College Faculty as an Inconsequential
                                                                        Agent of Political Socialization

           David L. George                                                    Jerry F. Medler
           Department of Political Science                              University of Oregon
           Cal Poly State University                                        Eugene, Oregon
           San Luis Obispo, California 93407


This research examines some of the agents of political socialization which may influence university students political attitudes. A comparative analysis of two ideologically diverse student samples was  undertaken in public universities located in California and Oregon. The data from this study suggests that: (1) student perceptions of faculty attitudes at both universities were not related to their attitudes across the issues of class, race, gender and the environment p.>05.(2) student attitudes were related to family and friends across these four issues p.<05. Our research indicates that while some components of the higher education environment are associated with student political attitudes-­friends and community-­they appear to be independent of the students political ideology.

KEY WORDS: agents of political socialization; college faculty; college students; political ideology; political attitudes.


     This paper examines some political socialization agents which influence young adults. In particular, this study examines possible effects exerted by the family, peer groups, professors and the surrounding community on college students from two state universities. A comparative study of two state universities--California Polytechnic State University at San Luis Obispo and the University of Oregon at Eugene was completed in February of 1996. The student samples taken from these two universities were drawn from two upper division courses taught by the authors. Students were asked to participate in the study as a class exercise and findings, as available in the course of the term, were presented to students for comment and discussion. Political socialization is defined by the authors as an ongoing process in which the individual is repeatedly exposed to the acceptable political norms through a mixture of indoctrination and learning.
     Political science has a substantial research literature in political socialization developed primarily from the late 1950s through the 1980s which examines how children and adolescents are socialized into our political system (see, for example, Hyman, 1959; Greenstein, 1965; Davies, 1965; Newcomb, Koenig, Flacks and Warwick, 1967; Easton and Dennis, 1969; Dawson and Prewitt, 1977; and Jennings and Niemi, 1974, 1981). Since then, studies of political socialization have declined markedly. Ironically, Greenstein (1970) provided a stock market analogy when he suggested that the subfield of political socialization was experiencing a "bull market" during the late 1960s. However, as is the case with the stock market, one author in the mid-1980s asserted that the field of political socialization was then in the midst of a "bear market" (Cook, 1985). What happened to transform this once important and burgeoning subfield into a virtual non-entity? Cook asserted that this bear market in political socialization was directly attributable to a lack of coherent and unified theory. Cook's critique of the political socialization literature is well-taken: the apparent contradiction between studies cited above suggests that better theory and further research are necessary to resolve these differences. Another possible explanation for different findings is that the studies use different samples, research designs, operational definitions and are applied to different environments.
     One recent controversial theoretical perspective has been put forth by Judith Rich Harris (1995,1998) who argues that a genetic component needs to be incorporated in our socialization models and that contrary to popular belief the peer group exerts more influence on the child than the family. While this theoretical approach is beyond the scope of the present study, especially her argument that the genetic factor needs to be considered, it offers additional explanatory power to both the study of political socialization and political behavior in general.
     One of the earlier studies of political socialization was completed in the late 1930s by Theodore Newcomb and became know as the Bennington study (Newcomb, 1943). More recently, the publication of Theodore Newcomb's final analysis of his famous Bennington study presents a remarkable finding: political beliefs instilled during the childhood and adolescent stages of development may be overturned during a student's collegiate experience and this attitudinal change tends to persist throughout life (Alwin, Cohen and Newcomb,1991). This finding derived from a longitudinal study argues strongly that political socialization continues into young adulthood and that higher education serves as a powerful catalyst, reshaping and changing political beliefs put in place by other, earlier agents of socialization such as the family.
     It was thought that among the most influential agents of political socialization operating on these young women at Bennington were their college professors and peer group and the predominantly liberal views of these two groups. From the perspective of this 1930s study, this conclusion was probably correct. However, Alwin realized the "historical boundedness" of the results of their long-term study and suggests that additional research is needed to assess the impact of other historical periods. Much to the surprise of the authors, we could find no extended discussion or empirical analysis of the Bennington hypothesis.(1) The failure to further explore this important hypothesis provided the impetus for the present study. One major question which the present study examines is to what extent can we generalize from the Bennington study to the university environment in the 1990s?
     The argument made in this study is that the findings obtained from the Bennington study were limited in the sense that the sample consisted of young women who attended a small, elite, private college. The question undertaken in the Bennington study was: what happens to the political beliefs of these young women from politically conservative backgrounds when they encounter an environment which is overwhelmingly liberal? Not surprisingly, Newcomb's findings suggest that a large proportion of these incoming students underwent a change of beliefs bringing them in line with their college environment (Newcomb,1943;and Newcomb, 1967). As previously noted, the longitudinal data derived from that study indicate that these cognitive changes endured over time (Alwin,1991). This finding is persuasive and may still apply to small-colleges with a homogenous political environment which differs from the student's background. However, to what extent do these findings obtain for larger, more politically heterogenous universities which characterize higher education for most students in the 1990s?
     Several other studies have been concerned with the influence exerted by the family and peer groups, however, we were unable to locate any studies which examined the impact of faculty on student political attitudes. In a panel-study of adolescents, Jennings and Niemi (1981) concluded that:

"Weak though some of these correlations are, they do not support an interpretation of a turning away from parents or even the milder statement that youths develop completely independent of parental attitudes" (p. 384).


The Bennington study suggests that under certain circumstances, environmental factors can overcome parental influence. However, more recent research, while not replicating the Bennington study, suggest that parents have a continuing influence over their offspring. Resnick, in a more recent study (1997), looks at the potential conflict between parents and peers and concurs with Jennings and Niemi's findings. "There is a perception that pretty much after adolescence, parents surrender their influence over kids and kids become beholden the peer group." "Everything in this study suggests the contrary."
     Following the implications of Newcomb's findings, this study undertakes a comparative analysis of two universities in order to explore the effects of higher education on the political socialization process in the 1990s. There are two central questions in this study: (1) to what extent does higher education, in general, and professors in particular, shape students' political beliefs in the 1990s; and (2) what role does the student's ideological orientation play in this socialization process?



     The Bennington study of the political socialization of college students focused upon ideological change and the persistence of this change over time. By contrast, the central concern of this study is with some major agents of political socialization and how these agents are associated with the student's positions on four issues. Of central concern to this study is to what extent do college professors influence their students' political attitudes? If faculty exert a significant degree of influence, we should expect to find a high degree of similarity or statistical association between the students' attitudes on specific issues and that of their professors. In this study we are asking: is there a strong socialization effect in the college environment of the 1990s? Does the effect that Bennington discovered in the 1930s obtain in today's institutions of higher education, or has the effect of higher education been significantly muted by other developments in contemporary American society?
     There exist at least two broad differences between the Bennington study and the present study: (1) external environmental differences and;(2) internal environmental differences. When comparing the students' external environment of the late 1930s with the late 1990s, dramatic changes have taken place. In particular, the role of technology has become more integral to the lives of today's student. Electronic media such as television, computers, CDs and audio tape devices were unheard of in the late 1930s. However, today they are standard fare for the typical college student. One implication of these external environmental influences is that the students are exposed to a much broader array of potential agents of socialization and therefore, the potency of the college environment as a socializing agent may exert less influence than was the case in the late 1930s. Since the Bennington study dealt exclusively with female students, it is necessary to examine the changing role of women in society and on campus. While today the majority of students who attend institutions of higher education are females, this was not the case in the 1930s. Some sixty years later, it can be argued that the student population, including women, is less compliant to authority and more questioning than was the case in the past.
     The internal environmental differences between these two studies are related to the fact that Bennington was a relatively small, private, all-women's college and both the University of Oregon and Cal Poly are much larger, more heterogeneous public universities.



     The natural maturation of humans provides an easily understood continuum for the understanding of the socialization process. Infants become children who become teenagers who become young adults. At each stage of this developmental process students of political socialization have attempted to ascertain which agents of influence the political attitudes and values of the individual (Newcomb, 1943;Davies,1965; Greenstein, 1965; Jennings and Niemi, 1968; Easton and Dennis, 1969; Dawson and Prewitt, 1977; Alwin and Cohen, 1991 and Dey, 1997). While there has been much research to suggest that the early childhood stages of socialization are important, the Bennington studies argue for consideration of young adulthood as an important period for fixing political attitudes and values. At this stage of development, the young person is presumably well-integrated into both family and community value systems. However, it is at this juncture that many young people experience an identity crisis which may motivate them to make a break with at least some of their past attitudes and beliefs as they assume more of an adult role and attempt to establish their identity (Erikson, 1968).
     Attending college is, of course, one of the traditional ways in which young people may alter their political cognitions. During their college experience, new and different agents of socialization may come into play. In the educational setting, college students are exposed to new sets of friends, new sources of information and possibly new sets of cognitions as expressed by their significant others such as their peers, faculty, and members of the surrounding community. Often, the student experiences intra-personal conflict as their new environment challenges the lessons of home and family, the media and their previous set of peers. It is at this juncture that there exists the potential for a change in attitudes and values as discovered by Newcomb and Bennington College students.
A variety of political socialization agents are thought to influence the thinking and behavior of the individual. Among the most important are the family, peer-group, college faculty and the surrounding community.


The Family

     An important agent of the political socialization process is the family. The early literature which focused upon the childhood stage of development indicates that the political socialization process begins at a relatively early age (Greenstein, 1967 and Easton and Dennis, 1969). Typically, the child acquires an attitude or opinion about something political--be it a police officer, or the President of the United States--at a relatively early age. If the child's experience with these political objects has been positive and their views have been reinforced by their parent(s), then it is likely that they will have a positive attitude toward these political objects. According to Greenstein, affect precedes information (Greenstein, 1967) In other words, a liking or disliking of a political object takes place before the child has in-depth information about that object.
     One recurring finding in the literature is that parents have an influence over their children's sense of party identification, political ideology and a generalized orientation toward politics (Dawson, Prewitt and Dawson, 1977). This is especially the case if the parent(s) are in agreement over political matters and they verbalize their opinions in the presence of their children. Of central importance to this process is the primacy principle. According to this principle, what individuals learn during the childhood stage of development tends to have an enduring influence over the rest of their life. Inglehart has been more specific in this area, hypothesizing that the socioeconomic conditions present during one's youth will have an enduring influence over the individual (Inglehart, 1990). A recent test of this hypothesis has supported the materialistic value orientations only among younger segments of a longitudinal sample of women (Sangster and Reynolds, 1996).
     Typically, the family is of paramount importance during the childhood stage of development and it is thought to exert a powerful influence--either positive or negative--which will have a lasting impact on the child. In order to exert political influence over an individual, a socializing agent must: (1) communicate clearly to the individual and; (2) the individual must be receptive to whatever message is being communicated (Williams and Minns, 1986). The family tends to accomplish both of these functions. Additional factors which contribute to the influence of the parents over their offspring include the amount of time spent together and the bonds that tie parents and their children together. If these bonds are strong it is likely that the offspring will attempt to please the parents by mimicking their political attitudes (Davies, 1965).

Peer Group Pressure


     Asch (1956) conducted one of the earliest empirical studies on peer group pressure. He demonstrated the potency of peer group pressure in an experimental setting by designing a study in which the subjects were placed in a conflict between trusting their own judgement and accepting the discrepant, and incorrect, judgment of peers. The experiment was "rigged" so that each "confederate" in the experiment openly expressed their views to the group prior to the subject. When faced with clearly false judgements by the confederates, the majority of the subjects expressed views consistent with their peers. Based upon a meta-review of the socialization literature, Harris (1995) argues that among environmental factors, the peer group is of paramount importance and has more influence than even the family.
     According to the early literature, the political attitudes and values of college students were linked directly to their collegiate experience. Moreover, this collegiate experience was viewed as having a more potent influence on the student than their background characteristics and initial political orientations.
     The Bennington study of the political socialization of college students has had a lasting effect on the literature. A major finding to emerge from that study was that the college environment--peers and professors--was able to influence students attitudes in a manner which was sharply divergent from that of their parents. For the most part, the women attending Bennington, came from wealthy, conservatively oriented families from the northeastern United States. By contrast, the Bennington faculty was decidedly liberal and even leftist in their political orientation. What Newcomb found in his initial study was this politically conservative group of female students had changed their political beliefs to a more liberal orientation--an orientation which was consistent with the rather homogeneous political atmosphere which existed at Bennington during the 1930s. Interestingly, Newcomb and his colleagues re-interviewed the original sample of five hundred in two follow-up studies: one in 1959 and 1960; and the second completed in the mid-1980s. Both of these studies suggest that the liberal political orientations adopted by these female students in the 1930s persisted for about five decades.
     As the individual begins to spend more time away from their family, their peer group can take on an important role as a socializing agent. Pressures to conform to group norms can be very powerful and have an impact upon students' political attitudes and values. As previously indicated, peer group pressure is very powerful and may well overcome the students' prior attitudes and values. According to Dey, the peer group agent of political socialization influences students to change in the direction of institutional peer norms irrespective of the social era in which students attended college (Dey, 1997).


     One result of the Bennington study was to foster a wide-spread belief that higher education in general, and college professors in particular, influence their students' political beliefs in a decidely liberal direction. Recent studies, however, indicate only a modest increase in liberalism among students (Dey, 1997). Moreover, the 1997 annual survey by UCLA's Higher Education Institute indicates that the entering freshmen class are continuing to embrace more conservative social values than college students from the last few decades (Sax, Astin, Korn, & Mahoney, 1997). This social conservativism includes attitudes about civil rights, abortion and gay rights and indicates that incoming freshmen value the educational goal "to be very well off financially" more so than their previous cohorts who valued "to develop a meaningful philosophy." Table I illustrates some indicators of political ideology derived from the recent UCLA study.


By comparing a two-decade period in can be argued that the incoming freshmen were not that liberal to begin with and they hold political views today which are even more conservative than their 1977 cohort. Faculty influence on students' political attitudes, beliefs and values is highly variable and depends upon a number of complex situational, individual and interactive variables. In short, this process of faculty influence is quite complex and is thought to be diluted as more potent agents of political socialization operate on the student. In sharp contrast to the late 1930s when the first Bennington study was completed, the student of today is exposed to a much wider variety of potential influences such as a much expanded media.

The Surrounding Community

     Students do not spend all of their time in the university environment. Certainly, a portion of their university experience consists of interactions with members of the surrounding community. In most settings, this interaction has typically been characterized as a "town-gown" conflict. With few exceptions, the university community has been portrayed as more liberal than the surrounding community. While this dimension of political socialization of college students has played a minor role in previous studies, this study examined the relationship between student attitudes and their perceptions of the surrounding communities across the issues of class, race, gender and the environment. The inclusion of this dimension provides a means to examine the extent to which town/gown attitudes are associated.


In order to begin to evaluate the relationship between the students and the four socialization agents discussed above, we asked students to rank themselves, as well as their family, friends, professors and members of the community on four issues: class, race, gender and the environment. A ten-point scale was utilized in this ranking which extended from very conservative­assigned a value of one--to very liberal­assigned a value of ten. The mean scores, difference scores and independent samples t-tests for these rankings are presented in Table II.


     In examining the table, a number of interesting findings emerge. First, the students' ideological perceptions are quite different from one university to the other. Since Cal Poly is a polytechnic university in which over half of the student body majors in agriculture, architecture, business and engineering, the student body is quite politically conservative when compared with other state universities in California. By contrast, students at the University of Oregon have a reputation of being politically more liberal. The data support this view. Across all twenty of the mean scores, the University of Oregon students classified themselves as more liberal than their counterparts at Cal Poly. The difference scores between the two campuses ranged from +0.2 on their respective ranking of their families on the issue of environment to +2.6 for their ranking of their college communities on the issue of class. With the exception of three pairings­-family on the issues of class, race and the environment­the t-tests indicate statistically significant differences between the ideological orientations of these two student samples.
     Second, there was a tendency for students at both universities to perceive themselves as located ideologically in between their family and their professors. Students at the University of Oregon expressed this relationship as a "tug-of-war" between their more conservative family and their more liberal professors.
     Thirdly, student perceptions of their surrounding communities was markedly different. Across all four issues, University of Oregon students perceived the city of Eugene, Oregon as left of center. By contrast, Cal Poly students tended to view San Luis Obispo, California as right of center.
     Do ideological variations among students lead to differences in how they rank order the influence of agents of political socialization? For example, are conservative students more subject to influence from their families than their liberal counterparts? Conversely, are liberal students less influenced by their families and more influenced by other agents of political socialization? In order to begin to examine these relationships, simple bivariate correlations were computed between the students and the various agents of socialization across the four issues. In Table III, Pearson correlation coefficients between the students and other potential socializing agents across these four issues are displayed. Clearly, these correlations indicate that the strongest relationships exist between students and their families.

This relationship obtains at both universities and across the four issues. The second strongest association exists between the students and their friends. In both cases, each relationship is statistically significant across the four issues. By sharp contrast, correlations for the community and the faculty are very small and for the most part insignificant. Students from both universities rank their communities and professors as considerably different from themselves in relation to these four issues. With the exception of Cal Poly students perceptions about themselves and the surrounding community on the environment, none of the other coefficients of these two agents is statistically significant. Thus, these finding suggest that the student's family is the most important agent of political socialization, followed by friends. What is significant in Table III is the absence of a statistically significant relationship between students and faculty across these four issues.
     This finding runs counter to the previous literature (Erikson and Tedin, 1995). While simple correlations have their limits, these data portray a picture which is in sharp contrast to the Bennington study in the sense that college faculty of today, at these two public universities, appear to have little impact on their students' political attitudes.
     The lack of a statistically significant relationship between students and faculty is consistent with a previous study of Cal Poly students. In that study, students were asked to describe three well-publicized events which had occurred before the study. These events included: (1) The first Rodney King verdict; (2) Anita Hill's testimony at Clarence Thomas' U.S. Supreme Court confirmation hearing; and (3) the Exxon-Valdez oil spill. After the students had written down their descriptions of these three events, they were asked to indicate: (1) their most important source of information about each of these three events and; (2) the source they ranked as most important in evaluating these events. Across all three of these events, they ranked faculty at or near the bottom of the information and evaluation scales.
     Based upon these simple bi-variate relations, students perceive themselves as most similar to their family and friends and least similar to their surrounding community and professors. Although these students tend to be more "liberal" than their parents, they are not following the lead of their professors. At both universities and across all four issues, there is an absence of statistical significance between students and faculty.
In Table IV these data are presented in a multivariate linear regression model in which student attitudes are regressed on family, friends, faculty and college communities.




    2 In Figure 1, a series of scatterplots for all four issues and the four agents of socialization across both universities indicate that the standard OLS assumption of normally distributed residuals is easily met by the data in this model. Note that the Loess smoother deviates little from the regression line (Cleveland,1979).

In this form the previous findings appear in a format which can estimate the simultaneous effects of these four agents of socialization. Here again, family and friends appear to have strong (and statistically significant) independent effects, while there is no evidence that faculty and community exercise an effect on student attitudes. The one exception is on the environmental issue between Cal Poly students and the surrounding community. These findings are quite robust and hold across issues of class, race, gender and the environment. Simply put, our data suggest that with the possible exception of friends, the environment of higher education in these two state universities appears to have little effect on students' political attitudes. This thesis is strengthened by the fact that students at both universities perceive significant differences between themselves and their professors across the four issues included in the study. These findings differ markedly from the Newcomb findings of the 1930s.



     There are major differences between the Bennington study and the study presented here. (1) The studies represent two markedly different eras. In the case of Newcomb's study, students did not have access to the wide array of technological devices which are available to today's student. (2) There are major differences between a small all-womens private college and the two public universities which were utilized in this study. Apart from the size differences, both the student body and faculty at Bennington appeared to be far more homogeneous than the student bodies and faculties at Cal Poly and the University of Oregon. In the Bennington study, the women were from predominantly conservative, upper-class families from the northeastern part of the United States. The faculty at Bennington was also fairly homogeneous in the sense that it was a progressive faculty with a decidedly liberal to "left-wing" bent. Students from both Cal Poly and the University of Oregon are far more diverse than the Bennington women. With an increasing emphasis upon diversity, the student bodies of both universities contained in this study are very heterogeneous in terms of social economic status, ethnicity, gender and political attitudes. Moreover, faculty at both state universities are far more diverse ideologically than the Bennington faculty. In a recent study done of Cal Poly faculty, only 18% of considered themselves to be either far left or strong liberals (Grafina, 1996). By contrast, 26% of the faculty considered themselves to be conservatives. Although similar data is not available for University of Oregon faculty we suspect that while they are more liberal politically, nevertheless, this ideological label does not characterize all or perhaps even most of them.
     When we compared the two student samples, we uncovered striking differences. The Cal Poly sample was consistently more conservative than their University of Oregon cohort and, with the exception of three pairings, the difference scores in their ideological rankings of self and other agents of political socialization were statistically significant. By contrast, the bivariate correlations and multiple regression models of the student attitudes across the issues of class, race, gender and the environment were found to be strikingly similar. Irrespective of their political ideology, statistically significant correlations were observed between both student samples and their friends and family across the issues. It did not matter whether the students' identified themselves and others as politically liberal or conservative, nevertheless, they viewed themselves as most closely allied with first, their family and second, their friends. No other potential agent of political socialization was viewed as consistent with them across the four issues. Importantly, in no case did we uncover a significant relationship between students and faculty.
     These findings suggest that in a relatively heterogeneous political environment: (1) the student's political ideology is unimportant in their ranking of the potency of the various agents of political socialization; and (2) contrary to prevailing belief, college professors appear to exert little, if any, influence over their students political attitudes. Theoretically, these findings could be explained on the basis of cognitive dissonance theory (Festinger,1957). According to Festinger, in an environment in which the individual is exposed to conflicting views, they will rely upon a "least effort principle" to attempt to reduce the conflict. That is, those agents of political socialization­-their family and friends­which tend to reinforce rather than conflict with their political attitudes. The primacy principle supports this theoretical interpretation of the potency of the family, while the notion that we tend to select friends who share similar viewpoints supports the efficacy of friends as an agent of political socialization.
     While we can only speculate on this matter, why do faculty appear to have little, if any, influence over their students political attitudes in the late 1990s? One possible explanation is the "Rodney Dangerfield" effect. This explanation argues that present-day college students have little respect for their professors. Faculty report increasing incidents of rude and disrespectful behavior which were unheard of in previous decades. Perhaps this phenomenon is related to a growing disrespect for authority figures in general. Second, and as previously mentioned, the college student of today is exposed to many more potentially influential agents of political socialization than was the case in the past. The net result of this phenomena is to dilute the influence of college faculty. Finally, the amount of time students spend interacting with faculty in the two state universities in this study is probably less than was the case at Bennington college during the late 1930s. A small private college such as Bennington afforded the faculty more time to interact with their students than is the case today at larger, less personal state universities. As a result, faculty influence over the student of today is minimized.



Alwin, D.F.,Cohen, R.L., & Newcomb, T.M. (1991). Political attitudes over the life span. Madison: University of Wisconsin Press.

Asch, S.E.,(1956). Studies of independence and conformity: a minority of one against a unanimous majority. Psychological Monograph, 70, No. 9 (Whole No.416).

Brim, O.G., Jr. (1966). Socialization through the life cycle. In O. G. Brim Jr. and S. Wheeler, Socialization after childhood.

Cleveland, W.S. (1979. Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association, 74, 829-36.

Davies, J.C. (1965). The family's role in political socialization. The annals, political socialization: Its role in the political process. Philadelphia: The American Academy of Political and Social Sciences.

Dawson, R., Prewitt, K. and Dawson, K. (1977). Political socialization. 2nd ed. Boston: Little.

Dey, E.L., (1997). Undergraduate political attitudes. Journal of Higher Education, 68, 398-413.

Easton, D. and Dennis, J. (1969). Children in the political system.  New York: McGraw-Hill.

Erikson, E.H. (1968). Identity Youth and Crisis. New York: W.W. Norton & Company.

Erikson, R.S. and Tedin, K.L. (1995). American public opinion. Needham Heights, Massachusetts: Allyn and Bacon.

Festinger, L. (1957). A theory of cognitive dissonance. Stanford California: Stanford University Press.

George, D.L. and Medler, J.F. (1993). The mass media as adult political socialization agents. Paper presented at the Western Political Science Association. Pasadena, CA.

Graffigna, M. (1993). Political socialization at cal poly: The political beliefs of professors and students. Unpublished senior paper.

Greenstein, F.I.(1965). Children and politics. New Haven, Connecticut: Yale University Press.

Harris, J.R. (1995). Where is the child's environment? A group socialization theory of development. Psychological Review, Vol. 102, No. 3., 458-489.

_______, (1998). The nurture assumption. New York: The Free Press.

Hyman, H.H. (1959). Political socialization. Glencoe, Illinois: The Free Press.

Inglehart, R.(1990). Culture shift in advanced industrial society.  Princeton: Princeton University Press.

Newcomb, T.M., Personality and social change: Attitude formation in a student community (1943). New York: John Wiley and sons.

Newcomb, T.M., Koenig, K.E., Flacks, R., and Warwick, D.P.(1967). Persistence and change: Bennington college and its students after twenty-five years. New York: John Wiley and Sons.

Pascarella, E.T. and Terizini, P.T. (1991). How college affects students. San Francisco: Jossey-Bass.

Resnick, M.D., Protecting adolescents from harm. The Journal of the American Medical Association, Vol. 278, 823-832.

Sax, L.J., Astin, A.W., Korn, W.S., and Mahoney, K.M. (1997). The american freshman: National norms for fall 1997. Los Angeles: The Higher Education Research Institute.

Singer, S. (1968). Katz (ed.), Growth and constraint in college students. Stanford, CA: Stanford University, Institute for the Study of Human Problems.

Sangster, R.L. and Reynolds, R.W. (1996). A test of inglehart's socialization hypothesis for the acquisition of materialist/postmodern values: The influence of childhood poverty on adult values. Political Psychology, 17(2), 253-269.

Williams, C.B. and Mimms, D.R.(1986). Agent credibility and receptivity influences in children's political learning.
Political Behavior, 8 (2): 175-200.

 1. We reviewed eight-four articles in The American Political Science Review and seventy-one articles in the American Sociological Review and other similar journals which mentioned Bennington in the text. Our APAR review extended back to 1970, while the other reviews extended back to the 1930's.