Hazardous Alcohol Use among University Students - Critical Appraisal and Data Analysis

This article provides a critical appraisal of Davoren et al. (2015) paper on hazardous alcohol use among university students and data analysis on alcohol use among university students. It includes descriptive statistics, regression analysis, and significant predictors of hazardous alcohol use.

Running head: HAZARDOUS ALCOHOL USE
Hazardous Alcohol Use among University Students
Name
Course Number
Date
Faculty Name

HAZARDOUS ALCOHOL USE 2
Introduction to Biostatistics: Hazardous Alcohol Use among University Students
Question 1: Critical Appraisal of Davoren et al. (2015) paper on Hazardous alcohol use among
university students
Introduction
Following the increasing rates of alcohol consumption among university students, Davoren et
al. (2015) did a cross-sectional study in the University College Cork, Ireland to investigate the
correlates and prevalence of hazardous alcohol use among university students.
Sample Size
The study followed a classroom data collection strategy were the sample was selected using the
proportional to size sampling technique. Only students who were on degree programmes were eligible
for inclusion. The inclusion criteria included students who were in undergraduate degree programmes.
Those who did not meet the inclusion criteria were not included in the final study size.
Statistical Methods
Stratified probability proportional to size sampling technique was used. Frequency and
descriptive statistical analysis techniques were used to describe the data. Further, univariate and
multivariate logistic analyses were conducted to investigate the association of the covariates with the
hazardous alcohol drinking. Due to the unbalanced gender representation in the study, stratified
analysis was used to minimize error. Also, there were cases of over and under-sampling for the first
and fourth years respectively. Therefore, year in college was used as a covariate to check whether it
was statistically associated with hazardous alcohol use. Age was also used as a confounder in the
multivariate analysis.
Results
Using a precision level of 1.5%, the prevalence of hazardous alcohol use of 0.73 and a total
population of 12,475 students meeting inclusion criteria, a sample of 2,686 participants was calculated.
Of the 2,686 students selected to participate in the study, 2,332 (94.3%) agreed to participate and 57 of

HAZARDOUS ALCOHOL USE 3
them were identified as post-graduate students. Therefore, the final sample for the study was attained
at 2,275 students.
The study included 63.1% female participants against 36.9% males. Of those under 18 years,
the prevalence of hazardous alcohol drinking was 67.9% and 74.2% for males and females
respectively. Women under 18 years (74.2% compared with 67.9%), 20 years (68.6% compared with
66.9%) and those above 22 years (60.9% compared with 53.5%) had higher prevalence of hazardous
alcohol use. Generally, female students had higher prevalence of hazardous alcohol use for the schools
of science (65.1% against 63.6%), arts (63.4% against 59.5%), law (79.4% against 77.5%) and
medicine (68.4% against 61.3%). Hazardous alcohol use among first years does not seem to be
significantly difference among male (65%) and females (65.6%). Similar, fourth years have almost
equal levels of hazardous drinking (72.2% males and 75.6% females). Comparing first and fourth
years with second and third-year students, the latter had a larger difference in the levels of hazardous
drinking among males and females. School of Law and business was significantly associated with
hazardous alcohol use compared to those in school of science/engineering/food science. Having a
sexual partner was significantly associated to hazardous alcohol compared to those without. Adjusting
for age, smoking was a significant covariate for male students. However, females smoking status was
significantly associated with hazardous alcohol use, even without controlling for age. Illicit drug users
had higher odds of being hazardous alcohol users. Controlling for age, third-year students had higher
odds of engaging in hazardous alcohol use compared to the first year males.
2 686 2 332
94.3% agreed to
participate
57
Were post-graduates
2 275
Final Sample

HAZARDOUS ALCOHOL USE 4
Question 2: Alcohol use among university students data analysis
Descriptive statistics
Table 1: Summary statistics of age, alcohol use and Warwick Edinburgh Mental Wellbeing Scale
(WEMWBS)
Age Alcohol Use WEMWBS
Minimum 17 0.5 36.2
1st Quarter 19 1.0 41.4
Median 20 5 43.4
Mean 19.69 7.66 44.03
3rd Quarter 21 10 46.40
Maximum 24 53 53.2
Group Age Alcohol Use WEMWBS
Male 19.84 9.23 44.22
Female 19.49 5.43 43.75

HAZARDOUS ALCOHOL USE 5
Figure 1: Gender distribution
Figure 2: Drug Use

HAZARDOUS ALCOHOL USE 6
Figure 3: Bar graph of distribution of study participants by school
58.6% of the study participants were males. 24.9% of the students in the study were in the
school of engineering, 28.2% from the school of medicine and health and 27.5% from the law and
business school. 61.9% of the study participants were not drug users.
Figure 4: WEMWBS by log alcohol use scatter plot

HAZARDOUS ALCOHOL USE 7
According to the descriptive statistics, males take higher amounts of alcohol on average
compared to their female counterparts. Also, the wellbeing scale differs between males and females
with an average of 44.22 and 43.75 respectively. This is an indication that gender might be a
significant confounder in predicting alcohol use. According to figure 4 above, the values of log of
alcohol use seems to be reducing as the mental well-being scores increases. Therefore, the mental
wellbeing score can be a significant predictor on the amounts of alcohol taken among students.
According to the scatter plot, males and females are equally distributed. However, the descriptive
statistics shows males and females have differing levels of alcohol intakes and their mental well-being
scores are also different. Therefore, since sex affects both, the predictor and the response variable, it
can be a significant confounder(Wilsnack et al., 2009).
Regression
Table 2: Model coefficients
Estimate Standard error T value p-value
Intercept 3.134 0.405 7.748 <0.001
WEMWBS -0.055 0.009 -6.041 <0.001
sexfemale -0.37 0.066 5.647 <0.001
The model is statistically significant with an F statistic of 32.12 associated with a p-value <
0.001. Therefore, sex and Warwick Edinburgh mental well-being scale significantly predict the
amount of alcohol taken by the university students. Both the mental well-being scale and sex explain
19.22% of the variation in amounts of alcohol students take. The WEMWBS score is a statistically
significant predictor in the model. Increasing the score of the mental wellbeing score is associated with
lower amounts of alcohol takes in a week by 0.055 after controlling for sex of the student. Sex is a
significant confounder with a p-value <0.001 – females intake lower levels of alcohol compared to
male university students. In conclusion, the prevalence of alcohol use is higher in male university
students compared to females and WEMWBS is a significant correlate of hazardous alcohol use.

HAZARDOUS ALCOHOL USE 8
References
Davoren, M. P. et al. (2015) ‘Hazardous alcohol consumption among university students in Ireland: A
cross-sectional study’, BMJ Open, 5(1). doi: 10.1136/bmjopen-2014-006045.
Wilsnack, R. W. et al. (2009) ‘Gender and alcohol consumption: patterns from the multinational
GENACIS project.’, Addiction (Abingdon, England). NIH Public Access, 104(9), pp. 1487–500. doi:
10.1111/j.1360-0443.2009.02696.x.