College Students' Perceptions of Professor Bullying

Although bullying research has burgeoned over the past two decades, only recently have studies begun to explore bullying of students by teachers. Preliminary findings suggest that teacher bullying and the maltreatment of students may result in loss of trust, feelings of hopelessness and depression, oppositional behavior and increased fighting amongst peers (Pottinger & Stair, 2009). To date, only one study (Chapell et al., 2004) has addressed the prevalence of teacher (professor/instructor) bullying in college student populations. Given the impact professor/instructor relations can have on college student outcomes (Wilson et al., 2010) and the severe consequences teacher bullying can have on primary and secondary students, it is important to identify whether college students report bullying by their professors/instructors. The present study examined the self-reported prevalence of instructor bullying among college students. Results revealed that 51% of students endorsed seeing another student being bullied by a professor/instructor at least once and 18% endorsed being bullied by a professor/instructor at least once. The findings also revealed a relationship between teacher bullying and professor/instructor bullying. Additional characteristics of student victims of teacher and professor/instructor bullying were explored; however, no significant differences were demonstrated between male and female students or between students with and without disabilities in their self-reported ratings of being bullied by teachers and professors/instructors. Finally, the psychometrics of a newly formed questionnaire addressing student perceptions of professor/instructor and teacher bullying were explored and established. Implications for universities and colleges are discussed and suggestions for future research are advanced.


List of Tables
Definition. Perhaps the greatest difficulty in the empirical study of bullying involves the definition of bullying, which has varied across studies (Bauman & Del Rio, 2006;Espelage & Swearer, 2003). Dan Olweus, who coined the term "mobbing" to first describe bullying in 1972 (Espelage & Swearer, 2003, p. 365), more recently defined bullying or victimization as the following: "A student is being bullied or victimized when he or she is exposed, repeatedly and over time, to negative actions on the part of one or more other persons" (Olweus, 1993, p. 9).
Upon reviewing the definitions of bullying over the previous ten years, Espelage and Swearer (2003) concluded that bullying involves repeated physical and verbal aggression that is enacted by an individual or group to attain a goal. Other definitions have followed similar formats. Twemlow, Sacco & Williams (1996) provided a definition that they explained is similar to Olweus's: The exposure of an individual, over and over again, to negative interactions on the part of one or more dominant persons, who gain in some way from the discomfort of the victims. These negative actions are intentional inflictions of injury or discomfort and may involve physical contact, words, or insulting gestures. Essential to the phenomenon of bullying is that there is an imbalance of strength, an asymmetrical coercive power relationship, and that the victims have problems defending themselves. Thus, fighting between two persons of similar strength and skill would not be defined as bullying. Forms of bullying 4 may be quite direct, that is, physical conflict, and others more indirect, such as ostracism, teasing, and other forms of social isolation (p. 2). Bauman & Del Rio (2006) offered another definition where bullying was defined as "a subset of more general aggression, distinguished by an intent to harm, the repetitive nature of the acts, and the power imbalance between bully and target" (p. 219).
Finally, Beran (2006) described bullying as "repetitive negative behaviors against another child who is unable to defend himself or herself" (p. 241).
Although a universally accepted definition of bullying is lacking in the literature (Espelage & Swearer, 2003), there are widely agreed upon aspects within most definitions (Nansel et al., 2001). For example, Nansel et al. (2001) highlighted three common aspects of bullying: Bullying is a specific type of aggression in which (1) the behavior is intended to harm or disturb, (2) the behavior occurs repeatedly over time, and (3) there is an imbalance of power, with a more powerful person or group attacking a less powerful one. This asymmetry of power may be physical or psychological, and the aggressive behavior may be verbal (e.g., name-calling, threats), physical (e.g., hitting), or psychological (e.g., rumors, shunning/exclusion) (p. 2094).
Both Olweus (2011) and Beran (2006) emphasized similar commonalities to Nansel et al. (2001), but Olweus expanded on the first aspect as unwanted and negative behavior that is aggressive and Beran described it as a "different affect between the aggressor and the targeted child" (p. 242).
Types of Bullying. Bullying can be direct/overt, involving in-person physical or verbal confrontations, or it can be indirect/covert, including rumor spreading, indirect name-calling (Espelage & Swearer, 2003;Olweus, 1993) or relational bullying, which involves the disruption of social relationships between victims and their peers (Berger, 2006). Bullying has also been subdivided into different types of aggression throughout the literature. Raine and colleagues (2006) described two forms of aggression, proactive and reactive, and Espelage & Swearer (2003) concluded that bullying fits into the former form of aggression. Proactive aggression involves aggression to attain a goal; bullying as a type of proactive aggression is unprovoked and used to attain a social goal (Beran, 2006). Olweus (2011) further delineated that bullying may be expressed in one of nine forms: verbal, social exclusion or isolation, physical, lying and rumor spreading, stealing or damaging, threatening, racially related, sexually related, or cyber incidents.
Four consistent categories of bullying identified within the literature include physical, verbal, cyber, and relational bullying. Physical bullying, a type of overt bullying, may be the easiest type of bullying to identify (Bauman & Del Rio, 2006). It typically involves hitting, kicking, or beating victims (Smith, 2011). Behavioral bullying, which is related to physical bullying, might include behaviors such as 6 stealing a peer's lunch, ruining homework, or holding one's nose while interacting with a peer (Berger, 2006).
Verbal bullying, especially name-calling (Whitney & Smith, 1993) occurs more commonly than other types of bullying and is even more prevalent in children of older ages (Berger, 2006;Wang, Iannotti & Nansel, 2009). Verbal bullying often involves teasing, taunting or threatening the victim (Smith, 2011). A similar construct identified in the literature involves verbal abuse, which includes ridiculing and teasing, name-calling, or yelling (Brendgen, Wanner & Vitaro 2006). Schaefer (1997) conducted a survey with mental health professionals and found that rejection, verbal put-downs, perfectionism, negative prediction, scapegoating, shaming, cursing or swearing, threats and guilt trips were generally agreed upon as parental verbalizations classified as verbal abuse.
A type of bullying that has become a prominent issue more recently is cyberbullying, sometimes referred to as electronic bullying (Smith & Slonje, 2010).
Cyberbullying has been defined as "the use of any electronic means to harm another person" (Trolley & Hanel, 2010, p. 33) and may involve the electronic spreading of inappropriate photographs of a victim or the online harassment on a webpage or social networking site (Berger, 2006;Swearer, Espelage & Napolitano, 2009). What is distinct about the operational definition of cyber bullying compared to other forms of bullying is that it can meet the criteria for repetitive occurrence through multiple viewings of one webpage or email (Smith & Slonje, 2010).
Relational bullying, which is a type of covert bullying, involves the disruption of social relationships between victims and their peers (Berger, 2006) and includes 7 ignoring or excluding children or spreading humiliating rumors about a victim (Bauman & Del Rio, 2006;Smith, 2011). Crick (1996) defined relational bullying to also include planned manipulation or sabotage of a peer relationship to obtain desired goals.
Roles. The roles involved within bullying are not necessarily exclusive to the traditional dyad of bully and victim (Salmivalli et al., 1996), but instead may fall on an intrinsically dynamic continuum (Espelage & Swearer, 2003). Bullies have been categorized as sadistic, depressed, or agitated (Twemlow et al., 1996) and described as aggressive towards peers and sometimes adults (Olweus, 1997).
Victims have been classified by Olweus (1997) as either passive/submissive or provocative/aggressive victims. The more common type of victims include passive victims, who are characterized as more anxious and insecure than average students and may be cautious, sensitive and quiet (Olweus, 1997). Research also indicates that passive victims may demonstrate lower levels of social skills and nonassertive behaviors compared to children not classified as any type of bully or victim by their peers (Toblin, Schwartz, Gorman, & Abou-ezzeddine, 2005). A far less common type of victim, the provocative victim, displays anxious and aggressive reactions and is characterized by irritating behaviors, quick tempers and hyperactive behaviors (Olweus, 1997). Provocative victims have been shown to be more likely to carry weapons, use alcohol, and become involved in physical fights than non-victims or victims who have nonaggressive attitudes (Brockenbrough, Cornell, & Loper, 2002) and to display impairments in self-regulation, such as higher levels of impulsivity, emotional dysregulation, and hyperactivity (Toblin et al., 2005). Additionally, 8 provocative victims have been shown to be more likely to use physical forms of bullying above verbal forms of bullying and more likely than passive victims to be victims of physical bullying (Unnever, 2005).
Preliminary evidence also suggests that children with physical disabilities (Dawkins, 1996), learning disabilities (Luciano & Savage, 2007;Martlew & Hodson, 1991;Norwich & Kelly, 2004), and attention problems (Brendgen et al., 2006;Brendgen, Wanner, Vitaro, Bukowski & Tremblay 2007;Unnever & Cornell, 2003) are more prone to being bullied. Specifically, children with physical disabilities have been shown to be twice as likely to be bullied compared to children without physical disabilities (Dawkins, 1996) and children with mild learning difficulties have been shown to be more prone to bullying (Luciano & Savage, 2007;Martlew & Hodson, 1991), teasing (Martlew & Hodson, 1991), and having fewer friends than students in mainstream classrooms (Martlew & Hodson, 1991). In one study, approximately half of students with learning difficulties who endorsed high rates of bullying reported the bullying was related to their learning difficulties (Norwich & Kelly, 2004). On the other hand, Kaukiainen and colleagues (2002) found that students with learning disabilities were more likely to be bullies, but not victims. Unnever and Cornell (2003) reported that students with ADHD were more likely to report bullying because of their difficulties with self-control, and more likely to report being bullied independent of any difficulties with self-control. Instead, Unnever and Cornell suggested that students with ADHD may be more vulnerable to being bullied because of their social status and tendency toward aggressive behavior with their peers.
9 Salmivalli and colleagues (1996) reported that bullying involves all school children and suggested that interventions should therefore address all students, not just bullies and victims. Specifically, children not directly involved in the bullying incident may play an important role in the acceptance of bullying (Salmivalli et al, 1996). Although Whitney and Smith (1993) reported that children may find it difficult to understand why others bully, the researchers also reported that only half of junior/middle school students and only one-third of high school students in their study reported trying to help a student victim of bullying and 20% of all students reported doing "nothing" (Whitney & Smith, 1993, p. 17). When bystanders do nothing to stop an incident of bullying, an acceptance of bullying behaviors may be conveyed (Espelage & Swearer, 2003;Salmivalli et al, 1996).
School and teacher roles in peer bullying have also been well documented. The attitudes adopted by the school concerning bullying may impact student bullying (Baker, 1998) and if adults in the school accept bullying behaviors -these behaviors may be more accepted by the students (Espelage & Swearer, 2003;Salmivalli et al., 1996). Teachers may play a part in bullying if they do not encourage appropriate interactions or neglect to stop a bullying behavior (Espelage & Swearer, 2003). For example, Stockdale, Hangaduambo, Duys, Larson and Sarvela (2002) reported that teachers and parents endorsed lower rates of bullying than students reported. Teachers of older students may not always identify bullies correctly and their ratings of bullies and victims do not well match student ratings (Leff, Kupersmidt, Patterson & Power, 1999). Multiple studies have found that teachers may misperceive the degree of seriousness of bullying (Craig, Henderson & Murphy, 2000;Bauman & Del Rio, 2006). Yoon & Kerber (2003) reported that teachers tended to have less empathy and become less involved when victims experienced social exclusion compared to physical or verbal bullying. Specifically, teachers were asked how they would respond to certain situations depicting different types of bullying. For social exclusion, only 10% of teacher responses involved interventions, but for physical and verbal bullying, 50% of teacher responses included interventions (Yoon & Kerber, 2003). Bauman & Del Rio (2006) reported similar results within a sample of pre-service teachers.
Specifically, when pre-service teachers were presented with situations involving relational bullying they were less likely to intervene and demonstrated less empathy compared to situations involving physical or verbal bullying (Bauman & Del Rio, 2006). Ellis & Shute (2007) reported that while a teacher's moral orientation as measured by the Sanctioning Voice Index (SVI; Caputo, 2000) does affect the way teachers intervene and if they comply with bullying policies, seriousness of the incidence may have a greater impact. Teachers may misperceive physical conflicts as more severe forms of bullying, whether meeting the criteria for bullying or not (Hazler, Miller, Carney & Green, 2001) and may underrate the number of students involved in bullying (Bradshaw, Sawyer, & O'Brennan, 2007). Bradshaw et al., (2007) reported that students were less likely than their teachers to believe their school had adequate preventative efforts in place and more likely to report teachers observing and not intervening on bullying. Similarly, Hazler, Hoover and Oliver (1991) reported that the majority of students (69%) felt their school professionals responded poorly towards incidents of bullying and some students felt this was due to ignorance or lack 11 of awareness.
Prevalence. Bullying within primary and secondary school systems unfortunately is a common problem (Espelage & Swearer, 2003) and may be the most pervasive forms of violence within the schools (Batsche, 2002). One of the first studies to assess bullying prevalence reported that a total of 15% of students in primary and secondary Norwegian schools were involved in bullying -9 % as victims and 7 % as bullies (Olweus, 1993). Reporting much larger figures, a retrospective study by Hoover, Oliver and Hazler (1992) conducted in the Midwestern and Southeastern parts of the United States with junior and high schools reported that 77% of students endorsed being bullied and 14% experienced severe reactions to the bullying. Nansel and colleagues (2001) assessed a large nationally representative sample of students in grades six through ten and reported that 29.9% of students endorsed being involved in bullying frequently; thirteen percent as a bully sometimes, 10.6% as a victim, and 6% as both. These findings were stable across schools located in towns, suburban areas, and urban areas. More recently, another nationally representative study conducted by Vaughn et al. (2010) reported that six percent of the population, based on interviews with over forty thousand individuals, reported lifelong bullying of others.
Preliminary research also indicates that student perception of teacher response to bullying may impact psychosocial functioning (Troop-Gordon & Quenette, 2010).
Troop-Gordon and Quenette (2010) found that student perceptions may moderate the relationship between harassment and internalizing symptoms and school avoidance.
Specifically, male students showed greater internalizing difficulties as bully victims only when they perceived that their teachers were encouraging independent coping strategies and avoidance or assertive behavior in response to aggressive peers. Female students, on the other hand, demonstrated a relationship between emotional dysfunction and victimization only when they saw their teachers encouraging 13 independent coping strategies and avoidance or assertive behavior in response to aggressive peers at low levels or rarely (Troop-Gordon and Quenette, 2010).

Measurement of Bullying
Although bullying was first operationalized in the 1970s, psychometric problems often characterize bullying assessment measures (Beran, 2006). Solberg & Olweus (2003) (Beran, 2006). Finally, the divergences in the scope of the questions across studies may result in divergent prevalence rates (Stockdale et al., 2002).
Selecting the appropriate method of assessment may help address these psychometric problems. Espelage & Swearer (2003) identify four ways of assessing bullying: self-report, peer and teacher nominations, and behavioral observations.
Teachers, parents, students, peers, and researchers are all possible sources for data collection (Espelage & Swearer, 2003) and bullying can be assessed with a global scale or it can comprise specific, behavior related questions gauging frequency (Stockdale et al., 2002). Solberg & Olweus (2003)  estimates that are more abstract than estimates generated from a single item.
Self-reports. In general, self-reports are the most commonly type of psychological instruments used within psychological research (Constantine & Ponterotto, 2006). Bullying self-report measures typically ask students to report their perceptions about bullying behaviors and bully-victim experiences over a specific length of time (Espelage & Swearer, 2003). An important limitation of self-report measures is that they may inflate and/or underrepresent the prevalence of bullying. Salmivalli et al. (1996) reported that only 23.9% of the students identified by their peers to be victims of bullying self-identified themselves as victims, suggesting selfreport measures for victims of bullying may underrepresent prevalence rates.
Similarly, bullies themselves may misreport their actions either as too high if they are proud of their actions or too low if they are ashamed (Berger, 2006). Finally, students may be more inclined to endorse being the victim of a verbal or physical bullying behavior than they are to endorse being bullied overall (Stockdale et al., 2002), suggesting students may be hesitant to self-endorse that they are victims of bullying.
These findings support the supplemental use of peer and/or teacher reports, as well as the use of questionnaires with multiple items, opposed to dichotomous self-report questionnaires alone.

15
The most frequently used self-report measure in the bullying literature is the Olweus Bully/Victim Questionnaire (BVQ, Olweus, 1989;Schafer, Werner & Crick, 2002), which can be administered anonymously or confidentially (Olweus, 2010). The BVQ measures two global constructs: bullying and victimization (Solberg & Olweus, 2003). Furthermore, the BVQ includes a detailed definition of bullying and specifies a reference period of the past couple of months to ensure its sensitivity to change (Olweus, 2010), with questions clearly identifying the context of the school (Solberg & Olweus, 2003). After an initial question for each construct is asked (e.g. "How often have you been bullied at the school in the past couple of months?"), seven to eight follow-up questions are asked to specify how often specific aspects of bullying occur (Solberg & Olweus, 2003). The questions address verbal, relational, and physical bullying (Solberg & Olweus, 2003). Questions also address the behaviors and reactions of others witnessing bullying events (Olweus, 2010). Answer choices are represented on a Likert-scale from never to several times a week and the sum of the items is used as an overall measure of bullying level (Beran, 2006). Solberg & Olweus (2003) identified the cut-off point to be two to three times per month for both bullying and victimization, although other cut-off points have been reported (e.g., Dawkins, 1996).
Studies assessing the validity and reliability of the BVQ have reported mixed results. Solberg and Olweus (2003) reported acceptable convergent and divergent validity as well as strong construct validity and psychometric properties. Solberg and Olweus also concluded that there was a linear relationship between internalizing problems and victimization and externalizing problems and bullying. Beran (2006), on the other hand, concluded that the BVQ demonstrated adequate convergent validity, but inadequate divergent validity because it did not distinguish between reactive aggression and bullying. Students and teachers also appeared to perceive bullying differently and therefore Beran (2006) suggested that questionnaires should include more items with a clearer definition of bullying.
Another self-report questionnaire that has been used to assess bullying is the  (Espelage & Swearer, 2003;Solberg, Olweus, 2003). Nomination measures will either require a minimum number of nominations or specify a certain standard deviation above the mean to classify a student as a bully or victim (Solberg & Olweus, 2003). Solberg and Olweus (2003) argued against using this method for prevalence rate estimations, because the cut-off points are often arbitrary and difficult to reproduce. Nomination measures have similar limitations to self-report measures; students' reports may be influenced by reputation rather than personal experience (Berger, 2006) and adults may not be reliable nominators as they may misperceive the severity of situations (Bauman & Del Rio, 2006;Yoon & Kerber, 2003).
An example of a nomination measure is Salmivalli and colleague's (1996) peer nomination questionnaire. The nomination instructions included a definition of bullying and students were instructed to identify how well each classmate fit bully behaviors on a scale from one to three. The questionnaire next instructed students to identify peers who might be victims of bullying; students were considered victims if at least 30% of the students identified them. A final component of this questionnaire assessed sociometric qualities; students identified peers they liked the most and liked the least (Salmivalli et al., 1996).

Teacher & Student Relations
Although literature directly addressing teacher bullying is sparse, the importance of teacher and student relations (Ladd, Birch, & Buhs, 1999;Liljeberg, Eklund, Fritz, & Klinteberg, 2011;Rueger, Malecki, & Demaray, 2010) and tension between students and teachers (Buxton & Brichard, 1973) have been well documented. For example, teachers have described students with attention difficulties (Batzle, Weyandt, Janusis, & DeVietti, 2009;Eisenberg & Schneider, 2007), antisocial tendencies (Ladd et al., 1999), and learning disabilities (Woodcock & Vialle, 2011) more negatively than students without these conditions and consequently 18 these students may be more at-risk for verbal abuse (Brendgen et al., 2006;Brendgen et al., 2007). A study conducted by Buxton & Brichard in 1973 reported that 81% of high school students believed their teachers were violating the rights of students by not respecting their opinion. These negative perceptions are especially concerning given difficult teacher-child relationships developed early on have been shown to have a negative connection with academic achievement (Hamre & Pianta, 2001;Ladd et al., 1999) and an even greater relationship with behavioral outcomes (Hamre & Pianta, 2001). Furthermore, research has found that negative teacher-student relationships are difficult to change by teachers later on (Howes, Phillipsen, Peisner-Feinberg, 2000;Jerome, Hamre, Pianta, 2008), possibly because of negative attributions and stereotypes developed by teachers based on early conflict (Ladd et al., 1999).

Teacher & Professor/Instructor Bullying
There is a body of literature suggesting that teacher bullying of students exists (Chapell et al., 2004;Olweus, 1996 as cited in Brendgen et al., 2006;Pottinger & Stair, 2009;Twemlow et al., 2006), although professor/instructor bullying of college students has been relatively unexplored. In addition to research addressing student perceptions of teacher bullying, the extant research includes teacher perceptions of teacher bullying (Twemlow et al., 2006) and student perceptions of psychological (Casarjian, 2000) and verbal abuse (Brendgen et al., 2006;Brendgen, et al., 2007) by teachers. Furthermore, there is a substantial body of research relating to physical and verbal maltreatment of students by school staff and students bullying teachers in other countries such as Israel (Benbenishty, Zeira, Astor, & Khoury-Kassabri, 2002;Khoury-Kassabri, 2006;Khoury-Kassabri, 2009;Khoury-Kassabri, 2011;Khoury-Kassabri et al., 2008;Terry, 1998 Therefore, the definition of professor/instructor bullying used in the present study draws on definitions of teacher bullying provided by previous researchers (Twemlow et al., 2006;Olweus, 1996as cited in Brendgen et al., 2006 and the broader definition of peer bullying (Olweus, 1993). Specifically, professor/instructor bullying of students is defined in the current study as the use of power to punish, manipulate or belittle a student beyond what would be a reasonable disciplinary procedure. For example, professor/instructor bullying may include saying hurtful things in general or specific to the student's character or ability, making obscene gestures to the student, ignoring or neglecting the student, physical actions or attacks that may involve hurting or pushing a student around, or telling lies or secrets that make others dislike the student or that get the student into trouble. For the present study, the definition of a professor includes academicians who are involved in teaching and research at a college or university and the definition of an instructor includes persons holding a teaching role at a college or university.
Roles. Although specific roles within professor/instructor bullying have not been explored, research investigating the roles within teacher bullying in primary and secondary education has been conducted. Twemlow and colleagues (2006) reported that teachers who endorsed being bullied by their students and teachers who reported being bullied as children were more likely to bully their own students. A study by Pottinger & Stair (2009) found that male teachers were more likely to bully students than female teachers were and that students reported embarrassment or humiliation and physical bullying as a worst experience.
Research investigating teacher verbal abuse of students, a closely related construct to teacher bullying, has shown that students with higher likelihoods of being verbally abused by teachers may comprise 15% of school children, and even as students change teachers from grade to grade, the probability of teacher verbal abuse remains relatively stable for these children (Brendgen et al., 2006). Boys appear to be more likely than girls to perceive themselves as victims of teacher maltreatment (Benbenishty et al., 2002;Casarjian, 2000;Khoury-Kassabri, 2009;Khoury-Kassabri et al., 2008) and female students from families with a high Socioeconomic Status (SES) are less frequently victims of teacher verbal abuse (Brendgen et al., 2007).
Additionally, students with prominent inattention and antisocial behaviors have been shown to be at more risk to teacher verbal abuse than students without learning disabilities as these behaviors may jeopardize a teacher's efficiency and systematic management of the classroom (Brendgen et al., 2006;Brendgen et al., 2007). Khoury-Kassabri (2011) most recently reported a relationship between students being bullied by other students and being maltreated by teachers and school staff. Finally, students who are both bullies and victims and students who are bullies alone, may be more likely to be maltreated by school teachers and staff than victims of bullying (Khoury-Kassabri, 2009).
Prevalence. Olweus (1996as cited in Brendgen et al., 2006 reported that 1.67% of school-aged children endorsed being bullied by teachers. Twemlow et al. (2006) reported that over 70% of teachers have recognized teacher bullying as a problem and 45% of teachers admitting to bullying students. In 1998, Terry reported that 37% of teachers reported that their students may have viewed their (teacher) behaviors as bullying at least once during the course of one term. In 2005, McEvoy investigated current and former high school students' perceptions of teacher bullying and reported that 93% of the respondents identified at least one teacher as a bully in their school. When asked about emotional and physical maltreatment by teachers and school staff, one-third of Israeli students endorsed being emotionally maltreated (Benbenishty et al., 2002;Khoury-Kassabri, 2009) and one-fifth endorsed being physically maltreated by teachers and school staff (Benbenishty et al., 2002;Khoury-Kassabri, 2006;Khoury-Kassabri, 2009). In the United States, Casarjian (2000) reported that nearly two-thirds of middle school students reported at least one occurrence of teacher psychological abuse during the school year. Similarly, Whitted 22 and Dupper (2008) investigated teacher bullying among adolescent students in alternative education programs in an urban school district in the United States and reported that the majority of students endorsed being physically (86%) and psychologically (88%) mistreated by an adult. In a study investigating teacher (professor/instructor) bullying of college students, approximately 30% of college students reported witnessing teacher (professor/instructor) bullying at least once, 12.8% reported witnessing it occasionally and 2% reported witnessing it very frequently (Chapell et al., 2004). Fifteen percent reported being bullied by college teachers (professors/instructors) once or twice, 4% reported being occasionally, and 2% reported being bullied frequently (Chapell et al., 2004).

Outcomes. Students who have been bullied by teachers have demonstrated
greater risk for oppositional behavior, increased fighting, loss of trust, feelings of hopelessness and suicidality, post-traumatic stress disorder (PTSD) and depression (Pottinger & Stair, 2009). Teacher bullying has also been shown to be related to oppositional defiant disorder for male students who endorse feeling threatened by teacher bullies and for female students who endorse being verbally humiliated by teacher bullies (Pottinger & Stair, 2009). Finally, Pottinger & Stair (2009) reported that the more frequently a student reports being bullied by a teacher the higher their perceived pathological symptoms may be. While the long-term negative effects of teacher bullying during childhood into early adulthood have been documented (Pottinger & Stair, 2009), to date, no studies have addressed the impact of professor/instructor bullying within the college student population.

Purpose of the Present Study
Although the negative impact of peer bullying and teacher maltreatment of students in primary and secondary education have been clearly documented, only one study has addressed the prevalence of teacher (professor/instructor) bullying in college populations (Chapell et al., 2004). Given the impact professor/instructor relations can have on college student outcomes and the severe consequences teacher bullying has on primary and secondary students, it is important to identify whether college students report bullying by their professors/instructors. Therefore, the primary purpose of the present study was to examine self-reported prevalence of instructor bullying among college students.
Additionally, the present study explored whether specific student characteristics were associated with professor bullying. For example, previous research suggests that male students compared to female students may be more susceptible to teacher bullying (Pottinger & Stair, 2009) and students with learning disabilities may be especially at risk to peer bullying in primary and secondary school (e.g., Brendgen et al., 2006). Consequently, the present study examined the effects of sex and disability status on college student victimization of instructor bullying.
Previous research has also indicated that school-aged children have reported being bullied by teachers (McEvoy, 2005;Olweus, 1996 as cited in Brendgen et al., 2006) and teachers have recognized teacher bullying as a problem (Twemlow et al., 2006). Furthermore, students more prone to teacher verbal abuse, a construct similar to teacher bullying, appear to be more likely to be verbally abused over time (Brendgen et al., 2006). Therefore, the present study also explored college students' perspectives of being bullied by teachers in elementary, middle, and high school.
Although a few studies have addressed similar constructs to teacher bullying (e.g., verbal abuse, psychological abuse, maltreatment), the extant literature specifically addressing teacher bullying has relied on measures assessing similar constructs (Pottinger & Stair, 2009;Whitted & Dupper, 2008) or has not made the questionnaires available (Olweus, 1996 as cited in Brendgen et al., 2006). In addition, the only published study addressing teacher (professor/instructor) bullying of college students relied on dichotomous yes/no questions (Chapell et al., 2004). Therefore, the final purpose of the present study was to examine the psychometric properties of a newly formed questionnaire, the Student Perception of Professor/Instructor Bullying Questionnaire -SPPBQ, designed to assess professor/instructor bullying.
In summary, the purposes of the present study were to: 1. Examine the relationship between college students' history of being bullied by teachers prior to college and current self-ratings of being bullied by professors/instructors in college. It was hypothesized that students who reported a history of being bullied by teachers prior to college were more likely to report being bullied by professors/instructors in college as measured by a global professor bullying score on the SPPBQ.
2. Explore the relationship between disability status and a) self-reported ratings of being bullied by teachers prior to college and b) self-reported ratings of being bullied by professors/instructors during college. It was hypothesized that students who reported having a current disability, including physical and learning 25 disabilities, were more likely to report a history of being bullied by teachers prior to college as measured by a question about frequency of teacher bullying on the SPPBQ and report being bullied by professors/instructors in college as measured by a global professor bullying score on the SPPBQ.
3. Investigate the relationship between sex and a) self-reported ratings of being bullied by teachers prior to college and b) self-reported ratings of being bullied by professors/instructors during college. It was hypothesized that male students were more likely to report being bullied by teachers prior to college as measured by a question about frequency of teacher bullying on the SPPBQ and report being bullied by professors/instructors in college as measured by a global professor bullying score on the SPPBQ.
4. Further explore sex as a potential moderator on the relationship between teacher bullying in high school and professor bullying in college. It was hypothesized that the relationship between student report of being bullied by teachers prior to college and student perceptions of being bullied by professors/instructors in college would be moderated by sex. 26

Pilot Study Procedure
Prior to use in the exploratory study, the properties of a newly formed questionnaire, the SPPBQ, were explored via a pilot study. The SPPBQ includes nine questions that inquire about experience of peer, teacher, and professor/instructor bullying and 17 questions that inquire about exposure to different types of bullying, including personal, academic and physically intimidating forms of bullying; this set of questions is repeated for each bullying incident reported by the participant. The purpose of the pilot study was to address any unanticipated problems with the questionnaire before beginning the main study as suggested by Ary, Jacobs, Razavieh & Sorensen (2006). Revisions included rewriting any poorly written or misleading questions to refine any abstract ideas and ensure a complete understanding of the intended content (Redding, Maddock & Rossi, 2006). A script to elicit feedback in the questionnaire from participants (See Appendix A) was used during this phase of the research.
For the pilot study, a convenience sample of four undergraduate students was recruited from the University of Rhode Island (URI). This number was based on a sample size used in a pilot study by Chen et al. (2002) that also modified questions of a new questionnaire. Participants contacted the student investigator who explained the purpose of the pilot study and reviewed the informed consent (see Appendix B).
Participants who provided consent then completed the SPBBQ and answered questions about their understanding of specific questions and their experience of completing the questionnaire based on a cognitive script to elicit feedback (see 27 Appendix A). Minor wording changes were made to the SPPBQ based on participant feedback and one question was added to address the number of teachers and/or professors/instructors by whom students endorsed being bullied.

Main Study Procedure
Participants for the proposed study included 337 college students recruited from general education courses, upper level college courses, emails through listservs, and flyers posted throughout campus at the URI. Courses that included students of all majors and years were targeted, but based on the majors endorsed by participants, it appears that psychology and communication courses in particular yielded the most participants. Information directed participants to a secure and encrypted screen hosted by the website for SurveyMonkey where the online survey was accessible. Once students accessed the site, they were instructed to read a consent form and confirm they understood the contents by clicking on a statement of endorsement. Participants who provided consent were then presented with electronic forms of the SPPBQ and a demographic questionnaire designed by the researcher. At the end of the survey, participants were provided with information about how to contact the researcher if desired.
Participants. A convenience sample of 337 participants was recruited and included college students from URI at different levels of education and with varying majors. A minimum sample size of 300 was chosen based on a power analysis (see Appendix C) and Nunnally's (1978) recommendation for a sample of at least 300 participants in assessments of internal reliability. Participants were expected to be representative of the sex and ethnicity demographics of the undergraduate population 28 at URI; the majority of participants were therefore expected to be white, and males and females were expected to be approximately even (Table 1 provides   The final sample consisted of 337 students, including 260 females and 65 males, of which 80.7% were white (n = 272), 5.6% were black or African American (n = 19), 6.8% were Latino/Hispanic (n = 23), 2.9% were Asian/Asian American (n = 10), and 3.5% self-identified as another ethnicity including American Indian or Alaskan (n = 3), Pacific Islander (n = 1), multiethnic (n = 5), and other (n = 3). The students ranged in age from 18 to 35, although the majority of students (90.8%) were between the ages of 18 and 22 and the mean age was 20. Eight percent of the participants were freshmen (n = 27), 18.7% were sophomores (n = 63), 37.4% were juniors (n = 126), and 32.1% were seniors (n = 108). Forty-seven percent of students had declared majors in the Arts & Sciences (n = 158), of which 53.8% were psychology majors (n = 85) and 24.7% were communications majors (n = 39); the remaining students represented majors across all colleges of the university (Business, n = 20; Education, n = 15, Engineering, n = 4; Human Science and Services, n = 64; Nursing, n = 17; Pharmacy, n = 6, Sciences, n = 30, and Undecided or Other, n = 12).
The mean GPA reported by participants was 3.129. Table 2 presents participants by year in college, ethnicity and sex. Twenty participants reported having a documented disability of which 7 students reported having a Learning Disability, 11 students reported having ADHD, 2 students reported having dyslexia, and 3 students reported having a mental disability (anxiety, Post-Traumatic Stress Disorder, and depression). One participant reported having a physical disability. The majority of students (approximately 80%) reported attending a public elementary school (n = 266) and a public high school (n = 270).
The remaining students endorsed attending private and religious schools.
Informed Consent. Students who accessed the website to participate in the present study were required to document that they had read and understood the consent form and were of at least eighteen years of age before beginning the surveys.
The consent form included a basic description of the project as well as any potential for harm, confidentiality, and benefits of participating. Participants were made aware that they could discontinue their involvement at any time by ending the survey. No identifying information was collected; however, participants were provided with the project director's contact information if they had any questions or concerns. Informed consent is presented in Appendix D and debriefing is provided in Appendix G.
Measures. Two measures were used in the present study. All participants completed a questionnaire assessing their experience of being bullied by teachers and professors/instructors and a questionnaire including demographic information.
To assess the prevalence of professor/instructor bullying among a sample of college students as well as the perception of teacher bullying retrospectively throughout primary and secondary education, a self-report questionnaire (SPPBQ) was developed. The SPPBQ includes a working definition of teacher and professor/instructor bullying followed by three questions inquiring about teacher and professor/instructor bullying experiences that were used by Chapell et al. (2004) to address student perceived prevalence of teacher (professor/instructor) bullying.
Follow-up questions then address when self-reported bullying incidents occurred and two questions address student intervention in situations of professor/instructor bullying. The remainder of the questionnaire follows a similar format to the NAQ-R (Einarsen et al., 2009) -a previously validated measure designed to assess exposure to workplace bullying. This questionnaire encompasses three underlying factors: personal, work-related (revised to academic-related) and physically intimidating forms of bullying and has been found to generate a single item measure of bullying (Einarsen et al., 2009). Questions address specific behaviors and answer choices are on a Likertscale for frequency, from never to daily. The SPPBQ ends with a definition of peer bullying followed by three questions that inquire about peer bullying experiences, which were used by Chapell et. al (2004). The SPPBQ is presented in Appendix F.
A demographic questionnaire included questions regarding student sex, age, years of education, GPA, ethnicity, disability status, and major. Although age, years of education, GPA, ethnicity and major were not variables included in the hypotheses, they were included in the demographic questionnaire for descriptive information, post hoc analyses, and potential covariates in future studies. The demographic questionnaire is presented in Appendix E

Design
The present research study a) investigated the psychometric properties of the SPPBQ, b) examined prevalence rates of bullying based on descriptive findings, c) explored the relationship between teacher bullying and professor/instructor bullying, and d) explored the characteristics of student victims of teacher and professor/instructor bullying. SPSS and EQS were used to conduct the data analyses.
To investigate the psychometric properties of the SPPBQ, internal consistency and dimensionality were assessed with an exploratory principal component analysis (PCA), followed by an item analysis and final PCA including 50% of the final sample.
Additionally, cross-validation was assessed with a split-half technique in which a Confirmatory Factor Analysis (CFA) including the remaining 50% of the final sample was conducted.
The assessment of the relationship between teacher bullying prior to college and professor/instructor bullying in college was conceptualized as a one-way between subjects design with one dichotomous independent variable (victimization status before college) and self-perceived professor/instructor bullying as measured by the SPPBQ's global bullying score as the dependent variable. A 2 x 2 factorial design was conceptualized for the assessment of the relationship between student characteristics and report of teacher and professor/bullying. Two dichotomous independent variables (disability status and sex) were included with two continuous dependent variables of self-perceived bullying by teachers prior to college and self-perceived professor/instructor bullying in college as measured by the global bullying scores from the SPPBQ. Finally, a 2 x 2 factorial design was conceptualized for the assessment of sex as a moderator of the relationship between student report of being bullied by teachers prior to college and student perceptions of bullying by professors/instructors in college. Two dichotomous independent variables (teacher bullying status and sex) were included with one continuous dependent variable of self-perceived bullying by professors/instructors in college as measured by the global bullying scores from the SPPBQ.

Chapter III: Results
Six different sets of analyses were conducted; the first two analyses related to the exploration of the psychometric properties of the SPPBQ, the third analysis was conducted to provide descriptive statistics of prevalence rates, and the final three analyses related to the four hypotheses of the study. Specifically, the analyses included: a) an analysis of the internal consistency and dimensionality of the SPPBQ for the professor bullying section and the teacher bullying section of the SPPBQ, b) a cross-validation of the SPPBQ for the professor bullying section and the teacher bullying section of the SPPBQ, c) prevalence analyses including descriptive data of students' reports of professor and instructor bullying, d) an analysis of the relationship between teacher bullying before college and professor/instructor bullying during college, e) a group comparison analysis between sex and bullying as well as disability status and bullying, and f) post hoc analyses to address some of the inconsistencies within the preliminary findings, exploring the role of sex as a potential moderator in the relationship between teacher bullying and professor/instructor bullying.

Psychometrics -Item Analysis & Dimensionality
SPSS version 21 was used to conduct the item analyses and PCAs on 50% of the sample completing all questions on the SPPBQ (n = 153). Items involving the professor/instructor section of the SPPBQ were assessed first and items involving the teacher section of the SPPBQ were conducted second. For both sets of analyses, Horn's parallel analysis and Velicer's MAP were run as outlined by O'Connor (2000) to assess the number of components to be specified within the PCA. Items that were complex (loading on more than one component with coefficients greater than .40), did not load onto any dimensions with coefficients greater than .40, and loaded on components that did not make conceptual sense in the initial PCA were removed.
Next, an item analysis involving a comparison of item and total-item correlations was conducted, where items that correlated with the total-item correlation less than .40 were removed. A final PCA with an orthogonal (varimax) rotation was run on the remaining items and yielded the final version of the questionnaire. Internal consistency was assessed with Cronbach's alpha and Pearson's bivariate correlation was then conducted to assess the criterion validity of the professor and teacher sections of the SPPBQ.
Professor/Instructor Bullying -SPPBQ. Horn's parallel analysis and Velicer's MAP identified three components within the professor/instructor bullying section of the SPPBQ. Factor loadings from the initial PCA are presented in Table 3. Seven items were removed at this stage. Items 3, and 14 were removed because they did not load strongly on any component and items 4, 8, 9, 11, and 18 were removed because they were complex.  The results of the PCA, shown in Table 6 Teacher Bullying -SPPBQ. Horn's parallel analysis and Velicer's MAP specified two components within the teacher bullying section of the SPPBQ. Factor loadings from the initial PCA are presented in Table 7. Item 21 ("having insulting or offensive remarks made about your private life by a professor/instructor"), which was included in the professor/instructor section of the SPPBQ, was omitted from the teacher section due to a clerical error. Item 21 was one of the first items eliminated in the professor/instructor section (it loaded with two other similar items on the same component and was considered redundant); therefore, this omission is considered only a minor limitation. Four items were removed at this stage. Items 3, 11, 14, and 17 were removed because they were complex.  41 elimination of the second component as Noar (2003) recommends retaining at least four items per construct. Therefore, items 8, 10, and 20 were retained. During this stage items that were redundant were also removed; specifically items 7 and 15 were removed because they represented similar constructs to other items within the same factor.  The results of the PCA, shown in

Psychometrics -Cross-Validation
A CFA was conducted in EQS as a cross-validation procedure, which allows for more confidence in a measure's psychometric structure (Redding et al., 2006).
Specifically, a split-half cross-validation technique as recommended by Redding and colleagues (2006) was used, where data gathered from the second half of the participants completing all items on the SPPBQ (n = 151) was used. CFAs were conducted separately for the professor/instructor bullying section and the teacher bullying section. For each section, four models were tested: 1. A null model suggests correlated. For each model, the first factor loading was fixed at one in order to allow the factor variance to be freely estimated. Based on the results of the PCA, the best fitting model was hypothesized to be the correlated model encompassing two factors.
The best fit for the models was based on previous theory and research and parsimony as recommended by Noar (2003) and measures of good fit for the models were established by a chi-square to degrees of freedom ratio of no more than two to one, a Comparative Fit Index (CFI) of at least .90 (Noar, 2003) or .95 (Hu & Bentler, 1999).
Additionally, models with Root Mean Square Error of Approximation (RMSEA) less than .05 also indicated a good fit and lower Type II error rates (Hu & Bentler, 1999).
Professor/Instructor Bullying -SPPBQ. items 6 and 10, which involves items from two different factors and therefore does not make conceptual sense. The one-factor model and the uncorrelated two-factor model, however, yielded many standardized residuals above .20, further strengthening the conclusion that the best-fit model is the correlated two-factor model.  Table 12 displays the factor loadings and error variances for the correlated model, which demonstrates how well the measured variables represent each of the two factors (Harlow & Newcomb, 1990). The factor loadings are all significant (p < .001) and the error variances are reasonably low. Teacher Bullying -SPPBQ. Table 13 displays a summary of the measures of fit for each model within the teacher bullying section of the SPPBQ that are similar to the results of the professor/instructor bullying section. Although none of the models fit the data according to the standards described earlier, the correlated two-factor model appears to fit the data the best. For all of the models, the chi-square to degrees of freedom ratios are higher than two to one, the CFIs are lower than 0.90 and the RMSEAs are above .05. The difference in chi-square values from the correlated model and the uncorrelated model is significant [χ 2 (1) = 13.77, p < 0.001, ΔCFI = .008], Bullying component, which would result in complex loadings. The one-factor model and the uncorrelated two-factor model, however, yielded even more standardized residuals above .20, further strengthening the conclusion that the best-fit model is the correlated two-factor model.   Table 14 displays the factor loadings and error variances for the correlated model, which demonstrates how well the measured variables represent each of the two factors (Harlow & Newcomb, 1990). The factor loadings are all significant (p < .001) and the error variances are reasonably low.

Prevalence
The primary purpose of the present study was to explore the prevalence estimates of professor/instructor bullying on a college campus. The following section provides prevalence estimates of bullying based on descriptive findings of the study.
Prevalence estimates of how often students have witnessed and experienced professor/instructor, teacher, and student bullying were assessed and are displayed in  Table 15. Approximately half of the participants (51%) endorsed seeing another student being bullied by a professor/instructor at least once, but only 18% endorsed being bullied by a professor/instructor themselves at least once. Nearly half of the participants (44%), however, reported being bullied by a teacher in elementary, middle or high school at least once. Very few students reported a time when another student stopped or attempted to stop a professor/instructor from bullying them (7%) or a time when they stopped or attempted to stop another student from being bullied by a professor/instructor (14%). Additionally, prevalence rates of peer bullying and peer attempts at preventing bullying were analyzed. Although the majority of students reported witnessing peer bullying in college at least once (64%), only 33% endorsed being bullied by a peer in college and only 15% endorsed bullying their peers in college.

*since college
If participants endorsed being bullied by a professor/instructor or a teacher, they were also asked to identify the number of professors/instructors and teachers who bullied them. As can be see by Table 16, for professor bullying (n = 57) answers 53 ranged from 1 (n = 31) to 6 (n = 1) professors/instructors, with a median and mode of 1 professor/instructor. For teacher bullying before college, students (n = 161) reported being bullied by between 1 (n = 67) and 7 (n = 1) teachers, with a median of 2 teachers and a mode of 1 teacher. Participants were also asked the grade in school or year in college in which they were bullied by a teacher or professor/instructor; these results are shown in Table   17. More than half of the participants (51%, n = 158) reported at least one grade or year in which they were bullied. Because students were instructed to report as many grades in which they recalled being bullied, a total of 335 endorsements of grades and years were given by the 158 students. Students endorsed between 1 (n = 68) and 9 (n = 1) different grades and years in which they reported being bullied by teachers or professors/instructors; the median number of grades/years students endorsed was 2 and the mode was 1.  Table 18 displays the grades and years in which students reported being bullied by teachers and professors/instructors. The median grade specific to teacher bullying was 8 th grade and the mode was 10 th grade and the median and mode year in college specific to professor/instructor bullying was sophomore year of college (note that the second year of college was the most common even when including a sample of first year students who had not yet experienced their second year). Finally, prevalence rates of teacher or professor/instructor bullying of students by sex and disability status are presented in Table 19 and Figures 3 and 4 respectively.
While 47% of female participants endorsed being bullied by teachers prior to college at least once, only 34% of male participants endorsed being bullied by teachers; these differences yielded a small effect size (d = .272). Similarly, 21% of female participants endorsed being bullied by a college professor/instructor at least once and only 9% of male participants endorsed being bullied by a college professor/instructor, yielding a small effect size (d = .331).
Although limited by the small sample of students with disabilities (n = 20), the descriptive differences between students with and without disabilities are notable.
Seventy-five percent of students with a documented disability, compared to 42% of students without disabilities, reported being bullied by a teacher prior to college and fifty percent of students with disabilities, compared to 16% of students without disabilities reported being bullied by a college professor/instructor in college. The 56 differences between students with and without disabilities and their endorsement of teacher and professor/instructor bullying yielded large effect sizes (d = .676 and d = .737 respectively).

. Professor/Instructor & Teacher Bullying by Disability Status Teacher & Professor/Instructor Bullying Relationship
The hypothesis that students who report a history of being bullied by teachers prior to college were more likely to report being bullied by professors/instructors in college was tested via a one-way between subjects ANOVA with one dichotomous independent variable (victimization status before college) and self-perceived professor/instructor bullying as measured by the SPPBQ's global bullying score as the dependent variable. Additionally, Cohen's d was calculated to measure effect size.
Participants were instructed to identify the most recent time they were bullied by a teacher prior to college.  Assumptions of the F test, used to assess an ANOVA, include independence, normality, homogeneity of variance, and consistent data to the underlying structural model (Myers, Well & Lorch, 2010). The assumption of independence refers to randomly assigning each participant to a group (Myers et al., 2010). A second assumption, involving normality (including skewness and kurtotis), is commonly violated by researchers within between groups designs (Myers et al., 2010). Although there is ample evidence that breaking this assumption has a relatively small impact on the validity of conclusions (Cohen, 1983) and in particular, Myers and colleagues (2010) explain non-normal data has little impact on the Type I error rate with moderately large sample sizes, breaking the normality assumption may reduce the statistical power of the study (Myers et al., 2010). For the present analysis, however, post hoc power was relatively strong (1-β = .727) even though the assumption of normality was broken. Homogeneity of variance is the assumption that the variability within the data is related to the intervention or independent variable (Myers et al., 2010); this assumption is broken when the unequal variance is related to another a difference in variability between groups, or when the data are subject to floor or 59 ceiling effects (Myers et al., 2010). The structural model assumption reflects the assumption that the only manipulated factor influencing the data is the factor of interest and that the residual variability reflects random error (Myers et al., 2010).
Due to the purpose of the present study, it was impossible -and unethical -to assign participants to groups of being bullied by teachers or not being bullied by teachers, and this correlational design reflects a violation of the assumption of independence. Means, standard deviations, skewness and kurtosis of overall global ratings of the SPPBQ -Professor/Instructor, which is a composite score on students' perceptions of being bullied by professors/instructors in college, are shown by teacher bullying status prior to college in Table 21. The table demonstrates that the data within the present study also breaks the violation of normality due to leptokurtic variation, but no transformations were made because of the low potential impact on Type I error rate and the importance of maintaining the nature of the relationships.
Another limitation of this analysis is related to the structural model assumption and involves the potential influence of the imbalance of males and females within the two groups of students endorsing being bullied by a teacher prior to college and students reporting having never been bullied by a teacher prior to college, which were reported in test that deals with heterogeneity of variance and works best when the data is not 60 highly skewed and has group sizes larger than 10 (Myers et al., 2010), was run in addition to the standard F test to account for this violation.

Sex, Disability Status & Bullying
The hypothesis that a) students who report having a current disability were more likely to report having been bullied by teachers prior to college and to report being bullied by professors/instructors in college and b) students that are male were more likely to report having been bullied by teachers prior to college and to report being bullied by professors/instructors in college, was tested with a 2x2 (disability status; sex) between subjects factorial MANOVA. Two dichotomous independent variables (disability status and sex) were included with two continuous dependent variables of self-perceived bullying by teachers prior to college and self-perceived professor/instructor bullying in college as measured by the global bullying scores from the SPPBQ.
Assumptions of between subjects MANOVAs are similar to those of ANOVAs and therefore the violations in this analysis are similar to those presented in the previous section. The assumption of independence was not met due to the nature of the independent variables -sex and disability status, which cannot be randomly assigned. An additional problem within this analysis involved unequal groups, especially for comparisons between students with and without disability status. Myers and colleagues (2010) suggest an ANOVA for unequal group sizes in the case of onefactor between-subjects designs.
Means and standard deviations of overall global professor and teacher ratings of the SPPBQ, which include the global score on students' perceptions of being bullied by teachers prior to college and professors/instructors in college, are shown by sex and disability status prior in Table 23. The table demonstrates that the data within the present study also breaks the violation of normality due to leptokurtic variation, but no transformations were made because of the low potential impact on Type I error rate and importance of maintaining the integrity of the relationships between the independent and dependent variables. Finally, homogeneity of variance was assessed via Levene's test, which was not significant for perception of teacher bullying prior to college [F(1, 282) = 0.819, p = 0.484], but was significant for perception of professor bullying in college at the .05 level [F(1, 282) = 6.032, p = 0.001] indicating this violation was broken for professor bullying only. Therefore, Welch's F tests were run in addition to the standard F test to account for this violation. conducted to further explore the relationship between these two variables within perception of professor/instructor bullying; no significant results were revealed.
Statistical power for these analyses were computed post hoc; power was low for the ANOVAs involving both disability status (1-β = 0.24) and sex (1-β = 0.35).

Post Hoc Analysis -Sex as a Moderator
Post hoc analyses were conducted to further explore sex as a potential moderator on the relationship between teacher bullying in high school and professor bullying in college. Moderators are variables that impact the strength and/or direction of the relationship of an independent and dependent variable (Baron & Kenny, 1986).
Sex was chosen as a moderator for two reasons. First, Table 27 demonstrates that even though the relationship between teacher bullying and professor/instructor bullying was significant and 66% of the students who reported being bullied by a professor in college also reported being bullied by a teacher prior to college, the low percent of students (28%) who endorsed being bullied by teachers prior to college who also endorsed being bullied in college merited further exploration. Second, given the contradictory finding that male student reports of being bullied as a single item question were lower than female student reports in college and male perceptions of being bullied by professors as assessed by the SPPBQ were higher than females in college, a deeper understanding of sex in relation to bullying is important (See Table   19). Therefore, the hypothesis that the relationship between student report of being bullied by teachers prior to college and student perceptions of being bullied by professors/instructors in college was moderated by sex was tested via was tested with a 2x2 (teacher bullying status; sex) between subjects factorial ANOVA. Two dichotomous independent variables (teacher bullying status and sex) were included with one continuous dependent variable of self-perceived bullying by professors/instructors in college as measured by the global bullying scores from the SPPBQ.
Assumptions of between subjects ANOVA, and the violations of these assumptions, are similar to those mentioned previously. These assumptions include the assumption of independence, the issue of unequal groups, normality, and homogeneity if variance. Means and standard deviations are shown in given the relation of interest is the interaction, no modifications were made. . There was also a significant interaction, shown in Figure 6, between teacher bullying status and sex with a small effect size [F(1, 303) = 7.37, p = 0.007, η 2 =.018,]. Therefore, a simple effects test, shown in Table 30, was conducted to further explore the relationship between these two variables within perception of professor/instructor bullying. Although there were no significant results for female students, a significant effect was demonstrated for male students with a moderate effect size [F (1, 303) = 1.667, p < 0.001, d = 0.612, 95% CI (.262, .711)]. Therefore, for male students, endorsement of being bullied by teachers prior to college led to higher ratings of perceptions of professor/instructor bullying in college; however, for female students, teacher bullying status prior to college demonstrated no effect.

Chapter IV: Discussion
Historically, bullying research has focused on student to student bullying and only recently have studies begun to explore teacher bullying of students. Preliminary findings suggest that teacher bullying and the maltreatment of students may result in loss of trust, feelings of hopelessness and depression, oppositional behavior and increased fighting amongst peers (Pottinger & Stair, 2009). Although much less is known about professor/instructor bullying of college students, research addressing the relationship of professors and students has demonstrated that college students' perception of their rapport with professors may predict motivation, perceptions of learning, and perceived grade (Wilson et al., 2010). After examining the psychometric properties of a questionnaire designed to assess college students' perspectives of professor/instructor bullying and teacher bullying, the present study identified the selfreported prevalence of professor and instructor bullying among college students and college students' perspective of being bullied by teachers in elementary, middle, and high school. Additionally, the present study explored whether specific characteristics were associated with professor bullying -including history of being bullied by teachers, sex and disability status.

Psychometric Findings of SPPBQ
To date, no research has examined the types of professor/instructor and teacher bullying that may exist. A large body of research, however, suggests that peer bullying may be subdivided into direct/overt bullying, involving in-person physical or verbal confrontations, and indirect/covert bullying, including rumor spreading and indirect name-calling (Espelage & Swearer, 2003;Olweus, 1993). Bullying has also been categorized as either verbal, physical, relational or cyber. Furthermore, no measures assessing college students' perceptions of professor/instructor bullying exist and studies addressing teacher bullying of students have either relied on measures involving similar constructs or have not made their questionnaires available. In addition to examining the psychometric properties of a newly formed questionnaire assessing professor/instructor bullying and retrospective teacher bullying, the present study examined the types of bullying professors/instructors and teachers may use.
The results from the psychometric analyses revealed that both sections of the questionnaire demonstrated satisfactory Cronbach's alpha ratings (greater than .70 as suggested by Nunnally, 1978) and revealed adequate loadings and inter-item correlations for each component. Additionally, the findings provided support for good criterion validation for overall teacher and professor/instructor bullying and its two subcomponents -Academic Bullying and Physical Bullying. Results from the CFA offered evidence for cross-validation of the components within the SPPBQ. Overall, the analyses offer a strong psychometric foundation for the SPPBQ. Additional research should explore the SPPBQ's test-retest reliability, as well as its reliability across different samples. Furthermore, the convergent and divergent validity of the SPPBQ should also be explored.
Because the questionnaires were developed from an existing workplace questionnaire (NAQ-R; Einarsen et al., 2009) that encompassed three underlying components of bullying -personal, work-related, and physically intimidating forms of bullying -it was hypothesized that similar components would be revealed for the professor/instructor bullying and retrospective teacher bullying questionnaire within the SPPBQ. Additional items were added, however, to address other aspects of bullying that might not exist within the workplace (e.g. acts of violence and being excluded). Interestingly, only two components were revealed for these new measures: Academic Bullying and Physical Bullying. Academic Bullying encompassed academic forms of bullying that occur within the classroom and are related to course performance or participation. The second component, labeled, Physical Bullying encompassed physical and sexual bullying only. Examples of items that loaded on the Academic Bullying component included, "A professor/instructor/teacher withholding information that affects your performance," "Being humiliated or ridiculed by a professor/instructor in connection with your course," and "Repeated reminders of your mistakes by a professor/instructor/teacher." Academic Bullying included items specific to professors/instructors and teachers that were congruent to verbal and relational bullying as described in the peer bullying literature. The items, however, related specifically to verbal bullying within the classroom setting or relational bullying that would impact a student's academics. Although previous research has addressed bullying as it relates to personal and work characteristics (e.g., Einarsen et al., 2009) and a large number of studies have explored the academic outcomes of victims of bullying (e.g., Holt, Finkelhor, & Kantor, 2007;Swearer, Espelage, Vaillancourt, & Hymel, 2010), no studies to date have explored bullying behaviors involving academics specifically.
Physical Bullying included items with more severe consequences that included both physical bullying and sexual bullying. Items loading on the Physical Bullying component included, "Threats of violence or physical abuse by a professor/instructor/teacher," "Acts of violent or physical abuse by a professor/instructor/teacher," and "having a professor/instructor/teacher gossip about your sex life or spread rumors about your sexual activities." The Physical Bullying component included items resembling physical bullying in the peer bullying literature, but also included sexual harassment. In the professor/instructor section of the SPPBQ only, the Physical Bullying component also includes making false allegations.
Although these items are not consistent with specific types of peer bullying or workplace bullying, they appear to represent bullying in the college context that is more severe and does not necessarily occur in the classroom setting. Future research should explore the generalizability of Academic and Physical Bullying in other college and primary and secondary school settings. Additionally, research should explore the criterion validity of each component separately by including additional questions related to the frequency of academically related bullying as well as the frequency of bullying that is physical or involves more serious threats.
Although item loadings for the SPPBQ did not remain consistent to the categories developed by Einarsen and colleagues (2009) within workplace bullying, this is not especially concerning because of the differences implicit within workplace and academic settings. Generalizing research from non-academic settings to college contexts has proved to be problematic in other areas as well. For example, Myers, Edwards, Wahl and Martin (2007) reported that attributes related to argumentative individuals in contexts relating to superior and subordinate relationships may not translate to the college context. In the present study, items that loaded on Academic Bullying on the SPPBQ included items from the NAQ-R in all three categories.
However, most of the items that fell into the Academic Bullying component on the SPPBQ came specifically from the work-related (academic-related) and person-related categories represented in the NAQ-R. The discrepancy between the two questionnaires may relate to the difference in relationships between supervisors and supervisees and professors and students. Specifically, professors have a substantially shorter length of time to get to know their students (e.g., classes do not meet every day and only last for one semester) and as a result may know less about their students personally. Therefore, on the SPPBQ items that previously related to personal bullying within workplace bullying may fit better within academic forms of bullying.
Although the Academic and Physical Bullying components appear to be unique to professor/instructor and teacher bullying, they were relatively consistent between the professor/instructor bullying section and retrospective teacher bullying section. Items that differed within the questionnaires may reflect the differences between the college setting and the primary and secondary school settings. It is possible that the nature of the relationship between teachers and students in primary and secondary school and those of professors/instructors and college students vary in duration. Future research should explore the differences between professor/instructor bullying in college and teacher bullying in high school.

Prevalence Rates of Bullying in College
The review of the literature explored previously suggested a dearth of research in the area of professor/instructor bullying of college students. It is interesting to note that the only study to date that has addressed professor/instructor bullying of college students (Chapell et al., 2004) did so based on preliminary interviews with students who told stories of teacher bullying and the original intent of the study was actually to assess peer bullying. Therefore the results of the present study, only the second to address this issue, add substantially to the literature.
Previous prevalence rates reported by the only study addressing professor/instructor bullying reported that 44% of college students endorsed witnessing a teacher (professor/instructor) bully other students at least once and 19% of college students endorsed being bullied by a professor/instructor themselves at least once (Chapell et al., 2004). The present study's findings are remarkably consistent with the rates reported by Chapell and colleagues in 2004. Specifically, 51% of the participants endorsed witnessing a professor/instructor bully other students at least once and 18% of the participants endorsed being bullied by a professor/instructor themselves at least once.
The estimates of peer bullying within college students are also similar to Chapell et al.'s findings. Chapell et al. (2004) reported that 60% of college students reported seeing peer bullying at least once and in the present study, 64% of the participants endorsed seeing a student be bullied in college at least once. Furthermore, the previous study reported that 18% of students admitted to bullying another college student and the present study reported that 15% admitted to doing so. The findings from the present study, however, demonstrate a much larger percentage of students endorsing being bullied themselves by another college student compared to Chapell et al.'s (2004) findings (33% compared to 15%), which may be due to differences in campuses or the timing of the questionnaires. It is important to note, however, that the demographics, including year in college, sex, age, and ethnicity did not appear highly different in each study.
The prevalence rates relating to students stopping or attempting to stop professor/bullying from occurring are unique to the present study; however, previous research has made estimates of middle school students attempting to stop peer bullying. Whitney and Smith (1993) reported that approximately one-third of high school students reported trying to help a student being bullied and 20% of all students reported doing nothing. Although the rates reported in the present study are even lower -only 7% endorsed having another student intervene in professor/instructor bullying and only 14% endorsed intervening in professor/instructor bullying themselves -these rates do suggest that some students may perceive themselves and others as student allies. Student allies may be crucial in the implementation of support groups and/or clubs to encourage anti-bullying climates on college campuses, especially considering that when bystanders do nothing in response to bullying a message of acceptance may be conveyed (Espelage & Swearer, 2003;Salmivalli et al, 1996). Future research needs to assess effective interventions and prevention programs that support student allies and help reduce professor/instructor and teacher bullying.
The prevalence rates associated with experiences of teacher bullying prior to college were alarmingly high. Forty-four percent of the college students sampled endorsed being bullied at some point by a teacher prior to college, but only 12% endorsed this to have occurred more than once or twice. Previous prevalence rates throughout primary and secondary school of teacher bullying and emotional maltreatment were much lower, ranging from 1.7% (Olweus, 1996as cited in Brendgen et al., 2006 to 33% (Benbenishty et al., 2002;Khoury-Kassabri, 2009). It is probable that the discrepancy in the present study's finding compared to other findings relates to this study's retrospective and inclusive nature. Previous studies have assessed prevalence rates of teacher bullying while children are still in elementary, middle and high school and have utilized measures assessing bullying within specific time periods. The present study, however, asked college students about their teacher bullying experiences at anytime in elementary, middle and high school retrospectively and was therefore more likely to demonstrate higher frequencies of teacher bullying.
One other important item to note about students' perspective of bullying prior to college is that the median grade students endorsed being bullied by teachers was 8 th grade and the most commonly reported grade students reported teacher bullying was 10 th grade. Future studies exploring teacher bullying might focus their efforts on middle school and early high school given students endorsed this year as a time of being bullied by teachers at high rates.

Characteristics of Victims of Teacher & Professor/Instructor Bullying
To explore whether there are specific characteristics within victims of teacher and professor/instructor bullying, group analyses were conducted based on college students' teacher bullying status before college, sex, and disability status. The present study found support for the first hypothesis, that students with a history of being bullied by teachers were more likely to report being bullied by professors/instructors in college; however, there was no evidence to support the remaining two hypotheses, that students with disabilities and students who are male were more likely to report being bullied by teachers prior to college and professors/instructors in college. The present study did, however, find support for an additional hypothesis, that the relationship between student report of being bullied by teachers prior to college and student perceptions of being bullied by professors/instructors in college was moderated by sex, which was added based on the preliminary findings of this study.
Students in the present study who reported being bullied by teachers before college endorsed significantly higher ratings on the professor SPBBQ suggesting that there was a relationship between teacher bullying status before college and perceptions of professor bullying in college. This finding is consistent with previous research, which reported that forty-percent of victims of bullying in college were also victims in primary and secondary school (Chapell et al., 2004). Within the present study, 66% of those students reporting being bullied by a professor/instructor in college at least once, also reported being by bullied by a teacher prior to college. However, only 28% of those students who reported being bullied by a teacher prior to college at least once, also reported being bullied by a professor/instructor in college -demonstrating the substantial decrease in rates of professor/instructor bullying compared to teacher bullying and the possibility of specific moderators on this relationship. It is plausible that students who reported teacher bullying may have reported less professor/instructor bullying upon entering college due to the size of university classes. University classes, especially those that meet general education requirements and are taken the first few years of college, are often larger than primary and secondary school classes. University courses may therefore offer more anonymity to 79 students and fewer opportunities for professor bullying of students to occur.
Although victims of teacher bullying may remain relatively stable throughout their pre-college education and college experience, future research should explore the different characteristics associated with students who endorse teacher bullying and professor/instructor bullying compared to those students who report being bullied only in high school or only in college. Findings from studies exploring these differences may help identify students most at risk for professor/instructor bullying. Additionally, uncovering factors associated with students who appear to overcome teacher bullying may help inform the development of interventions for students who may remain prone to bullying from adolescence into early adulthood.
Although previous research suggests male students may be more likely than female students to perceive themselves as victims of teacher maltreatment (Benbenishty et al., 2002;Casarjian, 2000;Khoury-Kassabri, 2009;Khoury-Kassabri et al., 2008) and may be more susceptible to teacher bullying (Pottinger & Stair, 2009), the present study did not find a relationship between sex and perception of bullying in school or college as a main effect. This finding was somewhat surprising for the results involving students' perceptions of teacher bullying before college based on the research previously mentioned; however, a recent review of peer bullying suggested sex within the roles of bullies and victims are complex (Carrera, DePalma & Lameiras, 2011). Specifically, boys and girls may be more likely to fall into different categories based on type of bullying (e.g. physical or relational; Carrera et al., 2011). In the present study, the prevalence rates reported by males of teacher and professor/instructor bullying were actually lower than those reported by females (34% 80 of males compared to 47% females for teacher bullying prior to college) as assessed by a question asking how frequently the participant had been bullied. Males, however, reported approximately the same level of bullying experiences as females on the overall ratings of teacher bullying prior to college within the SPPBQ, which includes the average of all items related to Academic and Physical Bullying. It is possible that the differences in question construction led to varying results. For example, Stockdale and colleague (2002) reported that students may be more inclined to endorse specific components of bullying compared to endorsing being bullied in general. The retrospective nature of the questionnaire may have also led to different results than previous research, which asked students questions within a specified time period.
The present study's findings that there were no differences between male and female students' perceptions of professor/instructor bullying in college are less surprising. The only study exploring sex differences in teacher (professor/instructor) bullying within college did report that males engaged in bullying behaviors significantly more than female college students; however, the study did not find a significant relationship between sex and other bullying variables and ultimately called for more research in the area (Chapell et al., 2004). In the present study, males actually reported higher means (although not significantly higher) than females did on the SPPBQ for professor bullying, but endorsed a much lower frequency of being bullied by professors/instructors when asked how often they experienced professor/instructor bullying (9% of males compared to 21% females for professor/instructor bullying in college). Even though the SPPBQ included a definition of bullying in an attempt to establish a consistent understanding of bullying, 81 these conflicting ratings might reflect a difference in perceptions in the definition of bullying and reflect an issue with the construct validity of using a one item question to assess prevalence. Preconceived notions of bullying may have influenced participant responses to the single question asking how frequently they were bullied, but may have been less influential on the global bullying score, which was comprised of multiple questions addressing independent experiences. As mentioned previously, students may be more hesitant to report being bullied in general compared to endorsing specific components of bullying (Stockdale et al., 2002). The types of students that succeed in high school and enter college may also relate to why males and females did not differ significantly in their ratings of professor/instructor bullying.
It is possible that a higher rate of male students, who may be more prone to being bullied by teachers prior to college (Pottinger & Stair, 2009), are not completing high school or entering college.
Based on the inconsistent findings related to sex and bullying over time, post hoc analyses were conducted to explore whether sex acted as a moderator to the relationship between endorsement of being bullied by teachers prior to college and perceptions of being bullied by professors in college. The present study did find support for this hypothesis. Specifically, for male students, endorsement of being bullied by teachers prior to college led to higher ratings of perceptions of professor/instructor bullying in college; however, for female students teacher bullying status prior to college demonstrated no effect. Therefore, sex may moderate the relationship between teacher bullying status prior to college and perceptions of professor/instructor bullying in college.

82
Although evidence did not support the final hypothesis that students who report having a disability were more likely to experience teacher and professor/instructor bullying, limited statistical power precludes any accurate conclusions. Specifically, the sample of students reporting disabilities was very small -only 15 students reported having a disability and completed both sections of the questionnaires completely. The high prevalence rate of students with disabilities who reported being bullied by teachers and professors/instructors suggests there may be differences between students with disabilities and students without disabilities in their reporting of teacher and professor/instructor bullying. In the present study, 75% of the students who reported having a documented disability, compared to 42% of students without a disability, reported being bullied by a teacher prior to college and fifty percent of students with disabilities, compared to 16% of students without disabilities, reported being bullied by a college professor/instructor in college. The differences between students with and without disabilities in their report of being bullied by teachers and professors also generated large and consistent effect sizes. Based on the present study's descriptive findings, as well as previous studies' findings indicating students with certain disabilities may be more at risk to teacher verbal abuse than students without disabilities (Brendgen et al., 2006;Brendgen et al., 2007), further research focusing on students with and without disabilities is merited.

Implications
Unfortunately, bullying within school systems is a common problem in the USA (Espelage & Swearer, 2003) and in other countries (e.g., Olweus, 1993) leading to a burgeoning area of research over the past two decades. Comparatively, very little attention has been drawn to the issue of teacher bullying, which may have severe consequences for student victims. Professor/instructor bullying in particular has been largely overlooked and before the present study, only one study had addressed the issue (e.g., Chapell et al., 2004). Therefore current findings offer important insight and implications for college campuses, as well as professors/instructors and students.
The present study offers a new measure, the SPPBQ, to assess professor/instructor bullying as perceived by college students. The SPPBQ could be used in a variety of ways. First, it could be used as part of an overall campus climate survey, helping college administrators, faculty and staff understand students' perceptions of professor/instructor bullying. Second, it might also be used as a screening tool for university early alert systems and assistance with retention, helping to identify college students who feel they are being bullied by professors/instructors or who may be at risk to being bullied by professors/instructors based on their experiences before college. The finding that students who report being bullied by teachers prior to college were more likely to report professor/instructor bullying in college may also indicate a need to screen for victims of teacher bullying within incoming first year students. Identifying the students most at-risk to professor/instructor bullying and arming them with resources and tools to prevent future bullying -for example a student support group to stop professor/instructor bullying -might aid in the prevention of professor/instructor bullying.
The prevalence rates reported in the present study about professor/instructor bullying also speak to the need to reduce professor/instructor and teacher bullying.
Students clearly perceive this phenomenon to be occurring yet little attention has been 84 given to this subject. Support systems designed to help students report professor/instructor bullying and mitigate the consequences of professor/instructor bullying should be explored. Additional interventions should target professors and instructors, ideally to prevent bullying from occurring. Furthermore, the low percentage of college students reporting that others or they themselves have attempted to stop professors/instructors from bullying students may indicate a need to encourage students to support one another. An acceptance of bullying may be conveyed when bystanders do nothing to stop an incident of bullying (Espelage & Swearer, 2003;Salmivalli et al, 1996), therefore encouraging students to speak out during incidents of bullying -peer and instructor -might help reduce the prevalence rates of bullying.
Finally, it is clear that future research is warranted in the area of teacher and professor/instructor bullying of students with disabilities. The present study's small representation of students with disabilities precludes any conclusions about the likelihood of increased bullying within this population; however, the high frequency at which the students with disabilities reported being bullied by teachers and professors/instructors and the resulting large and consistent effect sizes may serve as a pilot study supporting the need for further exploration.

Limitations
Although the present study attempted to explore the role teachers may play in bullying, one limitation is that the findings may be perceived as a persecution or condemnation of teachers and professors. Nonetheless, it is important to address whether students perceive if they are being bullied by teachers and uncover what they perceive to be teacher bullying. The present study was limited in a number of other 85 ways as well -including a limited sample, a weak study design and the violation of multiple statistical assumptions.
Although the present study attempted to include college students representative of the university at large, the final sample was not a perfect match to the student demographics at the university. The final sample comprised more students with majors in the College of Arts and Sciences than any other college and included more females than males. Male and female participants may have been unevenly represented because there were more female students (60%) than male students (40%) in the College of Arts and Sciences in 2012. Furthermore, the most common major of the participants in the present study was Psychology, which also included more females (56%) than males (44%). Although students of color and students with disabilities were close to the university wide demographics, the findings reflect a sample that is predominantly white/Caucasian and able-bodied, limiting the generalizability of the findings. In particular, teacher and professor/instructor bullying of college students representative of multiple ethnicities, socioeconomic statuses, and sexual orientations are critical areas for future research. The present study was conducted at only one university in the northeast region of the United States and does not represent universities across the country.
The cross-sectional nature of the present study prevents conclusions about change over time and stability of the findings. In addition, the retrospective questions related to teacher bullying are limited by students' memory and self-report biases. As mentioned previously, one limitation related to self-report measures is the over and/or under representation of participant endorsements. Specifically, the present study was not able to objectively assess if professor/instructor bullying was occurring, but instead relied on student report of their perceptions of professor/instructor bullying.
Prevalence rates, therefore, reflect student perception of professor/instructor bullying, which may differ from actual incidents of professor/instructor bullying. These issues influence the construct validity of a measure -with self-report measures it may be difficult to be sure the measure is measuring the construct it is intended to measure (Ellsworth & Gonzalez, 2007). An additional limitation involves the reactivity and social desirability of participant responses, which suggests participant responses may be influenced by how the participant would like or feels like they should be conveyed (Anastasi & Urbina, 1997;Ellsworth & Gonzalez, 2007).
Finally, there was limited statistical power for most of the analyses conducted in the present study. Specifically, much of this data was nonnormal (the predominant issue being that it was leptokurtotik), which may have led to a loss of power in the group analyses. This issue relates to the item distributions within the teacher and professor sections of the SPPBQ. There was a low frequency of endorsement for all items on the SPPBQ, which resulted kurtotik data. To accurately represent the low frequency of bullying experiences within the college student population, the decision was made to maintain raw data and perform analyses without transformations to meet basic statistical assumptions. Although a limitation, maintaining the data in its true form may generate more practical results than results generated from a transformed dataset.
Another limitation that resulted in loss of statistical power in the group analyses involved the small and discrepant sample sizes. Participants were more 87 represented by female students (n = 272) than male students (n = 65), and only 20 students reporting having a disability participated. The discrepancy in these samples limited the power of the analyses in comparing males to females and students with disabilities to students without disabilities; however, effect sizes for both sex and disability status were consistent. Furthermore, the violation of the assumption of independence, for all three independent variables in the present study, limits the interpretations of the findings. Ellsworth and Gonzalez (2007) explain that variables that cannot be manipulated (e.g., sex and disability status) are defined as 'found' variables. To account for 'found' variables, and self-report problems, Ellsworth and Gonzalez recommend ruling out similar constructs that are correlated with the construct of interest. Although including similar variables would have strengthened the findings, it was beyond the scope of the present study.

Future Directions
The present study adds to the dearth of literature concerning professor/instructor bullying and teacher bullying of students in several ways. Firstly, the establishment of the psychometrics properties of the newly formed questionnaire that explores college students' perceptions of professor/instructor and teacher bullying will help aid future studies and screenings related to bullying within college campuses.
The prevalence rate estimates of professor/instructor bullying of college students reported in the present study are consistent with the previously reported rates by Chapell et al. (2004) and draw attention to college students' perceived existence professor/instructor bullying. Because 14% of college students reported stopping or attempting to stop professors/instructors from bullying other college students, future 88 research should explore the characteristics associated with these student advocates and uncover the ways in which students are already working to stop professor/instructor bullying. The alarmingly high estimates of prevalence rates related to teacher bullying in primary and secondary school reported retrospectively by college students highlight the need to intervene in teacher bullying early, ideally in elementary school.
Furthermore, the finding that students reported higher incidences of teacher bullying between middle and high school helps guide the timing for future interventions and research when examining students' perceptions of teacher bullying. Exploratory analyses of the continuity of students as victims of teacher bullying before college and professor/instructor bullying in college suggests students with a history of being bullied by teachers are indeed more likely to perceive they are being bullied by professors/instructors in college. Bullying prevention efforts, therefore, should also be implemented in adult settings, such as the workplace and academia, outside of elementary, middle, and high school.
The findings from the current study indicate a need for additional research addressing teacher and professor/bullying. Furthermore, given the divergence in the bullying literature, methodological issues need to be resolved and an operational definition of bullying needs to be established (Espelage & Swearer, 2003). The present study included a definition of teacher and professor/instructor bullying that was based on previous definitions in the literature (Twemlow et al., 2006;Olweus, 1996 as cited in Brendgen et al., 2006); however, the definition did not explicitly state that teacher and professor/instructor bullying must occur repeatedly over time.
Although omitting a time reference was consistent to previous definitions of teacher 89 bullying, it was not consistent to more general definitions of bullying between peers.
Future research should explore the differences between definitions of peer bullying and teacher and professor/instructor bullying. In addition to more research exploring the psychometrics of the SPPBQ, a warranted area of research would explore a similar questionnaire designed to assess primary and secondary students' perspectives of teacher bullying. The SPPBQ was designed to assess college students' perspectives in college and prior to college retrospectively, and there is a need to assess students' perspectives of teacher during elementary, middle and high school. Although there is a substantial body of research relating to physical and verbal maltreatment of students by school staff in Israel (Benbenishty et al., 2002;Khoury-Kassabri, 2006;Khoury-Kassabri, 2009;Khoury-Kassabri, 2011;Khoury-Kassabri et al., 2008), to date, very few studies have investigated teacher bullying of students specifically. Studies that have explored teacher bullying have either relied on measures assessing similar constructs (Pottinger & Stair, 2009;Whitted & Dupper, 2008) or on dichotomous yes/no questions (Chapell et al., 2004). Researchers have clearly demonstrated an interest in exploring teacher bullying indicating the need for a valid measure to assess student perception of teacher bullying.
Furthermore, an important area that was not addressed in the present study includes the consequences related to professor/instructor bullying. Although research involving teacher bullying has shown that student victims of teacher bullying may be at greater risk for negative peer social preference, delinquent behavior, poor academics, lower rates of high school graduation rates, increased behavior problems in early adulthood (Brendgen et al., 2006;Brendgen et al., 2007), oppositional behavior, increased fighting, loss of trust, feelings of hopelessness and suicidality, PTSD and depression (Pottinger & Stair, 2009), presently there is a lack of research examining the consequences of professor/instructor bullying. The assessment of consequences of professor/instructor bullying is a valuable avenue for research given the importance of professor and student relations in college students' academic success (Wilson et al., 2010).
Finally, although significant differences within disability status and teacher and professor/instructor bullying were not found in the present study, given the compromised statistical power of the design, further research in this area is needed.
The high prevalence rates reported by the small number of students with disabilities compared to students without disabilities and their large effect sizes for both teacher and professor/instructor bullying in the present study suggest it is possible that a representative sample of students with disabilities may report higher rates of teacher and professor bullying than students without disabilities. One reason teacher and professor/instructor bullying may be especially important to consider amongst students with disabilities pertains to the power differential that is implicit to the teacher-student relationship. A power differential is considered to be a necessary component in the definition of bullying and it might include size, style of dress, money, appearance, ethnicity or any other valued social quality (Twemlow & Sacco, 2008). Few populations have struggled with power in ways that persons with disabilities have.
Critical disability theory argues that political issues around disabilities first and foremost involve who is valued and who is marginalized in a society, which leads to a group with power and a group that is powerless (Devlin & Pothier, 2006). Therefore, 91 students with disabilities may be particularly vulnerable to teacher and professor/instructor bullying.

Summary & Conclusions
The present study offers a new tool, the SPPBQ, for the assessment of professor/instructor bullying of college students and teacher bullying of primary and secondary education students. The SPPBQ is comprised of two underlying components of bullying -Academic Bullying and Physical Bullying -as well as a global component encompassing all of the items. Overall, the questionnaire demonstrated strong criterion validity and internal consistency. The SPPBQ may be used concurrently, to assess college students' perceptions of professor/instructor bullying, and retrospectively, to assess college students' perspectives on teacher bullying.
Prevalence rates reported in the present study suggest bullying of college students by professors/instructors is indeed a problem. The rates were consistent to previous research (Chapell et al., 2004) and revealed that half of college students endorsed witnessing a professor/instructor bully another student at least once and onefifth endorsed being bullied by a professor/instructor at least once. Prevalence rates associated with teacher bullying prior to college demonstrate that teacher bullying of students may be a common problem and needs to be addressed. Nearly half of the participants endorsed being bullied at some point by a teacher prior to college, and 12% endorsed being bullied by a teacher prior to college more than once or twice.
Additionally, nearly 10% of students endorsed having a peer stop or attempt to stop a professor/instructor from bullying them, and 13% endorsed stopping or attempting to stop a professor/instructor from bullying a peer. These rates suggest that not only are students endorsing being bullied by their teachers and professors, but some students are also trying to intervene in order to stop the bullying from occurring.
A characteristic of victims of professor/instructor bullying included a history of being bullied by teachers in the past, which may be true for male students, but not necessarily for female students. Sex may therefore act as a moderator variable for the relationship between teacher bullying prior to college and professor/instructor bullying in college. Additionally, students with disabilities endorsed high prevalence rates of being bullied by both teachers and professors, indicating one characteristic that may be associated with victims of professor/instructor and teacher bullying is disability status.
The present study supports that college students clearly perceive teacher and professor/instructor bullying as occurring but may not know how to properly address this problem when it occurs. Findings revealed that college students endorsed alarmingly high rates of being bullied by professors in college and by teachers prior to college. The SPPBQ was developed to aid universities and researchers in the identification of students being bullied by their teachers and professors. The SPPBQ may also be used in future studies to address additional characteristics of victims of teacher and professor/instructor bullying or as a screening measure to assist in the understanding, prevention and intervention of professor/instructor bullying. In conclusion, the present study supports that professor/instructor bullying of students is an issue of critical importance. University administrators, faculty and staff should be made aware of professor/instructor bullying and future research should identify effective methods to address this problem and preventing it from occurring. may discuss your concerns with Marisa Marraccini at (434) 409-0698 or her major professor, Lisa Weyandt, Ph.D. at (401) 874-2087. In addition, you may contact the office of the Vice President of Research, 70 Lower College Road, Suite 2, University of Rhode Island, Kingston, RI 02882 (401-874-4328).
You have read this Consent Form and currently have no further questions concerning your participation in this project. You understand that you may ask any additional questions at any time and that your participation in this project is voluntary. By participating in the project, you agree that your answers can be used without your signed consent.

Sample Size -Group Comparison
Sample size for the group comparison analyses were determined by using the formula n = 2[(Z α + Z 1-β )/d] 2 and estimating a small to medium effect size of Cohen's d (d = 0.30;Cohen, 1988). Setting α = .05, β = .80, and d = .30, we find that n = 175 per group. information will not be shared with any organization. To ensure the confidentiality of participant data entered via the Internet, the data will be saved with unique nonidentifying user ID and passwords. Data collected online do not contain identifying information.
Decision to quit at any time: You may choose not to participate at any time and your decision will in no way affect your status with the University of Rhode Island.  (401) 874-2087 and they will discuss them with you.

Rights and Complaints
I have read the consent form and have no further questions about my participation in this project at this time. I understand that I may ask any additional questions at any time, that my participation in this project is voluntary, and that I may withdraw from this project at any time.
• I am at least 18 years old and I have read the consent form and agree to participate • I choose not to participate or I am not at least 18 years old