a masculine gay male is the antithesis of what it means to be a "real man" but instead that he challenges the notion that masculinity is an exclusive characteristic of male heterosexuality. Clearly, masculine gay "Michael" values many of the traditional ideas of masculinity that participants who highly endorse traditional masculinity do as well, except for the fact that he is engaged in a romantic relationship with another man. The similarity to a gay male who exudes traditional masculinity may therefore be threatening because of the possibility for masculine straight men to be perceived as gay. In other words, if being traditionally masculine is not an exclusive component of heterosexuality, what does this mean for traditionally masculine men who are straight? Kimmel and Mahler (2003) capture this fear when they argue that "homophobia is far less about the irrational fears of gay people, or the fears that one might actually be gay or have gay tendencies, and more the fears heterosexuals have that others might (mis)perceive them as gay" (p. 1446).
The other interesting finding was the hostile sentiment toward effeminate gay "Michael." Since no previous study had examined both gender expression and sexual orientation components together, no a priori predictions were made regarding negative attitudes toward specific gender expression/sexual orientation combinations (e.g., masculine/gay or masculine/straight). Results suggested that effeminate gay males evoked more hostile attitudes from men who highly endorsed traditional masculinity than effeminate straight males. During her time spent at a public high school, Pascoe (2007) observed and often commented on the hostility experienced by effeminate gay males. In particular she described the daily torment suffered by "Ricky," who "embodied the fag because of his homosexuality and his less normative gender identification and self-presentation" (p. 65) and who ultimately dropped out of school because of the abuse he faced on a daily basis that ultimately became "unlivable" (p. 71). This overly negative sentiment toward effeminate gay men also exists within the gay community suggesting that traditional masculinity not only conveys the image of an acceptable straight man but also of an acceptable gay man (Moritz, 2013). This acceptable gay man is a man who "is the opposite of faggy, the opposite of the femme gay man who gestures, speaks quickly in a high-pitched voice and says "darling" (Moritz, 2013, p. 1). Pascoe (2007) also observed this sentiment in gay male personal ads where gay men actively seek "straight-appearing, straight-acting men'" (p. 59).
In some ways this finding is curious because although an effeminate gay male represents a man who in almost every conceivable way violates traditional masculine characteristics and is stereotyped as such, an effeminate straight male challenges this stereotype and may be harder for individuals to understand or accept. Although there are examples of ways in which heterosexual men have found acceptable ways to engage in feminine behaviors or have a more feminine appearance without facing harsh criticism: ergo, the "metrosexual." Pascoe (2007) describes how heterosexual men have adopted the term "metrosexual" to describe themselves as straight men who value their appearance and grooming behaviors. She remarks that "because these sorts of grooming practices are associated with gay men, straight men developed a new moniker to differentiate themselves from other straight men and from gay men" (p. 63). Even though these men engage in feminine behaviors they still want to separate themselves from the stereotype of the effeminate gay male.
Relation to previous findings. Previous research has found support for the relationship between traditional masculinity and negative attitudes toward homosexuality, specifically male homosexuality. Herek (1988) for example, used the Attitudes Toward Lesbians (ATL) and Attitudes Toward Gay Men (ATG) to assess hostility of heterosexuals (both male and female) toward homosexuals (both male and female). Herek (1988) found that heterosexual males expressed greater hostile attitudes in general than heterosexual females and that they had greater hostility toward gay males than toward lesbians. To help explain the results, Herek (1988) examined the correlation between attitudes of psychological defensiveness to attitudes of hostility toward gay males. He found that participants with negative attitudes tended to perceive themselves to be quite different from gay men and that their negative attitudes were associated with traditional views of sex roles and conservative religious ideology. The present study expands upon these results by suggesting that traditional masculinity is associated with hostility toward masculine gay males and discomfort toward gay males regardless of their gender expression.
The present study also expands on Glick et al.'s (2007) findings of hostility toward an effeminate gay male target by suggesting that a) an explicit threat to one's masculinity does not have to be present in order for negative attitudes (such as hostility) to be exhibited toward other males and b) hostility toward effeminate gay males is not equally likely to occur for all men.

Negative Attitudes Toward Traditional Masculinity
Lastly, this study found negative attitudes were also evoked by a male peer's expressions of traditional masculinity. Specifically, this study found that participants who did not highly endorse traditional masculine ideology or characteristics expressed general negative attitudes and discomfort toward masculine "Michael" regardless of his sexual orientation. This was not the case however for feelings of hostility.
Interestingly, males who did not highly endorse traditional masculinity expressed less hostile attitudes toward masculine gay males than masculine straight males. Males who did not highly endorse traditional masculinity tended to express greater hostile attitudes toward masculine straight "Michael" than masculine gay "Michael," but showed the opposite pattern toward effeminate "Michael." With the effeminate targets, high and low levels of endorsement predicted greater hostile attitudes toward effeminate gay males than effeminate straight males although those who highly endorsed traditional masculinity expressed far greater hostile attitudes.

Relation to previous research findings. While a plethora of research
indicates that men who are often seen as less masculine may experience feelings of inadequacy (Pleck, 1981) and other negative emotions associated with an ought/other discrepancy (Higgins, 1987), no research has examined how these feelings translate into negative attitudes toward traditional masculinity. Astute observations, particularly regarding school violence, suggest that there may be a link between men who do not "measure up" to traditional masculine standards and negative attitudes toward men who embody traditional masculine characteristics. For example, in discussing major commonalities of school shooters between 1982 to 2001 in the United States, Kimmel and Mahler (2003) noted that almost all of them were "gaybaited," not because they were actually homosexual but because "they were different from the other boys-shy, bookish, honor students, artistic, musical, theatrical, nonathletic, "geekish," or "weird" (p. 1445).
Of course this is not to suggest that all men who are bullied or harassed in this way have thoughts of executing acts of extreme violence toward their aggressors, but this observation is important as it emphasizes the complex relationship between the different types of men as defined by Connell's (1995) masculine hierarchy. It is not simply that negative attitudes are directed downward from men characterized as hegemonically or complicitally masculine to men who exist within the subordinate or marginalized categories, but also in the opposite direction as well. And likely, even more complex than upward or downward types of negativity there is also within group negativity too (e.g., negativity within the gay male community toward effeminate gay men).

Limitations
One major limitation of the present study concerns the lack of diversity within the sample. Racially and ethnically, the participants were predominantly Caucasian limiting the generalizability of the present results. Interestingly in Kimmel and Mahler's (2003) commentary on school shootings, they remark that Caucasian males are more likely to perpetrate such extreme acts of violence in relation to gay-baiting than males of other racial and ethnic backgrounds. Therefore, it is of utmost importance to understand why there are such drastic differences in how males of different racial and ethnic backgrounds respond to this type of harassment from other males, and how they may be more or less vulnerable to its effects.
Also concerning the generalizability of the results is the fact that the present study focused only on college males. Not only does this limit the generalizability in terms of age but also of the fact that the sample did not include males between the ages of 18 and 27 who did not attend college. Also, it is important to note that the sample was taken from a university in the Northeast and therefore many not be generalizable to a college population of males in another region of the United States.
Another concern is the lack of diversity in terms of sexual orientation. Only 6.3% of the sample identified as either homosexual or bisexual. Therefore, given the other limitations of the sample population, the results are not generalizable to sexual minority males. As evidenced in Moritz's article (2013), negative attitudes toward effeminacy also exist within the gay male community, underscoring the importance of ensuring that the perspectives of homosexual males are included in future analyses to better understand the intersections between traditional masculinity and male homosexuality. It is important however, to consider these results in the context of other sample characteristics, specifically the mean age of the sample which was 19 years old. Research by Floyd and Stein (2002) supports the notion that there are multiple developmental trajectories of sexual orientation identity formation for gay, lesbian, and bisexual youths (ages [16][17][18][19][20][21][22][23][24][25][26][27]. As such, there are some LGBT youth for whom their sexual orientation identity does not develop until late. It is possible then that the percentage of homosexual or bisexual participants is different than it would be with a predominantly older sample. Another considerable limitation of the present study is the relative absence of a consideration of social class, both as a participant demographic variable to examine alongside race, ethnicity, age, and sexual orientation, but also in the way this variable might have been represented in each vignette. Social class is an important cultural variable in the construction of masculinity (Morgan, 2005) and likely influenced the way in which participants a) thought about their own masculine identity and b) manner in which they responded to the vignettes. The intersection of social class with other cultural variables on the construction of masculinity should be considered in future research.
In thinking about possible explanations for why attitudes of fear were not evoked by reading the effeminate target descriptions, it is possible that the vignettes were not vivid enough in describing the four types of "Michael" descriptions.
Although the vignettes were modeled after those used in Glick et al.'s (2007) study, they were rather brief containing only seven short sentences. Perhaps, a more detailed description of "Michael" would have impacted the results, as the participants would have been provided with a more substantial description with which to react to.
In terms of methodological limitations, two issues concern the validity of the Masculine Self-Discrepancy Scale that was used to measure participants' sense of selfdiscrepancy along traditional masculine attributes. In the present study, alterations were made to the original Ought Self Questionnaire scale (Theodore & Basow, 2000) including the incorporation of items "aggressive" and "dominant" as part of the present study's masculine scale. The instructions were also altered to prompt participants to think specifically about their masculine identity in contrast with that proscribed by mainstream American society, in terms of a percentage of how alike or not alike they thought the two were. This alteration was made after initial pilot testing was done with four male undergraduates to improve the readability of scale. Although reliability in the present study was high (α = .83), some participants indicated in their answer that they were unsure of what they were being asked to do. In asking participants to think about the "kind of man you think you are versus what kind of man 'society' thinks you should be," participants may have been confused for a number of reasons. For example, it may be incorrect to assume that individuals are fully aware of societal expectations of them and that these societal expectations are imbued with cultural influences that this questionnaire may not have fully accounted for. Pascoe (2007) described this issue using the following example: "both a rich, slim, soft-spoken businessman and a poor, muscular, violent gang member might be described as hegemonically masculine" (p. 8). In other words, a man's cultural identity is likely to impact his internal definition of what society's idea of traditional masculinity is. This limitation also reflects one of the disadvantages of conducting online survey research in which the researcher is unable to clarify participant questions. However, this disadvantage is offset by the advantage of anonymity inherent in online questionnaires that studies have found reduces measurement bias and are less susceptible to social desirability response bias (e.g., Holbrook & Krosnick, 2010;Joinson, 1999).

Future Research
The present study did not specifically measure participants' attitudes toward homosexuality. As a result, it cannot be determined from the present study if participants' negative attitudes toward the gay male targets are reflective of anti-gay sentiment or of negativity towards effeminacy which is a common stereotype associated with gay men (Kimmel, 2003) or a combination of both. Future analyses should seek to correlate participants' scores on the MRNI-SF "Negativity toward Sexual Minorities" factor with their attitudes toward the target descriptions.
Although this study's purpose was to begin to understand some of the complexities in the connection between traditional masculinity and the relationships between men, there are many additional nuances that this study does not address.
Given the limitations inherent in a sample that lacks diversity, future research should strive to include male participants of different cultural backgrounds including race, ethnicity, class, and sexual orientation. Because of the complexities of masculinity's intersection with culture (Doss & Hopkins, 1998;Pascoe, 2007), it is imperative that future research continues to try and understand masculinity within the context of cultural diversity.
This study provided some evidence to support the notion that men who do not highly endorse traditional masculinity or feel that traditional masculine attributes are important to their sense of masculinity are less likely to express negative attitudes toward effeminate or gay males than those who feel stronger about these ideas.
However, this study does not provide an understanding into the mechanisms underlying why some men resist traditional masculine ideology and do not incorporate traditional masculine attributes into their own masculine identity. At the present, we are left to speculate why some men are not phased by notions of traditional masculinity. If adherence to these norms is toxic for so many men, we need to understand more about these men. Any attempts to challenge and/or re-define masculinity would benefit from taking these distinctions into account.

Male Target Descriptions Masculine Gay Male:
My name is Michael. I am a 24-year old male student and I am studying economics at college. When I graduate, I hope to pursue a career in accounting where I have opportunities for advancement. My ultimate career goal is to become a CEO. I think I would excel in this career as my boyfriend and friends tell me that I am a leader and have no problem making decisions about what needs to get done. While at school, I have become involved in my fraternity and intramural basketball. In my spare time I like to hang out with my boyfriend and play Xbox.

Effeminate Gay Male:
My name is Michael. I am a 24-year old male student and I am studying musical theater at college. When I graduate, I hope to pursue a career in theater where I have opportunities to try my hand in a variety of theater genres. My ultimate career goal is to become a lead actor on Broadway. I think I would excel in this career as my boyfriend and friends tell me that I am creative and have a way of expressing emotion through song and dance. While at school, I have become involved in a dance group and costuming club. In my spare time I like to hang out with my boyfriend and go shopping.

Masculine Straight Male:
My name is Michael. I am a 24-year old male student and I am studying economics at college. When I graduate, I hope to pursue a career in accounting where I have opportunities for advancement. My ultimate career goal is to become a CEO. I think I would excel in this career as my girlfriend and friends tell me that I am a leader and have no problem making decisions about what needs to get done. While at school, I have become involved in my fraternity and intramural basketball. In my spare time I like to hang out with my girlfriend and play Xbox.

Effeminate Straight Male:
My name is Michael. I am a 24-year old male student and I am studying musical theater at college. When I graduate, I hope to pursue a career in theater where I have opportunities to try my hand in a variety of theater genres. My ultimate career goal is to become a lead actor on Broadway. I think I would excel in this career as my girlfriend and friends tell me that I am creative and have a way of expressing emotion through song and dance. While at school, I have become involved in a dance group and costuming club. In my spare time I like to hang out with my girlfriend and go shopping.
Appendix D Masculine Self-Discrepancy Scale Consider the following: Mainstream American society often reflects the idea that men should act in certain ways and possess certain characteristics in order to be considered masculine, such as to be physically and emotionally strong or to act aggressively. Keeping this in mind, the items below inquire about what kind of man you think you are versus what kind of man 'society' thinks you should be. YOURSELF = Yourself as YOU see through your own eyes SOCIETY = Yourself as SOCIETY thinks you -as a man -should be Please look at the display below in considering the following example: Example: "Athletic." If you think you are not athletic at all but you believe that mainstream American society thinks you, as a man, should be athletic, you might choose the 0% option indicating that these two circles (or perspectives) are very different.
On the contrary, if you think you are very athletic and you believe that mainstream American society thinks you, as a man, should be athletic, you might choose the 100% option indicating that these two circles (or perspectives) are completely alike.
To the extent that you feel that you are somewhat athletic in comparison to how athletic mainstream American society thinks you, as a man, should be, you might choose one of the options in the middle of the scale indicating that at times you feel very athletic, but there may be other times when you do not. Now, for the following 21 items follow the same procedure described in the example. Rate yourself relative to how society expects you, as a male, should be.