Sponsored by Missouri Western State University Sponsored by a grant from the National Science Foundation DUE-97-51113
Home |
The proper APA Style reference for this manuscript is:
JARGO, K. E. (2005). Changing Drinking Behaviors in College Students: an Application of the Theory of Planned Behavior. National Undergraduate Research Clearinghouse, 8. Available online at http://www.webclearinghouse.net/volume/. Retrieved December 9, 2023 .

Changing Drinking Behaviors in College Students: an Application of the Theory of Planned Behavior

Sponsored by: CARRIE FRIED (cfried@winona.edu)
AbstractMany different research projects have tried to explain and alter the drinking habits of college students. This research project, which was a 2x2 factorial design, attempted to use the principles behind the Theory of Planned Behavior to decrease the drinking quantity and frequency of college students living in on-campus residencies. A general drinking survey was administered to all participants and then two manipulations were introduced to different groups of the participants. The manipulations were self-efficacy, where poster with positive slogans and thought inducing worksheets were used, and normative influence, where posters were displayed stating desirable drinking norms for college students. Four weeks later a second survey was administered. There was a statistically significant interaction found between the two manipulations when considering the amount of drinking, F(1,40) = 2.95, p = .09, but no other statistically significant results were found. The means indicated that the self-efficacy manipulation did change the drinking behavior.

Changing Drinking Behaviors Through Social Norms and Self-Efficacy Almost every college or university will say that their campus has a drinking problem. The community may have more police officers out during the evening hours in order to discourage students who are out drinking. Residence Halls try to implement programs that teach you safe ways to drink alcohol and also how to recognize alcohol related problems. Some halls try to provide activities so that students have something to do, other than drinking. In all of these situations the campus and sometimes the students are trying to reduce the behavior of drinking. One theory, called the theory of planned behavior, was developed in attempts to break down how we choose our behavior. This theory was a revision of the theory of reasoned action. The Theory of Planned Behavior lists three factors that people use in order to engage in a certain behavior. These factors are their attitude towards the behavior, the relevant subjective norms and what their perceived behavioral control is (Ajzen, 1991). These factors show that people rationally think about the consequences of their behavior and that behavior is intended. This perspective hypothesizes that behavioral intentions are what influence our decisions, not attitudes. Attitudes are only influential on behavior because they influence intentions. Attitudes towards behavior can be broken down into two determinants (Ajzen, 1991). The first is one’s beliefs about the consequences of performing the particular behavior. When deciding to eat an ice cream cone, for example, you might consider how many calories you’ll be taking on and how much weight you might gain from it. These consequences may be logical or illogical, but you perceive them as important. The second determinant is one’s evaluation of those possible consequences. If you were really worried about those extra calories than your evaluation of the consequences will lead you to not eat the ice cream. If you think the consequences aren’t that severe, chances are you’ll eat the ice cream. One study by Johnson and Hall (2004) looked at the application of the Theory of Planned Behavior on the prediction of safe lifting behavior for manufacturing companies. The researchers gave a survey that contained questions that were designed to measure intent, attitude, past behavior, perceived behavioral control and subjective norms. This survey was given to materials management employees at a heavy manufacturing organization. Johnson and Hall found that attitudes did not directly contribute to the prediction of safe-lifting behavior, but the remaining constructs did. They explain that their results don’t mean attitudes are unimportant, but they might be less important then the other factors. The second factor in the theory of planned behavior, the relevant subjective norms, involves how you think others would respond to your judgment. This factor also involves two parts, the perceived expectations of significant others and one’s motivation to conform to those expectations. Those closest to you will have the most influence on your decisions about behaviors. How concerned you are in general with the approval of your loved one also plays a part in your decision. If you have no need for approval from your mother about whether ice cream is good or bad for you, chances are you won’t really weigh in her opinion. When considering subjective norms the setting of the behavior matters. Those in a self-awareness setting tend to weigh a little more on personal attitudes towards the behavior. Those in a public-awareness setting tend to be more influenced by the subjective norms. Perceived control, which is the last part to the theory of planned behavior, is one’s perception of how easy or difficult it is to perform the behavior. You might consider self-efficacy in this factor. Ajken (1991) states that, “when people believe they have little control over performing a behavior because of lack of ability or resources, then their behavioral intentions will be low regardless of their attitudes or subjective norms.” Perceived control seems to have the biggest impact on our decisions about behavior. Self-efficacy is frequently tied to perceived control. Perceived self-efficacy is concerned with judgments of how well one can execute courses of actions to deal with the situation at hand (Azjen, 1991). In other words, how confident we are in ourselves that we can actually perform the action (perceived self-efficacy) has a great impact on the behavior itself. Dale Schunk (1991) looks at how one develops and maintains their level of self-efficacy. The main way individuals acquire information to appraise efficacy is from their performances. Failing at a task will lower one’s self-efficacy and succeeding at a task will raise it. Also, assessments are made through information from others. Students often receive persuasory information that they possess the capabilities to perform a task. For example, one might shout, “You can do it!” to the players of a soccer game. When discussing norms one must differentiate between descriptive norms and injunctive norms. Descriptive norms would pertain to the actual behavior, for example, the behavior of drinking alcoholic beverages. An injunctive norm would be the approval of the drinking behavior (Borsari & Borsari, 2003). Borsari and Borsari conducted a meta-analysis on the influences of injunctive norms and normative norms in relation to drinking behaviors in college students and self-other discrepancies, which are the differences in beliefs between one’s self and the judgment of others. The meta-analysis integrated 23 studies evaluating the influence of five predictors: norm type, gender, reference group, question specificity and campus size. He found that a greater self-other discrepancy was shown by injunctive norms and that injunctive norms had a greater impact on drinking behavior.Larimer and Turner (2004) looked at the role of descriptive and injunctive norms when predicting drinking behavior and alcohol-related problems among members of a fraternity or sorority. Perception of alcohol consumption in the pledge classes was the descriptive norm and acceptability of drinking was their injunctive norms. Larimer and Turner found that injunctive norms assume greater importance as an added risk factor for current and future alcohol-related problems. Jody Mattern and Clayton Neighbors (2004) looked at the relationship between changes in perceived norms and changes in drinking levels among college students. The researchers used different campaign methods, such as prominently displayed posters, a table tent and postcards, in residential halls to convey specific social norms about drinking. Mattern and Neighbors had participants fill out a quantity and frequency of drinking survey. They also asked participants to estimate how often the average college student drank and how much they drank. Their results found that by reducing what is considered the normal amount and frequency of drinking, drinking frequency among those tested was reduced. Previous research show us that injunctive norms, or what participants think is actually happening, are really important to consider when looking at manipulating behavior. Also, when trying to change behavior using a few different approaches at once can produce more of a change. For the research reported in this article I looked at injunctive norms and self-efficacy in relation to the frequency and amount of drinking among college students living in the residential halls. I attempted to see if there was an interaction between the injunctive norms and self-efficacy and if there were any main effects. Since previous research has shown that injunctive norms produce more of a change, I predicted that the injunctive norms manipulation would have a greater influence on reducing drinking frequency and quantity than self-efficacy.

MethodParticipants The sample for this experiment consisted of 215 participants of which 116 were female and 99 were male. Ages ranged from 18 to 23 with the average being 19. Participants were undergraduate students living in two residence halls on campus. Year in school for the participants ranged from 1st year to 5th year with 91% of them being first year students. Participants were later grouped into drinking (67) and non-drinking (148) participants based on their responses to the first survey. MaterialsDrinking Measures. For all groups a simple questionnaire was used to determine how often participants drank alcoholic beverages and how many they had in one sitting. Questions asked how many times a week, on average, the participant had 1-3 drinks in one sitting and also how many times per week, on average, the participant drank on campus or off campus. The survey stated what is considered one alcoholic drink. Participants checked a box of none, once, twice or three or more times for their answers. The questionnaire also asked questions pertaining to behaviors while and after drinking. Participants answered questions like, “I often drink until I pass out” with their responses based on a Lykert’s scale with the scale going from strongly disagree to strongly agree. Self-efficacy Worksheets and Scales. The worksheet given in the self-efficacy manipulation explained a hypothetical situation where a student had to make a decision between going out with friends to drink or staying home to work on a paper. Participants were asked to give examples of how they personally control their drinking and slogans that might promote personal responsibility. The self-efficacy test used was the Drinking-Related Locus of Control Scale developed by Donovan and O’Leary (1978). Only twelve questions were used from the scale. Design and Procedure General Design. This was a two by two factorial manipulation. The groups in the self-efficacy manipulation received the basic frequency survey, the worksheet and the locus of control survey. On the initial frequency survey participants were asked to pick a code, numbers or letter, that they would remember, and write it down on their survey sheets in order to match it with the follow-up survey. The participants were put into four groups based on which floor of the residence hall they lived on. Each group comprised of one floor of male participants and one floor of female participants. The first surveys and worksheets were given to all participants at their floor meetings. The first floors received the basic drinking frequency survey and were the control group. They did not receive any manipulation. The second floors received the self-efficacy manipulation, the third floor received the normative influence manipulation and the fourth floor received both the self-efficacy manipulation and the normative influence manipulation.All floors received the frequency of drinking survey at their floor meeting, which was run in cooperation with their Residence Assistants (RA). Consent forms were colleted for all participants. Four weeks after they were again asked, by their RAs to fill out the follow-up frequency survey.Normative Influence: Following the initial meeting, the five social norm posters were placed around the two floors that were receiving the normative influence. These posters were intended to reduce the perceived normal amount and frequency of drinking for the average college student.Self-Efficacy. At the initial meeting, those floors involved in the self-efficacy manipulation received a self-efficacy survey and the worksheet. This worksheet was a way for participants to think of realistic ways in which they could control their behavior when drinking. It also asked for slogans that might trigger an “in control” mood. One week after the initial meeting, participants were given their worksheets back with positive comments on them, such as, “good suggestion” and “this could really work.” The comments let participants know that their ideas for invoking control were high quality and realistic for a typical college students. The comments were also intended to remind the participants that control over drinking is possible. The self-efficacy drinking survey was not returned to the participants.

ResultsScales. A single measure of the amount of drinking was created by combining 5 items of drinking behavior, a = .931. Higher scores on this 4 point scale indicate more drinking. A second measure was created of the negative consequences that result from drinking, a = .792. On this 5 point scale, higher scores indicate more negative consequences due to drinking. Also, for subjects whose pre-test and post-test could be matched together, which was 44 subjects, changed scores were formed for both the amount and consequence scales. In both cases post-test scores were subtracted from pre-test scores, so higher, positive numbers indicate and increase in drinking behavior.Changed Scores of Amount. A 2 (self-efficacy) x 2 (normative influence) factorial ANOVA was conducted on the changed scores for amount. There was a marginally significant interaction F(1,40) = 2.95, p = .09. There was also a statistically significant main effect for normative influence, F(1,40) = 5.18, p = .03, however this appears to be an artifact of the interaction. There was no statistically significant main effect for self-efficacy, F(1,40) = 1.42, p = .241. The means for amount changed scores, which can be seen in Table 1, indicate that there was a change in behavior for the group that had no normative influence manipulation, but that did have the self-efficacy manipulation. All other groups show no significant change in behavior.Post-Test Only Amount Scores. A 2x2 factorial ANOVA was conducted on amount scores for all post-test surveys that were returned, regardless if they could be matched to a pre-test. This resulted in a much bigger sample than the one used for the changed scores. No statistically significant interaction was found, F(1,113) = .1, p = .747. Statistically significant main effects were not found for self-efficacy, F(1,113) = 1.46, p = .23, nor were they found for normative influence, F(1,113) = .003, p = .955. The means, listed in Table 2, again show more change in behavior for the group that received the self-efficacy manipulation only. Changed Scores of Consequences. A 2 (self-efficacy) x 2 (normative influence) factorial ANOVA was conducted on the changed scores for consequence. There was no statistically significant interaction, F(1,40) = .77, p = .39. Also, there was not a statistically significant main effect for self-efficacy, F(1,40) = 2.45, p = .125, nor for normative influence, F(1,40) = .11, p = .75. Post-Test Only Consequence Scores. Again, this was a 2 (self-efficacy) x 2 (normative influence) factorial ANOVA conducted on all of the surveys returned, so this was a much bigger sample that the consequence changed scores. There was no statistically significant interaction, F(1, 113) = 1.08, p = .30. There were no statistically significant main effects for self-efficacy, F(1,113) = 1.07, p = .30, or for normative influence, F(1,113) = .07, p = .80.

DiscussionIt was predicted that the normative influence manipulation would have a greater influence on reducing drinking frequency and quantity than the self-efficacy manipulation. The results showed that the self-efficacy manipulation changed the behavior of the participants, not the normative influence manipulation and that there was an interaction between the two manipulations. When looking at the means for the pre-test, post-test changed scores for amount and the post-test only amount the self-efficacy only group was the one group that seemed to show a significant change in behavior. The self-efficacy manipulation on its own had an impact on the behavior of the participants, but when it was combined with the normative influence manipulation the effect was virtually non-existent. The posters and slogans used to stimulate new thought about what is normal for college drinking behavior may have been so extreme that the messages were disregarded. The participants may have thought there were so outlandish that the slogans were a hoax. The main problem with the data for this experiment was the lack of post-surveys that could be matched with the pre-surveys. Without a substantial data set, statistically significant results could not be formed. Some participants simply did not fill out the follow-up survey, but most either didn’t remember their code or remembered the wrong one. The code was some set of numbers or letters that the participants picked to put on their initial survey and would remember as significant so that they could put it on their follow-up survey. In the future, if someone were to attempt a similar study, a better system of matching up the initial surveys with the follow-up surveys would be beneficial. Perhaps the researcher could use their full names, but take all precautions to keep them confidential. Maybe codes could be kept with the names but only a separate party that doesn’t see the actual survey results would keep track of this list. Another improvement would involve the normative influence manipulation. Research should be done about actual drinking habits of the population being studied and even the boundaries of what is believable information about drinking habits. It may also be beneficial for a stronger result if the study was conducted at the start of the school year instead of part way through the second semester. At this time you might catch first year students who have not experienced the college life yet and are absolutely full of misconceptions about college and drinking.



ReferencesAjzen, I (1991). The Theory of Planned Behavior. Organizational Behavior and HumanDecision Processes, 50;179-211Borsari, B. & Carey, Kate B. (2003). Descriptive and injunctive norms in collegedrinking: A meta-analytic integration. Journal of Studies on Alcohol, 64(3); 331-341.Donovan, D.M. & O’Leary, M.R. (1978). The Drinking –Related Locus of ControlScale. Journal of Studies on Alcohol, 39, 759-784.Johnson, S & Hall, A (2004). The prediction of safe lifting behavior: An application ofthe theory of planned behavior. Journal of Safety Research, 36(1); 63-73.Larimer, M.E. & Turner, A.P. (2004). Predicting Drinking Behavior and Alsohol-Related Problems Among Fraternity and Sorority Members: Examining the Role of Descriptive and Injunctive Norms. Psychology of Addictive Behavior, 18(3); 203-212.Mattern, J. & Neighbors, C. (2004). Social norms campaigns: Examining therelationship between changes in perceived norms and changes in drinking levels. Journal of Studies on Alcohol, 65(4); 489-493.Schunk. D (1991). Self-Efficacy and Academic Motivation. Educational Psychologist,26(3); 207-231Sprott, D., Smith, R.J., Spangenberg, E. & Freson, T. (2004). Specificity of PredictionRequests: Evidence for the Differential Effects of Self-Prophecy on Commitment to a Health Assessment. Journal of Applied Social Psychology, 34(6); 1176-1190.

Table 1

Submitted 6/14/2005 11:09:16 PM
Last Edited 6/14/2005 11:13:37 PM
Converted to New Site 03/09/2009

Rated by 0 users. Users who logon can rate manuscripts and write reviews.

© 2023 National Undergraduate Research Clearinghouse. All rights reserved. The National Undergraduate Research Clearinghouse is not responsible for the content posted on this site. If you discover material that violates copyright law, please notify the administrator. This site receives money through the Google AdSense program when users are directed to useful commercial sites. We do not encourage or condone clicking on the displayed ads unless you have a legitimate interest in the advertisement.