Statistical Study of Alcoholism Among Students

Topics: Alcohol, Alcohol Abuse, Alcoholism, Students Words: 1398 Pages: 5

Introduction: Research Question

This research paper investigates the relationship between workday alcohol consumption and several characteristics of students’ social, economic, and academic status. In particular, a large set of gender, demographic, and family data allows several research questions to form.

Does the number of absences from school affect workday alcohol use?
Does a final semester grade (in math) affect workday alcohol use?
Is there an increase in the combined effect of school absences and final semester grades on workday alcohol use?
Is there a relationship between the frequency of meetings with friends and alcohol consumption during the workday?
Is there a relationship between age and alcohol consumption during the workday?

The Significance of this Question

It is not difficult to conclude that alcohol is a severe problem not only for the college community but for humanity as a whole. As a detrimental factor, alcohol mainly affects the physical and emotional health of the individual. The choice of the student community to study the connections described above is not accidental: in youth, alcohol is often synonymous with fun and partying. Large amounts of drinking on a regular basis are often perceived as a natural part of a student’s everyday life, but this behavior has destructive consequences. As a young student, they develop chronic alcoholism, the effects of which will tell on their well-being later in life — in addition, drinking while studying leads to a drop in academic performance and a decline in the student’s social skills (Flagel, 2021). In this regard, examining some of the patterns seems like a quite meaningful strategy and has academic relevance.

Hypotheses

The more often a student misses schoolwork, the more likely he or she is to consume alcohol during the workday.
The higher the student’s final grade for the semester, the lower the student’s propensity to drink alcohol during the workday.
Skipping classes and a higher final grade have no apparent change from the individual strength of each variable.
The more often a student goes out with friends, the more likely they are to drink alcohol during the workday.
There is no significant relationship between age and alcohol consumption.

Data Description

In this paper, the student-mat data were obtained from publicly available sources. It is data collected from a large-scale survey of students who are enrolled in a high school mathematics course. The overall survey results are 33 variables, each describing a respondent’s social, economic, behavioral, or academic status. The selected data are the following set of variables:

Dalc – workday alcohol consumption (numeric: from 1 – extremely low to 5 – extremely high).
absences – number of school absences (numeric: from 0 to 93).
G3 – final grade (numeric: from 0 to 20, output target).
age – student’s age (numeric: from 15 to 22).
goout – going out with friends (numeric: from 1 – exceptionally low to 5 – exceedingly high).

Thus, this paper uses five variables, of which Dalc is the dependent variable for all measurements, and the other variables are treated as influential parameters.

Justification for the Method Used

All five variables were numeric and thus represented a value from a predetermined interval. Additionally, they were discrete variables. At the same time, an essential predictor for choosing a specific statistical testing methodology was the nature of the relationship of interest: the effect of one — or more — variables on the amount of alcohol consumed. Given the theoretical considerations learned in this course, the best solution for such a data set would be to use correlation analysis and linear regression. Recall, correlation determines the strength and direction of the relationship between two variables (Vedantu, 2020). In contrast, regression describes how a change in one variable affects a change in another variable. Regression and correlation are not interchangeable tests but using them together will produce exciting results for this paper. In addition, if the regression fits the data well (as determined by the coefficient of determination R²), it becomes possible to use the equation to predict results for a particular measurement. A prediction is an essential tool in the academic setting, and for this reason, regression analysis is an integral component of this statistical study.

Results

As general findings, five different correlations and regressions — depending on the purpose — were plotted according to each of the research questions. Each of the corresponding tables is listed below.

Model Summary – Dalc
Model	R	R²		Adjusted R²	RMSE
H₀	0.000	0.000		0.000	0.891
H₁	0.112	0.013		0.010	0.886

Coefficients
Model			Unstandardized			Standard Error	Standardized	t	p
H₀	(Intercept)		1.481			0.045		33.045	<.001
H₁	(Intercept)		1.410			0.055		25.728	<.001
	absences		0.012			0.006	0.112	2.233	0.026

Table 1. Linear regression test results for the relationship between alcohol use and school absenteeism.

Model Summary – Dalc
Model	R	R²		Adjusted R²	RMSE
H₀	0.000	0.000		0.000	0.891
H₁	0.055	0.003		0.000	0.891

Coefficients
Model			Unstandardized			Standard Error	Standardized	t	p
H₀	(Intercept)		1.481			0.045		33.045	<.001
H₁	(Intercept)		1.592			0.111		14.288	<.001
	G3		-0.011			0.010	-0.055	-1.085	0.278

Table 2. Linear regression test results for the relationship between alcohol use and math final grades.

Model Summary – Dalc
Model	R	R²	Adjusted R²	RMSE
H₀	0.000	0.000	0.000	0.891
H₁	0.126	0.016	0.011	0.886

Table 3. Linear regression test results for the joint relationship between alcohol use and absenteeism in school and math final grades.

Pearson’s Correlations
Variable		goout	Dalc
1. goout	Pearson’s r	—
	p-value	—
2. Dalc	Pearson’s r	0.267	—
	p-value	<.001	—

Table 4. Correlation analysis between alcohol consumption and meetings with friends.

Pearson’s Correlations
Variable		Dalc	age
1. Dalc	Pearson’s r	—
	p-value	—
2. age	Pearson’s r	0.131	—
	p-value	0.009	—

Table 5. Correlation analysis between alcohol consumption and age.

Interpretation of Results

Alcohol Consumption and Absenteeism from Class

Table 1 shows that the correlation coefficient (R) between these variables is 0.112, indicating a weak positive relationship. In terms of the unstandardized beta coefficient, each 1 unit increase in school absenteeism resulted in a 0.012 unit increase in the frequency of alcohol use. The P-value for this beta was below the 0.05 level of significance, which means that the coefficient is statistically significant.

Alcohol Use and Math Scores

Table 2 clearly shows that there is an inverse relationship between the variables. An increase in the semester grades each step led to a 0.011 decrease in the intensity of alcohol use at a p-value more significant than the critical level. Consequently, it can be concluded that there is no statistically significant relationship between these two variables.

Co-influence on Alcohol Consumption

It was interesting to clarify whether there was a common effect. According to Table 3, the joint effect for alcohol consumption is weak: the R coefficient was 0.126. It is already evident that the main contribution was from absenteeism since the math was shown not to present statistically significant patterns for alcohol.

Correlation of Alcohol use and Meetings with Friends

The correlation analysis shown in Table 4 shows a correlation coefficient between the variables of 0.267 with a statistically significant result (p-value<0.001). This is a weak relationship, showing that the two variables were virtually unrelated.

Relationship of Alcohol use and Age

Table 5 also shows that there was not even a moderate relationship between the student’s age and alcohol use: the r coefficient was 0.131. This is a critically low level of correlation, showing that age cannot directly influence the frequency of alcohol use.

Conclusion

The statistical analysis conducted with respect to high school students’ alcohol use suggests several important conclusions. First, increasing absences had virtually no effect on increasing alcohol consumption in any of the parameters studied. The most significant effects were found for frequency of meetings with friends (r=0.267), age (r=0.131), and absenteeism (r=0.112). Second, increases in final math grades showed no statistically significant relationship with increases in alcohol use in high school.

The results obtained may be helpful for school principals, parents, and social industry workers. It was shown that neither meetings with friends, age, nor school truancy were reliable predictors of alcohol use, which means that restricting any of these factors — as a punishment for the student, for instance — would not have the expected result. It also means that the cause of alcohol use is more profound, which means that one needs to continue to look for critical predictors to improve the academic agenda in high schools.

References

Flagel, J. 2021. Who alcohol use disorder affects — college students. ARH.

Vedantu. (2020). Difference between correlation and regression. Vedantu.