Personality and Situational Factors as Predictors of Alcohol Use by College Students

(http://www.haverford.edu/psych/alcohol_study/alcoholstudy.00.summary.html)

Clara Gruen & Karen Hooker
Doug Davis, Advisor
Haverford College
Department of Psychology
May 2000

Introduction

Social activities are a significant aspect of students' college life. Alcohol has always been a part of campus social life, and decisions about whether and how much to drink are important parts of many students' college experience. In recent years, however, alcohol consumption at colleges has been a matter of national concern, as surveys across the country have revealed large percentages of binge drinkers and a significant relationship between drinking and student social problems.

This research is addressing alcohol use at Haverford College. It comes at a particularly interesting time for the school because it has experienced numerous alcohol poisonings and other problems resulting from high levels of alcohol use. In recent years the college as a whole has tried to address what might be possible causes for the increase in this type of behavior but only hypotheses have been developed to account for it. This study aimed to learn not only statistics about alcohol use in this community but also why these problems might be occurring.

Personality

Personality has not been looked at extensively in the previous research on alcohol use by college students. Certain subscales of personality have been examined as correlates of alcohol use. For example, Wood, Nagoshi, & Dennis (1992) looked at the Eysenck Personality subscales of impulsivity, sensation seeking, and empathy. This research extended the work on personality by using the Big Five Factor personality inventory, which measures the five dimensions of personality - Neuroticism or emotional stability, Extraversion, Openness, Agreeableness, and Conscientiousness. This is a widely used personality test, and therefore it would be valuable to be able to correlate alcohol use with such a widely used and accepted test.

The first of the Five Factors is Neuroticism, or emotional stability. It includes the traits of anxiety, anger and hostility, depression, self-consciousness, impulsiveness, and vulnerability (Costa & McCrae, 1992). The second factor, Extraversion includes traits such as "warmth, gregariousness, assertiveness, activity, excitement seeking, and positive emotions" (Costa & McCrae, 1992). A person who is high on the Extraversion scale would rate phrases such as, "I am the life of the party", or "I like to draw attention to myself" as "Very Accurate" as descriptions of themselves (Goldberg). A person who is low on the Extraversion scale, or more introverted, would answer "Very Accurate" in rating phrases such as, "I feel uncomfortable around people," or "I don't talk a lot" (Goldberg). Openness, sometimes referred to as Intellect, is the third factor in this Five Factor Model. Openness to Experience includes the facets "fantasy, aesthetics, feelings, actions, ideas, and values" (Costa & McCrae, 1992). Costa & McCrae (1992) define Openness as "aesthetically reactive, values intellectual matters, wide range of interests, rebellious and nonconforming." A person who is high on the Openness dimension would rate the phrases "I spend time reflecting on things," "I have a vivid imagination," or "I have excellent ideas" as "Very Accurate" descriptions of themselves (Goldberg). The fourth factor is Agreeableness. Agreeableness includes the traits "trust, straightforwardness, altruism, compliance, modesty, and tender mindedness" (Costa & McCrae, 1992). A person who is high on the Agreeableness dimension would rate "Very Accurate" phrases such as "I feel others' emotions" or "I have a soft heart" (Goldberg). The final factor in the Five Factor Model is Conscientiousness. Conscientiousness includes the personality traits of "confidence, order, dutifulness, achievement striving, self-discipline, and deliberation" (Costa & McCrae, 1992). A person who is high on the Conscientiousness scale will rate the following phrases as "Very Accurate": "I pay attention to details" and "I like order" (Goldberg). It was predicted that there would be a main effect for personality, such that a certain profile on the personality dimensions would predict drinking behavior.

Situational Factors

Past research has found that situational factors are predictors of students' alcohol use (Lo 1991, Nagoshi et al. 1994, Schall et al. 1991, Schall et al. 1992, Wood et al. 1992). Surveys have asked questions about reasons students give for choosing whether or not to consume alcohol, and situations in which they choose to drink. This study looked at both of these factors as predictors of alcohol use expecting to find a main effect. Based on past findings, there should be certain situational factors that make drinking, and drinking heavily, more likely - such as stress, a pre-party, or certain reasons for drinking.

Interaction

In addition to main effects for personality and situational factors, the researchers were interested in looking at the interaction between them. This study wanted to look at the questions: what combination of personality and situational factors leads to heavy drinking and binge drinking behavior, what causes a person with a particular personality to drink moderately in one situation and to binge in another, and why is it that given two people in the same situation, one will drink moderately and the other will binge.

Methods

Subjects

Subjects for this study were all Haverford College students. There were 97 participants: 44 males, 53 females. Twenty-nine of the subjects were first-year students, 33 were sophomores, 11 were juniors and 24 were seniors. The subjects' mean scores on the personality dimension scales were 32.5 (SD=8.1) for Extraversion, 41.6 (SD=4.7) for Agreeableness, 40.4 (SD=5.5) for Openness, 36.2 (SD=7.3) for Conscientiousness, and 30.1 (SD=8.6) for Emotional Stability on a scale of one to fifty. Subjects were compensated for their time by being entered in a lottery for cash and gift certificate prizes ranging from $15 gift certificates to $100 cash prizes. The prizes totaled $1500 and subject had better than one in two chance of winning a prize.

Materials

All information was collected on the web and all responses were kept confidential.

Initial Questionnaire: Subjects were given a questionnaire that is based in part on Nagoshi's alcohol survey and in part on the CORE survey. This questionnaire asked questions about frequency and quantity of alcohol consumption, consequences of drinking, reasons students gave for choosing to drink (i.e. because you friends are drinking, because you are bored) or not to drink (i.e. against personal values, or have more fun without drinking, and situations in which people drink (i.e. at a bar, in your room, pre-parties, campus parties). The questionnaire also asked about membership in school social groups and activities. Subjects also took a Big Five personality inventory. Click here to view the Initial Questionnaire.

Weekly Loggings: Subjects completed weekly questionnaires about their drinking behavior each week for 9 weeks while at school. The log included asking students how many drinks they consumed each night of the week, where they were, who they were with, and their reasons for choosing to drink or not to drink on any given day. This was done in order to get information that was as close to accurate as possible about drinking behavior as opposed to an average from the initial questionnaire. The weekly logging questionnaire also asked subjects to report their average stress level for the week on a scale of 1-10, where 1 is the least stressed the subject ever feels, 5 is average personal stress level, and 10 is the most stressed the subject ever feels. During the 9 weeks of logging, a period of relatively high stress before and during finals, a period of low stress during the beginning of second semester and a period of medium stress before spring break in second semester were anticipated. It was also expected that during this 9 week period there would be a variety of social situations in which students would decide whether to drink or not (i.e. fluctuations in amount of parties, big social events on campus before which pre-parties often occur). Click here to view the Weekly Logging Form.

Procedure

All data was collected on the web from the subjects' personal computers. There was no direct communication between the researchers and subjects; all communication was done through a third party using e-mail. Subjects signed up to participate in this study through a Web page. A general description of the study was given, the procedures to protect their confidentiality were explained, and subjects were told that signing up for the study implied informed consent. Subjects then received an identification number and the URL of the Initial Questionnaire through e-mail. The Initial Questionnaire, as described above, was completed. Then subjects reported their daily alcohol use through the Weekly Logging Questionnaires for nine weeks. They received an e-mail on each Monday reminding them to fill out the questionnaire for the previous Monday through Sunday week. The weekly logging took place during the last three weeks of Semester One and the first six weeks of Semester Two. This period also spanned across a generally high academic stress period (end of Semester One and Exam Period), a low academic stress period (start of Semester Two), and a middle academic stress period (middle weeks of Semester Two). At the completion of the study, subjects were thanked for their participation and prizes were awarded through the lottery, as described in the subjects section.

Results

Dividing Subjects

At the end of the study, subjects were divided into groups of abstainers, light drinkers, moderate drinkers, and heavy drinkers for the purpose of analyzing the results. This was done by multiplying their quantity of drinking (the number of drinks consumed throughout the entire study) by their frequency (the number of days on which they drank). This quantity x frequency variable was used to classify subjects because only using the total quantity variable or the total frequency variable hid the binge events. It did not distinguish between subjects who binged on occasions and subjects who perhaps drank more frequently in normal amounts.

Personality

Main effects for personality were found. The dimensions of Extraversion, Agreeableness and Openness were found to both correlate with and predict alcohol use. Extraversion and Openness were found to positively correlate with the quantity of alcohol consumed and binge frequency. Agreeableness was negatively correlated with alcohol use. Multiple regression analysis demonstrated that alcohol consumption and binging behavior could be predicted using all five personality dimensions with Extraversion, Openness and Agreeableness being the significant contributors. Conscientiousness and Emotional Stability were not found to be significant correlates with alcohol use. (See table 1).

Table 1. Personality factors correlated with quantity x frequency and binge frequency.

 

Extraversion

Agreeableness

Conscientiousness

Stability

Openness

Q x F

r=.25*

r=-.26*

r=.00

r=.11

r=.24*

Binge

r=.28*

r=-.21*

r=-.03

r=.16

r=.23*

*p<.05 (n=75)

Situations

Drinking Situations: The top six reported situations for drinking were small party (3-10 people), room with friends, pre-party, campus party, bar or club, and restaurant. These situations were analyzed using a ratio of total number of times the subject drank in the situation over the total number of times the subject reported being in that situation with alcohol present, whether or not they drank. This was done because just looking at the total number of times the subject drank in the situation contained an artifact, because heavy drinkers in general drank more frequently, so they had more opportunity to report the situations they were in than light drinkers. An ANOVA was done comparing the ratios of drinking in the situation to being in the situation comparing heavy, moderate, and light drinkers. As can be seen by the differences in the ratios, there were many significant differences between the groups in situations. Heavy drinkers are more likely to drink given that they are in any of these situations. However, in some of the situations, light and moderate drinkers were also very likely to drink. (See Table 2) The following pie charts show the percentage of drinking events in each of the top six situations for each of the groups of drinkers.


Table 2. Ratios of drinking in a situation to being in that situation with alcohol present.

 

Light Drinkers

n=25

Moderate Drinkers

n=25

Heavy Drinkers

n=25

Small party

.33

.67

.83

Room with friends

.57

.77

.86

Pre-party

.49

.86

.95

Party

.30

.84

.95

Bar or club

.59

.74

.88

Restaurant

.34

.31

.71

Binge Situations: The top six reported situations for binge drinking were small party, pre-party, room with friends, bar or club, and drinking game. The order of the top situations changed from top six situations for drinking, and also drinking game is introduced as a top situation for binging, while restaurant is no longer one of the top situations. As in the situations, the ratios showed many significant differences between the likelihood of bingeing in the situation given that the subject was in the situation. Heavy drinkers were more likely to binge given that they were in any of the situations. However, there were fewer light drinkers who binged at all, whereas most of the moderate drinkers and all of the heavy drinkers binged. (See Table 3)


Table 3. Ratios of binge drinking in a situation to being in that situation with alcohol present.

 

Light Drinkers

n=7

Moderate Drinkers

n=19

Heavy Drinkers

n=25

Small party

.02

.18

.50

Pre-party

.17

.39

.72

Party

.22

.27

.66

Room with friends

.11

.11

.36

Bar or club

.00

.26

.45

Drinking Game

.00

.56

.70

Reasons

The top six reasons: to have fun, to relax, peers were drinking, for physical effects, to celebrate, and to get drunk. Reasons were found to play a role in how subjects drank. An overall analysis of variance found significant differences between the reasons given by heavy moderate, and light drinkers. One striking difference was that heavy drinkers used "to get drunk" as a reason for drinking far more than any other group. Also, heavy drinkers reported "to celebrate" as a reason for drinking less frequently in comparison to their other reasons than did moderate and light drinkers. (See Table 4)


Table 4. Number of times the top six reasons were reported by each drinking level.

 

Light Drinkers

n=25

Moderate Drinkers

n=25

Heavy Drinkers

n=25

To have fun

45

138

322

To relax

41

141

235

Peers were drinking

60

126

215

For physical effects

29

67

177

To celebrate

35

73

123

To get drunk

10

33

163

Interaction

An overall ANOVA and multiple regression found no interaction between personality and situation or reasons. However, this analysis was limited because the questionnaire did not provide us with adequate measures of subjects most important situations or reasons.

Stress

Significant differences in subjects' stress ratings were found in t-tests comparing the weeks. As compared to quantity of alcohol consumed, as stress went up, quantity went down. This can be seen especially in looking at the quantity means of alcohol consumed in weeks 2 and 3, which were periods of high stress at the end of first semester, then in week 4 which was the first week of second semester when stress ratings were very low, as compared to the following weeks where the work increased and stress ratings went up. This was opposite of what was predicted and what is commonly believed at Haverford. When subjects are reporting higher levels of personal stress, they drink less. (See Table 5)

Table 5. Means of stress rating and quantity of alcohol consumed by week.

 

Stress

Quantity

Week 1

5.4

5.2

Week 2

7.1

3.5

Week 3

7.2

6.0

Week 4

3.9

7.5

Week 5

5.3

6.8

Week 6

5.4

3.8

Week 7

6.2

5.0

Week 8

5.8

5.0

Week 9

6.5

4.7

Reasons for Abstaining

The top reasons for abstaining reported in the initial questionnaire were not liking the physical effects and for personal values, and the lowest reason was religious beliefs. In general, personal beliefs and values had more influence on abstaining than external forces.

Gender

The results showed some striking results for gender and alcohol use. There was a ratio of over 2 to 1 of males drinking to females drinking. Males had an average of 53.7 drinks throughout the entire study, while the means for females was 21.8 drinks. Males also had 566 (61.6%) of the 919 total drinking events in the study. Females had 353 events (38.4%). The males were responsible for over 2/3 of the binge events. Also, of the heavy drinkers, 76% were male. (See Table 6)

Table 6. Mean quantity, total binge events, and total drinking events for male and female drinkers.

 

Mean Quantity

Binge Events

Drinking Events

Males

(n=35)

53.7

231

566

Females

(n=40)

21.8

93

353

Class

The results showed differences between classes and their distribution among drinking levels. For example, of the 24 of the seniors in the study, there were twelve heavy drinkers, only two moderate drinkers, five light drinkers, and five abstainers. Half of the seniors studied were grouped as heavy drinkers and half of the heavy drinkers in the study were seniors. It was also interesting to note that the freshmen were fairly evenly spread throughout the four groups, with eight heavy drinkers, eight moderate drinkers, eight light drinkers, and five abstainers.

Table 7. Distribution of subjects by class among drinking levels.

 

Heavy Drinkers

Moderate Drinkers

Light Drinkers

Abstainers

Seniors

12

2

5

5

Juniors

3

4

3

1

Sophomores

2

11

9

11

Freshmen

8

8

8

5

Discussion

Limitations of the Sample

The results of this study were somewhat limited because of certain limitations of the sample. First, the personality means of the sample were higher than average. For example, the sample mean for Agreeableness was 41.6 and the mean for Openness was 40.4. The lowest mean on the personality dimensions was 30.1 for Emotional Stability. These are all high compared to the population mean of 25 on all scales. Another limitation is that Haverford is a fairly homogeneous sample of students and therefore these results may not generalize well to the larger population of college students. Also, drinking behavior may not demonstrate the full range of drinking that occurs at most colleges. The CORE survey revealed that Haverford College is very similar to other colleges in the percentage of people who drink at the school however, we have no knowledge of how Haverford compares in the range of quantity of alcohol that students consume.

Future Research

This study gives many implications for future research. First, considering our striking results with differences between male and female drinking patterns gender should be looked at further. Second, strong results were found demonstrating an inverse relationship between stress level and drinking, therefore this finding should be confirmed in future research as well as exploring a possible relationship between stress and binge drinking. A mixed result in the binge-stress relationship was found such that binging events increased as stress increased. However this finding only held for a short time in the study and otherwise followed the same pattern as stress and drinking. Third, because of the way the questionnaire was written, it was difficult to look at how subjects treated individual situations and reasons on any given day. Subjects were asked to report all situations and reason for drinking each day but did not ask which of these was the most important. On most days subjects gave multiple responses and therefore it was not possible to ascertain the principle motivators in their drinking. Finally, personality should be looked at in future research. Agreeableness was found to be an important factor in predicting drinking behavior. This had been previously unexplored in past research. Therefore in future research this dimension should be investigated further. Past research had found mixed results on the dimension of Emotional Stability. This study found this factor to not be significant, however this may be because we did not have a full range of drinkers in the sample. Vaillant (1983) had stated that Neuroticism was associated with alcoholic drinking patterns. Since we were working with a non-clinical sample of drinkers it follows that we would not find this to be a significant predictor of drinking behavior. It may also be because we didn’t have a normal range of scores on Emotional Stability in our sample. Because this was the first look at alcohol use using the complete Five-Factor model, and because past research focused more on situation as the predictor instead of personality, future research is necessary to confirm these findings.

Conclusions

This research found main effects for personality and situational factors as predictors of alcohol use by college students. This study is particularly valuable with its addition of a complete personality measure to the study of alcohol use by college students. It also used a more in depth look at situations than some previous studies. Its significant findings for the different personality scales and for situations and reasons is an important addition to the previous research on this topic. Future research should be done to explore some of the factors not concentrated on in this study, such as gender, to find a better measure of situations and reasons to predict alcohol use, and to compare this sample to other samples in terms of personality and alcohol use.

References

Costa, P. T., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual Differences, 13, 653-665.

Goldberg, L. R. International Personality Item Pool. Online. Internet. http://ipip.ori.org/ipip/index.htm

Lo, C. (1991), Psychosocial correlates of problem drinking behavior and multiple drug using behavior. College Student Journal, 25(2), 141-148.

Nagoshi, C. T., Wood, M. D., Cote, C. C., & Abbit, S. M. (1994). College drinking game participation within the context of other predictors of alcohol use and problems. Psychology of Addictive Behaviors, (8)4, 203-213.

Schall, M., Weede, T. J., & Maltzman, I. (1991). Predictors of alcohol consumption by university students. Journal of Alcohol and Drug Education, 37(1), 72-80.

Schall, M., Kemeny, A., & Maltzman, I. (1992). Factors associated with alcohol use in university students. Journal of Studies on Alcohol, 53(2), 122-136.

Vaillant, G. E. (1983). The Natural History of Alcoholism: Causes, Patterns, and Paths to Recovery. Cambridge, Massachusetts: Harvard U Press.

Wood, M. D., Nagoshi, C. T. , & Dennis, D. A. (1992). Alcohol norms and expectations as predictors of alcohol use and problems in a college student sample. American Journal of Drug and Alcohol Abuse, 18(4), 461-476.

___________________________________________________________________________________________
This document is a draft report of a study of alcohol consumption at Haverford College under the sponsorship of the Department of Psychology and the Office of the Dean of the College. Do not reprint, distribute, or link to this report without permission. For further information, or for a full copy of the thesis report, e-mail Doug Davis at ddavis@haverford.edu.