Anchoring and estimation of alcohol consumption: implications for social norm interventions.
Subject: College students (Case studies)
Alcoholism (Case studies)
Authors: Lombardi, Megan M.
Choplin, Jessica M.
Pub Date: 08/01/2010
Publication: Name: Journal of Alcohol & Drug Education Publisher: American Alcohol & Drug Information Foundation Audience: Academic; Professional Format: Magazine/Journal Subject: Health; Psychology and mental health; Social sciences Copyright: COPYRIGHT 2010 American Alcohol & Drug Information Foundation ISSN: 0090-1482
Issue: Date: August, 2010 Source Volume: 54 Source Issue: 2
Topic: Event Code: 830 Sales, profits & dividends; 242 Advertising Advertising Code: 52 Advertising Activity Computer Subject: Company earnings/profit
Product: Product Code: E197500 Students, College
Accession Number: 236247804
Full Text: ABSTRACT

Three experiments investigated the impact of anchors on students' estimates of personal alcohol consumption to better understand the role that this form of bias might have in social norm intervention programs. Experiments I anal II found that estimates of consumption were susceptible to anchoring effects when an open-answer and a scale-response format were used. Experiment III utilized a design that communicated social norm information as a previous social norm intervention had done and found that self-reported binge drinking was reduced though actual consumption could not have changed. Implications for the use and assessment of social norm intervention as a component of alcohol education are discussed including the pessimistic possibility that social norm interventions may not be affecting students 'actual consumption.

**********

Social norm intervention is an important tool that many colleges and universities across the country have utilized to educate students about their peers' drinking habits. The idea behind implementing a social norm intervention is that dangerous alcohol consumption can be reduced by educating students about the actual norm for their peer group which tends to be much smaller than the norm that students perceive (Perkins & Berkowitz, 1986). Since students' drinking behaviors are often correlated with what is normal for their peer group (Clapp & McDonnell, 2000; Perkins, 1985; Perkins & Berkowitz, 1996), providing students with factual information about their peers' drinking habits is thought to have a direct effect of reducing student consumption (Haines & Spear, 1996). This strategy is often compared to traditional alcohol education methods such as counseling or health information sessions highlighting the dangers of drug and alcohol use (Perkins, 1997), which appear to have little positive effect on students' behaviors (Rosenbaum & Hansen, 1998; Rosenbaum, Flewelling, Bailey, Ringwalt, & Wilkinson, 1994; Tobler, 1986). The research reported here suggests that studies of social norm intervention effectiveness may be confounded by anchoring effects (Tversky & Kahneman, 1974) and so might not be effectively assessing the success of these programs. This pessimistic possibility would imply that social norm interventions might be no more successful than traditional alcohol education methods such as counseling or health information sessions in reducing dangerous drinking behaviors.

Social norm intervention programs aimed at reducing problematic drinking on college campuses typically consist of a social norm marketing campaign in which counselors provide students with peer norm information via posters, print advertisements, or word of mouth (Gomberg, Kessel Schneider, & DeJong, 2001; Haines & Spear, 1996; Johannessen, Collins, Mills-Novoa, & Glider, 1999; LaBrie, Pedersen, Huchting, Thompson, & Hummer, 2008). Regardless of how the norm information is presented, identical surveys are collected before and after the intervention, which typically lasts from a few weeks (Mattern & Neighbors, 2003) to a few months (Gomberg et al., 2001) or even to a number of years (Haines & Spear, 1996).

A main component of many social norm intervention programs, especially those with long intervention duration, is a follow-up with students to ensure that the campaign message and norm information are being attended to and processed. Follow-up procedures sometimes involve offering money or prizes to students who accurately recall the campaign information (Gomberg et al., 2001; Haines & Spear, 1996); or mailing postcards with norm information to students (Mattern & Neighbors, 2003) to reinforce the campaign message and norm information; or involve a sufficient number of posters designed to convey social norm information so that students will not miss the message (LaBrie et al., 2008). At the end of the intervention period counselors determine whether the intervention successfully led to a reduction in norm misperception as well as a reduction in dangerous drinking habits (Haines, 1996; Haines, Perkins, Rice, & Barker, 2005; Johannessen et al., 1999). This assessment of intervention effectiveness is made by comparing students' pre and postintervention survey responses (Haines, 1996; Haines, Perkins, Rice, & Barker, 2005; Johannessen et al., 1999). If students self-report lower levels of personal alcohol consumption postintervention, researchers conclude that the intervention effectively reduced dangerous drinking behavior on campus.

Studies assessing the effectiveness of these programs have shown mixed results. Many of these studies report that drinking behavior is reduced after students have been exposed to the intervention (Gomberg et al., 2001; Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003; Neighbors et al., 2004; Perkins & Craig, 2006). In these studies, students who report drinking more than the norm prior to the intervention tend to report drinking less after the intervention. Some studies also report that drinking behavior is increased in some students after they have been exposed to the intervention (Mattern & Neighbors, 2003). These studies show an increase in drinking behavior among students that initially report drinking less than the norm. While both findings may seem consistent with social norm perspectives, which suggest that students adjust their drinking toward the norm, it may be that neither finding is indicative of changes in student consumption.

Instead of having the effect of educating students about the norm for their peer group and modifying student drinking behavior, it may be that social norm intervention programs are merely providing students with anchors that they then utilize to estimate their personal consumption (Tversky & Kahneman, 1974). If this is the case, then social norm intervention programs might not be effectively assessing the success of their campaigns; the education that students are receiving from social norm interventions may merely be affecting their estimates of alcohol consumption through an anchoring effect rather than changing actual behavior. Furthermore, just as some students may disbelieve the presented norm (Perkins, 2007), which could thereby reduce social norm effects, so some may disbelieve anchor values, which could thereby reduce anchoring effects (Mussweiler, Strack, and Pfeiffer, 2000). The findings typically observed in studies of social norm intervention effectiveness are, therefore, consistent with the possibility that norms act as anchors that bias estimates instead of changing student behavior.

In addition, the most common method for assessing students' drinking behavior in these studies is through self-report questionnaires (Gomberg et al., 2001 ; Haines & Spear, 1996; Johannessen et al., 1999) in which students are asked to report details of their alcohol consumption. Typically, students are asked questions related to the frequency and/or quantity of their alcohol consumption (Gomberg et al., 2001; Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003) such as the number of drinks they consume when they 'party' (Haines & Spear, 1996; Johannessen et al., 1999). Responses are most often gathered using frequency scales in which students estimate the number of drinks along a pre-determined scale (Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003), although some studies have used an open-answer format in which students provide an estimate of the exact number of drinks they consume (Gomberg et al., 2001).

Since the use of self-report increases the likelihood that students would rely on estimation when reporting their alcohol consumption, there are likely to be unintentional biases such as anchoring effects (Tversky & Kahneman, 1974) on the participants' responses. Previous researchers noted that self-reporting is vulnerable to social desirability effects whereby responses are biased toward values participants believe are socially desirable (Walsh & Braithwaite, 2008). However, anchoring effects are even more pernicious in that they occur on occasions where social desirability is not a factor, such as when estimating neutral values (Jacowitz & Kahneman, 1995; Tversky & Kahneman, 1974) and when the anchor is presented in an irrelevant task (Wilson, Houston, Etling, & Brekke, 1996). Controlling for social desirability effects would, therefore, still leave participants vulnerable to anchoring effects.

Anchoring is a phenomenon in which people's estimates of quantitative values are biased towards values known as "anchors" (Tversky & Kahneman, 1974). An anchoring effect is said to have occurred when the numerical estimates of individuals who are exposed to different anchors are biased toward the anchor (Jacowitz & Kahneman, 1995). Individuals who have been exposed to a higher anchor provide estimates that are closer to that high anchor while those who have been exposed to a low anchor provide estimates that are closer to the low anchor.

Tversky and Kahneman (1974) demonstrated these effects in a study in which they asked participants to identify the percentage of United Nations (UN) member nations that were African. They provided their participants with an arbitrary initial percentage (anchor) and asked whether the percentage of UN countries that were in Africa was greater or less than the provided percentage. Participants were then asked to estimate the actual percent of UN nations that were African. The researchers found that participants' responses were biased toward the initial percentage that the participants received. For example, when participants were provided with an initial percentage of 10% they estimated that 25% of UN nations were African whereas they estimated that 45% were African when given 65% as an initial percentage. Simply changing the anchor that is provided can change participants' responses.

Anchoring effects have been found to influence many types of estimates including those related to general knowledge such as the year of DaVinci's birth (Strack & Mussweiler, 1997) as well as attributes such as size, age, or amount (Strack & Mussweiler, 1997) and estimates of price (Mussweiler, Strack, & Pfeiffer, 2000). These effects have been found in both laboratory and real-world settings and in cases where anchors were provided by an outside source, such as an experimenter, and cases where anchors were self-generated. Research has shown that anchors are robust enough to influence estimates made several days to one week after the anchor is presented to an individual (Mussweiler, 2001), and the effect is so strong that individuals are affected by it even when they have been pre-warned of the effect (Wilson, et al., 1996).

Research on anchoring effects has also found that the size of the anchoring effect is reduced with greater certainty (Tversky and Kahneman, 1974; Jacowitz and Kahneman, 1995). According to this finding, the more persons know about the value they are trying to estimate, the less they will rely on anchors. Since people have a great deal of knowledge about their personal drinking habits, they might be less vulnerable to anchoring effects when reporting their own drinking behaviors than when reporting less well-known dimensions, which may explain why previous research has not explored the relationship between anchoring and reports of alcohol consumption. Individuals may, nevertheless, not know their exact amounts of consumption, which would require them to estimate and leave them vulnerable to anchoring effects.

In addition to providing students with numerical norm information that may essentially act as an anchor, social norm intervention programs also seem to hinge on frequent and powerful presentation of this information (Perkins, 1997). The most effective studies presented norm information for an extended time frame and utilized some follow-up method (Gomberg et al., 2001; Haines & Spear, 1996) to ensure that the information was frequently available and could be easily recalled from memory. Highly prevalent norm messages and implementation of follow-up procedures increase the likelihood that individuals will recall the norm information when making an estimate of their own alcohol consumption, which is an important element behind the anchoring effect. While anchoring studies have shown that an anchor can persist for a few days to one week after presentation (Mussweiler, 2001), social norm intervention studies assessed alcohol consumption while the norm information was still available in memory. By ensuring that norm information is salient and easy to remember to possibly affect students' actual drinking behaviors, social norm intervention programs also increase the likelihood that the numerical information will be salient and well-remembered enough to affect estimates of drinking behavior over an extended period of time, which would thereby confound any observed effects.

In short, social norm intervention studies that have reported successful reductions in college students' alcohol consumption have utilized methodologies that make it difficult to discount the possibility that an anchoring effect is influential in the outcome. Three experiments examined the influence of anchors and social norms that could act as anchors on estimates of personal consumption. Experiments I and II investigated whether anchors influence estimates of personal alcohol consumption despite the possibility that participants have a great deal of knowledge about their own behaviors such that reported consumption would be higher in the high anchor condition than in the low anchor condition. Experiment I utilized an open-answer response format (see Gomberg et al., 2001) while Experiment II utilized a scale-response format (see Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003).

Experiment III presented social norm information in a format that was almost identical to a previous social norm intervention study and assessed reports of binge drinking. In particular, this experiment used a poster adapted directly from Johannessen et al. (1999) to investigate whether exposure to this poster would lower self-reports of binge drinking when there was no difference in actual consumption. It was expected that the results of this experiment would parallel those of studies of social norm intervention effectiveness that demonstrate a reduction in the percentage of students reporting binge drinking after the intervention (Haines & Spear, 1996) albeit without any actual change in binge drinking behavior. It is important to explore the influence of anchors on reports of binge drinking since there are many negative consequences to this drinking pattern (see Ham & Hope, 2003 for a review) which leads many alcohol education programs to specifically focus on reducing this form of drinking. Also, examining the anchoring effect in this context may provide an explanation for a discrepancy between social norm intervention reports of reductions in the percentage of students reporting binge drinking (Haines & Spear, 1996; Johannessen et al., 1999) and reports by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) that indicate binge drinking has continually increased among college students over the past few years (Mitka, 2009).

EXPERIMENT I

The purpose of Experiment I was to investigate whether social norm information might serve as anchors that affect estimates of alcohol intake when responses are collected using an open-answer response format (see Gomberg et al., 2001).

METHOD

Participants

The participants were 155 (75 male, 80 female) individuals who were approached by the experimenter at public locations on campus and randomly assigned to one of three conditions. Participation was voluntary. Five responses were excluded from the analysis, because they were two standard deviations or more away from the mean for each group, leaving a sample of 150 participants (71 male, 79 female). There were 50 participants in the control group, 50 in the high-anchor group, and 50 in the low-anchor group.

Materials

The questionnaire consisted of three questions. The first asked whether participants think they drink more than, less than, or approximately the same amount as the average student at the university. The second was an open-ended question that asked how many drinks they consume per week on average. A box at the top of the questionnaire defined one drink as being equal to 4 oz. wine, 10oz. wine cooler, 12 oz. beer, 1 cocktail with 1 oz. of 100 proof liquor or 1 1/4 oz. of 80 proof liquor. The consumption question and definition were worded similarly to the types of questions and definitions utilized in social norm intervention studies (see Haines & Spear, 1996; Gomberg et al., 2001). A third, and final, question asked participants their gender.

Three versions of this questionnaire were used to manipulate the type of anchor that each participant received. A high anchor version of the questionnaire included a statement at the top that claimed the norm for students at the university was drinking 19 drinks per week. A low anchor version of the questionnaire included a statement at the top that claimed the norm for students at the university was drinking 1 drink per week. A control version of the questionnaire did not include a norm statement at the top.

Norm information for these questionnaires was adapted from previously collected drinking information for students at the university without exposure to norm information. In a pretest, participants reported drinking a mean of 9.89 drinks per week, on average. This number was rounded to 10 drinks and the high and low anchors were developed by adding and subtracting 9 drinks from the mean.

Procedure

Participants were approached by an experimenter and asked if they would be willing to participate in a study on students' drinking habits. Participants were informed that their responses would be completely anonymous. Participants received one of the three randomly presorted questionnaires and were asked to fill it out as completely and honestly as possible. In order to reduce experimenter effects, the experimenter walked a few feet away from participants while they completed the questionnaire. Participants were also approached individually to ensure that participants in the control condition did not accidentally view the norm message and that peers did not influence estimates. Once they finished the questionnaire, participants placed it in a large manila envelope and returned it to the researcher. The questionnaire typically took fewer than five minutes to complete. Once the questionnaire was returned to the researcher, each participant was debriefed to ensure that he or she was aware that the information provided regarding the norm for the average student was fictional, and that it was provided in order to assess its impact on his or her personal consumption estimate. The university institutional review board (IRB) reviewed all procedures prior to data collection.

RESULTS

A one-way, 3-group analysis of variance (ANOVA) revealed a significant difference in the mean estimates of the total number of drinks consumed in an average week for the three conditions (F (2, 147) = 4.07,p = .019). A post hoe Least Squared Differences (LSD) analysis revealed that there was a significant difference between the low (n = 50, mean = 4.28, SD = 4.56) and control (n = 50, mean = 6.90, SD = 6.40) groups. There was also a significant difference between the high (n = 50, mean = 7.44, SD = 6.61) and low groups. There was not a significant difference between the high group and the control group. These results support the suggestion that alcohol estimates can be influenced by anchors despite the high level of knowledge that individuals are likely to have about their own behaviors.

EXPERIMENT II

The purpose of Experiment II was to investigate whether social norm information might serve as anchors that affect estimates of alcohol consumption when responses are collected using a scale response format (see Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003).

METHOD

Participants

Participants for this study were 150 (75 male, 74 female, 1 unspecified) individuals on a mid-western university campus. Participants were recruited from several locations such as the student and academic centers on campus and were randomly assigned to one of three conditions. Participation was voluntary. Three responses were excluded from the final analysis because they were two standard deviations or more away from the mean for each group resulting in a sample of 147 (72 male, 74 female, 1 unspecified) participants. There were 47 participants in the low group, 50 in the control group, and 50 in the high group.

Materials & Procedure

The questionnaire for Experiment II was almost identical to the questionnaire for Experiment I except that the second question which asked participants how many drinks they consume per week on average used a scaled-answer format. The scale was anchored with zero on the low end and "more" on the high end with options of 1-3, 4-6, 7-9, 10-12, 13-15, 16-18, and 19-21 in between. The high-anchor version of the questionnaire included a statement at the top that claimed that the norm for students at the university was drinking 16 drinks per week. The low-anchor version included a statement at the top that claimed the norm for students at the university was drinking 4 drinks per week. These values were used rather than the values that were used in Experiment I, so that neither anchor would represent the highest or lowest values on the scale. A control version of the questionnaire did not include a norm statement at the top. The procedure for this experiment was identical to the procedure followed in Experiment I.

RESULTS AND DISCUSSION

Since consumption estimates were made on a response scale, the overall mean consumption estimates were calculated based on the average for each range on the scale. For example, an estimate in the range of one to three drinks was entered as two drinks. Estimates of "more" were entered as 22 drinks (one more than the highest numerical option on the scale). A one-way ANOVA revealed a significant difference between the three conditions (F (2, 144) = 6.24, p = .003). A post hoc Least Squared Differences (LSD) analysis revealed a significant difference between mean estimates for the low (n = 47, mean = 4.53, SD = 4.11) and the control group (n = 50, mean = 8.26, SD = 6.91) and a significant difference in mean estimates for the high (n = 50, mean = 8.62, SD = 7.22) and the low group. There was not a significant difference between the high and the control group. These results demonstrate that anchors can influence alcohol consumption estimates made on a response scale (see Haines & Spear, 1996; Johannessen et al., 1999; Mattern & Neighbors, 2003).

The results of Experiments I and II suggest that estimates might be influenced in a downward direction more than in an upward direction. This finding might initially seem inconsistent with previous research on anchoring effects (Tversky & Kahneman, 1974). One possible explanation for this finding is that the high anchor may have been too high and, as a result, was not perceived as plausible. The norm used in Experiment I, 19 drinks, may have fallen outside of the distribution of plausible values (Mussweiler, Strack, & Pfeiffer, 2000) for college students' alcohol consumption. As a result, students may have discounted the anchor outright or the anchor may not have resulted in an adequate memory search for instances of behavior that were consistent with the anchor (Mussweiler, Strack, & Pfeiffer, 2000). In selective accessibility (Mussweiler, 2001) terms, it may have been more difficult for students to activate anchor consistent knowledge of drinking 19 or 16 drinks than it was to activate instances of drinking 1 or 4 drinks.

EXPERIMENT III

The purpose of Experiment III was to investigate whether the presentation of social norm information in a poster adapted directly from a previous study (Johannessen et al., 1999) would affect self-reported binge drinking behavior when there was no difference in actual alcohol consumption. Some social norm intervention studies have reported a decrease in the percentage of students reporting binge drinking after the intervention compared to those reporting binge drinking before the intervention and have used this information to support their conclusion that the intervention resulted in a successful reduction in dangerous student drinking habits (Haines & Spear, 1996; Johannessen et al., 1999).

Binge drinking is most commonly defined as consuming five or more drinks per occasion for men and four or more drinks per occasion for women (Wechsler & Nelson, 2001) though this definition is controversial (Segrist & Pettibone, 2009). Some researchers argue that this definition is arbitrary in that it does not correspond to excessively high blood alcohol levels for most individuals (Lange & Voas, 2001) or because many individuals, especially college students, do not consider this definition to be representative of 'problematic' behavior (Goodhart, Lederman, Stewart, & Laitman, 2003; Segrist & Pettibone, 2009). Many college students agree that the 5/4 definition of binge drinking is significantly lower than the number of drinks they would consider to be problematic (Segrist & Pettibone, 2009); college students are more likely to perceive excessive or dangerous drinking as closer to six drinks over a two-hour time period for both males and females (Posavac, 1993). Since many social norm intervention programs utilize the 5/4 definition of binge drinking (including Johannessen et al., 1999), the current experiment uses the 5/4 definition in an effort to demonstrate that it is possible to find the effect observed by previous studies without changing actual behavior.

METHOD

Participants

The participants were 100 individuals (54 male, 45 female, 1 unspecified) recruited by the experimenter from public locations on campus and randomly assigned to one of three conditions. Participation was voluntary. Three responses were excluded from the final analysis because they were three standard deviations away from the mean for each group resulting in a sample of 97 participants (54 male, 42 female, 1 unspecified). There were 47 participants in the poster group and 50 in the control group.

Materials & Procedure

The presentation of social norm information for Experiment III was almost identical to a poster used in a previous social norm intervention study (i.e., Johannessen et al., 1999), except that the words "DePaul University" replaced the words "University of Arizona" on the poster utilized for the current experiment. The poster included the text "Most DePaul students have 4 or fewer drinks when they party" with the numeric information (the number 4) being most prominent (appearing in the center of the poster in the largest font size). The bottom of the poster included information regarding the common definition of one drink as being equal to 12 oz. of beer, 4-5 oz. of wine, or 1 oz. liquor. The poster also contained information suggesting it was created by "Campus Health Service". The questionnaire for this experiment was identical to the questionnaire utilized in Experiments I and II with one exception. Instead of being asked how much they consume per week on average, participants were instead asked to provide the number of drinks they consume when they party in order to be consistent with the norm information on the poster and the questions utilized in the Johannessen et al. (1999) study. For the experimental condition, the poster was attached to the front of the questionnaire. The procedure for Experiment III was identical to that of Experiments I and II.

RESULTS

Binge drinking was defined as consuming 5 or more drinks per occasion for males and 4 or more drinks per occasion for females (Johannessen et al., 1999). It was expected that the percentage of students reporting binge drinking would be lower in the poster group than in the control group. In the control group, twenty-six out of fifty, or 52%, of students reported binge drinking. In the poster group, seventeen out of forty-seven, or 36%, reported binge drinking. The percentage of students reporting binge drinking in the poster group was significantly different from the percentage of students reporting binge drinking in the control group ([chi square] (1, n = 97) = 5.2, p = .02). These results are consistent with the findings of social norm intervention studies that report a decrease in binge drinking following exposure to social norm campaigns that utilize posters to disseminate norms (Haines & Spear, 1996; Johannessen et al., 1999).

GENERAL DISCUSSION

The results of Experiments I and II demonstrate that students' estimates of their personal alcohol consumption can be influenced by anchors despite the student's knowledge of his or her personal drinking behavior. In Experiment III, the finding that the percentage of students reporting binge drinking was lower in the poster condition compared to the control condition demonstrates that it is possible to replicate the results of social norm intervention studies utilizing the same format of norm administration without changing actual binge drinking behavior. Taken together, these results raise the possibility that social norm intervention programs aimed at combating problematic drinking on college campuses may not be effectively educating students and reducing dangerous drinking behavior. In order to accurately assess the effectiveness of these programs, and determine whether they are more successful than traditional alcohol education methods, more objective forms of assessment may be necessary to measure true changes in student drinking behavior post-intervention.

Given the possibility that social norm intervention programs inadvertently introduce anchoring biases into students' self-reports of their personal alcohol consumption, methods of assessing alcohol education programs that are not susceptible to anchoring effects should be developed. While methods have been developed to address social desirability concerns in the use of self-report (see Walsh & Braithwaite, 2008 for an example), these methods cannot specifically address the biases that may occur as a result of anchoring, and new methods must be utilized to account for this concern as well. One suggestion to educators intending to utilize a social norm intervention program would be to rely more on behavioral assessment (e.g., analyzing students' consumption 'diaries' or assessing alcohol sales in the areas surrounding a college campus) than on measures of consumption that rely on estimation. Schultz, Nolan, Cialdini, Goldstein, and Griskevicius (2007), for example, ran a social norm intervention study on electricity use that did not rely upon estimation. Another suggestion would be for social norm messages to present peer drinking habit information in such a way as to avoid the use of numeric information that could essentially act as an anchor.

It is important to note that the current findings are not intended to suggest that past or current alcohol education programs that utilize the social norm intervention method are not successfully educating students and reducing dangerous consumption. The Schultz et al. (2007) electricity-use intervention, for example, found reductions in energy consumption that could not have been attributed to an anchoring effect since their measures did not involve estimation. Instead, the research presented here should allow educators to consider the possible effects of anchors in interpreting previously reported alcohol-related social norm intervention programs. We also hope these results will allow educators to consider the possible effects of anchors when planning and assessing future programs utilizing the social norm intervention methodology to reduce dangerous drinking behavior.

Any correspondence concerning this article should be addressed to: Megan M. Lombardi, Department of Psychology, DePaul University, 2219 North Kenmore Avenue, Chicago, IL 60614-3504, Phone: (224) 656-5212 (773) 325-2052; Fax: (773) 325-7888; Email: mlombar7@depaul.edu.

REFERENCES

Clapp, J.D. & McDonnell, A.L. (2000). The relationship of perceptions of alcohol promotion and peer drinking norms to alcohol problems reported by college students. Journal of College Student Development, 41, 19-26.

Gomberg, L., Kessel Schneider, S., DeJong, W. (2001). Evaluation of a social norms marketing campaign to reduce high-risk drinking at the University of Mississippi. American Journal of Drug and Alcohol Abuse, 27, 375-389.

Goodhart, F.W., Lederman, L.C., Stewart, L.P., & Laitman, L. (2003). Binge drinking: Not the word of choice. Journal of American College Health, 32, 44-46.

Haines, M. (1996). A Social Norms Approach to Preventing Binge Drinking at Colleges and Universities. The Higher Education Center for Alcohol and Other Drug Prevention, U.S. Department of Education.

Haines, M., Perkins, H.W., Rice, R.M. and Barker, G. (2005). A Guide to Marketing Social Norms for Health Promotion in Schools and Communities. National Social Norms Resource Center.

Haines, M.P. and Spear, S.F. (1996). Changing the perception of the norm: A strategy to decrease binge drinking among college students. Journal of American College Health, 43, 134-140.

Ham, L.S. and Hope, D.A. (2003). College students and problematic drinking: A review of the literature. Clinical Psychology Review, 23, 719-759.

Jacowitz, K. and Kahneman, D. (1995). Measures of anchoring in estimation tasks. Personality and Social Psychology Bulletin, 21, 1161-1166.

Johannessen, K., Collins, C., Mills-Novoa, B. and Glider, P. (1999). A Practical Guide to Alcohol Abuse Prevention: A Campus Case Study in Implementing Social Norms and Environmental Management Approaches. Tucson, AZ: Campus Health Services, The University of Arizona.

LaBrie, J.W., Pedersen, E.R., Huchting, K., Thompson, A.D., and Hummer, J.F. (1998). Letter to the editor: Correcting misperceptions and reducing risky drinking through a student-designed poster campaign. Journal of Alcohol and Drug Education, 51, 7-16.

Lange, J.E. & Voas, R.B. (2001). Defining binge drinking quantities through resulting blood alcohol concentrations. Psychology of Addictive Behaviors, 75, 310-316.

Mattern, J. and Neighbors, C. (2003). Social norms campaigns: Examining the relationship between changes in perceived norms and changes in drinking levels. Journal of Studies on Alcohol, 65, 489-493.

Mitka, M. (2009). College binge drinking still on the rise. Journal of the American Medical Association, 302, 836-837.

Mussweiler, T. (2001). The durability of anchoring effects. European Journal of Social Psychology, 31, 431-442.

Mussweiler, T., Strack, F., and Pfeiffer, T. (2000). Overcoming the inevitable anchoring effect: Considering the opposite compensates for selective accessibility. Personality and Social Psychology Bulletin, 26, 1142-1150.

Perkins, H.W. (1985). Religious traditions, parents, and peers as determinants of alcohol and drug use among college students. Review of Religious Research, 27, 15-31.

Perkins, H.W. (1997). College student misperceptions of alcohol and other drug norms among peers: Exploring causes, consequences, and implications for prevention programs. In Designing alcohol and other drug prevention programs in higher education: Bringing theory into practice (pp. 177-206). The Higher Education Center for Alcohol and Other Drug Prevention, U.S. Department of Education

Perkins, H.W. and Berkowitz, A.D. (1986). Perceiving the community norms of alcohol use among students: Some research implications for campus alcohol education programming. International Journal of Addiction, 21, 961-976.

Posavac, E.J. (1993). College students' views of excessive drinking and the university role. Journal of Drug Education, 23, 237-245.

Rosenbaum, D. P., & Hanson, G. S. (1998). Assessing the effects of school-based drug education: A six-year multi-level analysis of project D.A.R.E. Journal of Research in Crime and Delinquency, 35, 381-412.

Rosenbaum, D. P., Flewelling, R. L., Bailey, S. L., Ringwalt, C. L., & Wilkinson, D. L. (1994). Cops in the classroom: A longitudinal evaluation of Drug Abuse Resistance Education (DARE). Journal of Research in Crime and Delinquency, 31, 3-31.

Schultz, P.W., Nolan, J.M., Cialdini, R.B., Goldstein, N.J. and Griskevicius, V. (2007). The constructive, deconstructive, and reconstructive power of social norms. Psychological Science, 18, 429-434.

Segrist, D.J. and Pettibone, J.C. (2009). Where's the bar? Perceptions of heavy and problem drinking among college students. Journal of Alcohol and Drug Education, 53, 35-53.

Strack, F., & Mussweiler, T. (1997). Explaining the enigmatic anchoring effect: Mechanisms of selective accessibility. Journal of Personality and Social Psychology, 73, 437-446.

Tobler, N.S. (1986). Meta-analysis of 143 adolescent drug prevention programs: Quantitative outcome results of program participants compared to a control or comparison group. Journal of Drug Issues, 16, 537-567.

Tversky, A. and Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124-1131.

Walsh, J.A. and Braithwaite, J. (2008). Self-reported alcohol consumption and sexual behavior in males and females: Using the unmatched-count technique to examine reporting practices of socially sensitive subjects in a sample of university students. Journal of Alcohol and Drug Education, 52, 49-72.

Wechsler, H. and Nelson, T.F. (2001). Binge drinking and the American college student: What's five drinks? Psychology of Addictive Behaviors, 5, 287-291.

Wilson, T. D., Houston, C., Etling, K. M., & Brekke, N. (1996). A new look at anchoring effects: Basic anchoring and its antecedents. Journal of Experimental Psychology: General, 4, 387-402.

Megan M. Lombardi and Jessica M. Choplin

DePaul University
Gale Copyright: Copyright 2010 Gale, Cengage Learning. All rights reserved.