A Statistical Analysis of Greek vs. Non-Greek Habits and Lifestyles

Introduction

Upon entering college, students must make the decision to join a Greek organization or remain independent. There are different lifestyles and habits that develop out of the decision to join a Greek organization or remain independent. When discussing lifestyles and habits, we are including drinking, smoking, sleeping, studying, dating, quality of life, GPA, extra-curricular activity time, and evenings spent out per week. Due to the belief of a large difference between habits and lifestyles of the two organizations, a survey was semi-randomly distributed to the campus to try to determine if this difference truly does exist. Through hypothesis testing of Greeks and Non-Greeks the statistics show that in some areas there are differences between habits and lifestyles of the two groups. A regression analysis was also done to determine the factors which affect the quality of life for Hanover College students.

Data

There were 248 surveys distributed to students through the Greek leaders, Resident Assistants and theme house leaders. The distribution within these areas can be seen in the chart below.

Of these 248 surveys distributed, 145 were returned and included in the database. In order to better understand the statistical analysis of our information some of the variables must be defined. Students who responded that they were Greek received a 1 and those responding Non-Greek received a 0. When completing a regression model analysis Greek will be a dummy variable. In the gender category, males received a 1 and females received a 0. The GPA was a cumulative grade point average on a four-point scale. Those students who are currently dating someone received a 1 and those who are not received a 0. When completing a regression model analysis, dating someone will be a dummy variable. As for alcoholic beverages consumed, the quantitative data is number of alcoholic beverages consumed in a week. Another question with data in the weekly form is the amount of hours outside of class spent studying, including reading textbooks, completing homework, writing papers, labs, and research group meetings. Also in a weekly time frame, we asked how much out of pocket money is spent in dollars. Yet another category which has quantitative data in a weekly format is the number of evenings individuals spend out, meaning leaving Hanover College, going to a movie, out to a fraternity, or other activities. The category of extra-curricular activities takes into account the number of hours spent per week on these activities not including work, but including varsity sports, intramural sports, Greek organizations, and other student organizations. On a daily basis the survey asked for the average hours spent sleeping on a typical weeknight and the number of cigarettes smoked on a typical day. In order to gauge the quality of life at Hanover College, the survey requested a numeric response based on a scale from 1 to 10.

Our survey also consisted of other questions so that the main purpose of our project would not be disclosed, resulting in biased answers. The questions regarding class standing, whether the person one’s dating is enrolled at Hanover College, and if students were satisfied with the educational backgrounds of the professors at Hanover College will not be included in the statistical analysis. The reason for this is that they filled the position in the survey as distracter questions.

Summary Statistics

Summary statistics provide valuable information when analyzing data. Listed below are all of the summary statistic charts from the relevant topics on the survey. To avoid repetition, the following paragraph explains the significance and meaning of each item that relates to this project in the summary statistics chart. The same concept can then be carried through when analyzing all of the categories listed below.

Not all of the information given in the chart is relevant to this project and therefore it will not be discussed. The measures of location are mean, median, and mode. The mean gives the average. The median is the number at which 50% of the values are above and 50% of the values are below. The mode is the number which occurs with the most frequency. The measures of dispersion show how widely spread the data is from the mean and includes, range, variance, and standard deviation. The range is taken from the difference between the maximum and minimum values in the data set. The variance measures deviations from the mean, but is not very useful because it is not yet in the unit form of the observations. Finally, the standard deviation is how far, on average, each observation is from the mean.

Non-Greek How Many Drinks a Week		Greek How Many Drinks a Week
Mean	2.443181818	Mean	7.737373737
Standard Error	0.63091093	Standard Error	1.032764208
Median	0	Median	3
Mode	0	Mode	0
Standard Deviation	4.184989663	Standard Deviation	10.27587412
Sample Variance	17.51413848	Sample Variance	105.593589
Kurtosis	6.546079577	Kurtosis	2.693720026
Skewness	2.3581628	Skewness	1.72246934
Range	20	Range	45
Minimum	0	Minimum	0
Maximum	20	Maximum	45
Sum	107.5	Sum	766
Count	44	Count	99
Non-Greek How Much Sleep a Night		Greek How Much Sleep a Night
Mean	7.139534884	Mean	6.8775
Standard Error	0.196400035	Standard Error	0.460134134
Median	7	Median	6
Mode	6	Mode	6
Standard Deviation	1.287881157	Standard Deviation	4.601341345
Sample Variance	1.658637874	Sample Variance	21.17234217
Kurtosis	3.514232053	Kurtosis	79.87760792
Skewness	1.121895848	Skewness	8.539228166
Range	7.5	Range	46.5
Minimum	4.5	Minimum	3.5
Maximum	12	Maximum	50
Sum	307	Sum	687.75
Count	43	Count	100
Non-Greek How Much Study a Week		Greek How Much Study a Week
Mean	15.61363636	Mean	16.005
Standard Error	1.586050418	Standard Error	0.918262935
Median	15	Median	14.5
Mode	15	Mode	10
Standard Deviation	10.52066827	Standard Deviation	9.182629352
Sample Variance	110.6844609	Sample Variance	84.32068182
Kurtosis	1.047581334	Kurtosis	1.086171051
Skewness	1.118231052	Skewness	0.978700722
Range	43.5	Range	49
Minimum	1.5	Minimum	1
Maximum	45	Maximum	50
Sum	687	Sum	1600.5
Count	44	Count	100
Non-Greek Cigarettes a Day		Greek Cigarettes a Day
Mean	0.363636364	Mean	1.775
Standard Error	0.205862567	Standard Error	0.562299347
Median	0	Median	0
Mode	0	Mode	0
Standard Deviation	1.365537786	Standard Deviation	5.622993469
Sample Variance	1.864693446	Sample Variance	31.61805556
Kurtosis	11.65278026	Kurtosis	24.26179397
Skewness	3.601543258	Skewness	4.533139617
Range	6	Range	40
Minimum	0	Minimum	0
Maximum	6	Maximum	40
Sum	16	Sum	177.5
Count	44	Count	100
Non-Greek How Many Evenings Go Out in a Week		Greek How Many Evenings Go Out in a Week
Mean	1.886363636	Mean	2.515151515
Standard Error	0.230396086	Standard Error	0.135346305
Median	2	Median	2
Mode	2	Mode	2
Standard Deviation	1.528274739	Standard Deviation	1.346678729
Sample Variance	2.335623679	Sample Variance	1.813543599
Kurtosis	2.71322423	Kurtosis	1.443924561
Skewness	1.474654553	Skewness	0.977782214
Range	7	Range	7
Minimum	0	Minimum	0
Maximum	7	Maximum	7
Sum	83	Sum	249
Count	44	Count	99
Non-Greek How Much Time Spent on Extra-Curricular Activities in a Week		Greek How Much Time Spent on Extra-Curricular Activities in a Week
Mean	6.647727273	Mean	9.816326531
Standard Error	0.974920654	Standard Error	0.711976671
Median	5	Median	6.75
Mode	0	Mode	5
Standard Deviation	6.466892019	Standard Deviation	7.048209451
Sample Variance	41.82069239	Sample Variance	49.67725647
Kurtosis	-0.54072602	Kurtosis	0.535401446
Skewness	0.755378872	Skewness	1.08661
Range	21	Range	30
Minimum	0	Minimum	0
Maximum	21	Maximum	30
Sum	292.5	Sum	962
Count	44	Count	98
Non-Greek Cumulative GPA		Greek Cumulative GPA
Mean	3.139909091	Mean	3.002020202
Standard Error	0.085855253	Standard Error	0.047782184
Median	3.07	Median	2.98
Mode	3.6	Mode	2.8
Standard Deviation	0.569499324	Standard Deviation	0.475426725
Sample Variance	0.32432948	Sample Variance	0.226030571
Kurtosis	-1.17743921	Kurtosis	-0.49084082
Skewness	-0.19911422	Skewness	0.001958121
Range	1.982	Range	2
Minimum	1.998	Minimum	2
Maximum	3.98	Maximum	4
Sum	138.156	Sum	297.2
Count	44	Count	99
Non-Greek Dating Someone		Greek Dating Someone
Mean	0.613636364	Mean	0.46
Standard Error	0.074253929	Standard Error	0.050090827
Median	1	Median	0
Mode	1	Mode	0
Standard Deviation	0.492544843	Standard Deviation	0.500908266
Sample Variance	0.242600423	Sample Variance	0.250909091
Kurtosis	-1.8527476	Kurtosis	-2.014136
Skewness	-0.48339809	Skewness	0.162969267
Range	1	Range	1
Minimum	0	Minimum	0
Maximum	1	Maximum	1
Sum	27	Sum	46
Count	44	Count	100
Non-Greek Quality of Life at HC		Greek Quality of Life
Mean	6.909090909	Mean	7.405
Standard Error	0.3029422	Standard Error	0.157065096
Median	7	Median	8
Mode	8	Mode	8
Standard Deviation	2.009491221	Standard Deviation	1.570650962
Sample Variance	4.038054968	Sample Variance	2.466944444
Kurtosis	1.478540765	Kurtosis	2.720518974
Skewness	-1.18480371	Skewness	-1.42817189
Range	9	Range	8
Minimum	1	Minimum	2
Maximum	10	Maximum	10
Sum	304	Sum	740.5
Count	44	Count	100

Although the summary statistics did not play an integral part in the objectives of the analysis, they were useful in checking data entry. This was done through examination of the maximums and minimums in order to detect any extreme outliers, answers that were not within the scale allowed.

Do Habits/Lifestyles of Greeks and Non-Greeks Differ?

In order to determine if Greek and Non-Greek habits and lifestyles do differ, it is necessary to run a series of hypothesis tests. In order to avoid repetition, the process of arriving at a conclusion will be described. Then, the conclusion of each test will be explained.

Since the project is to determine if the two differ, with it being of no significance to this project which way they differ, a two-tailed hypothesis test was conducted. Each hypothesis test was conducted at a 95% confidence level. The reasoning behind the 95% confidence level was that going higher would include such a large number of the values, as it would have a larger standard deviation, that it would make it very difficult for there to be any difference between any two groups. Since there are no significant consequences if a value falls outside the specified confidence interval, the 95% confidence level provides a high level of confidence without losing all statistical interest. In all cases, the question is if Greeks and Non-Greeks differ on the habit or lifestyle. Therefore the same hypothesis test is used for all categories.

H_o: (μ_Non-Greek – μ_Greek) = 0

H_a: (μ_Non-Greek – μ_Greek) g 0

Since there are over 30 observations for both Greeks and Non-Greeks, the hypothesis test was conducted using z-scores. This process was done through Excel “z-test: Two Sample for Means.” If the critical z-score is closer to the expected value of the true mean than the absolute value of the test statistic z-score then the conclusion is to “Reject the H_o.” If the critical z-score is not closer to the expected value of the true mean than the absolute value of the test statistic z-score then the conclusion is to “Fail to Reject the H_o .” The critical z-score for all of the hypothesis tests conducted in this analysis at a 95% confidence interval is 1.959961082. The conclusions follow below.

Smoking

The test statistic for this hypothesis test was -2.356991478. Therefore, since this value is in the rejection range, the conclusion at a 95% confidence level is to reject the H_o. This means we are 95% confident that there is a difference in the smoking habits of Non-Greeks and Greeks.

Drinking

The test statistic for this hypothesis test was -4.374542654. Therefore, since this value is in the rejection range, the conclusion at a 95% confidence level is to reject the H_o. This means we are 95% confident that there is a difference in the drinking habits of Non-Greeks and Greeks.

Sleeping

The test statistic for this hypothesis test was 0.523759399. Therefore, since this value is not in the rejection range, the conclusion at a 95% confidence level is to fail to reject the H_o. This means we are 95% confident that there is not a difference in the sleeping habits of Non-Greeks and Greeks.

Studying

The test statistic for this hypothesis test was -0.213545661. Therefore, since this value is not in the rejection range, the conclusion at a 95% confidence level is to fail to reject the H_o. This means we are 95% confident that there is not a difference in the studying habits of Non-Greeks and Greeks.

Extra-Curricular Activities

The test statistic for this hypothesis test was -2.624706519. Therefore, since this value is in the rejection range, the conclusion at a 95% confidence level is to reject the H_o. This means we are 95% confident that there is a difference in the time spent on extra-curricular activities of Non-Greeks and Greeks.

Dating

The test statistic for this hypothesis test was 1.715272124. Therefore, since this value is not in the rejection range, the conclusion at a 95% confidence level is to fail to reject the H_o. This means we are 95% confident that there is not a difference in the dating habits of Non-Greeks and Greeks.

Quality of Life

The test statistic for this hypothesis test was -1.453264279. Therefore, since this value is not in the rejection range, the conclusion at a 95% confidence level is to fail to reject the H_o. This means we are 95% confident that there is not a difference in the views of quality of life Non-Greeks and Greeks.

GPA

The test statistic for this hypothesis test was 1.403362567. Therefore, since this value is not in the rejection range, the conclusion at a 95% confidence level is to fail to reject the H_o. This means we are 95% confident that there is not a difference in the GPA of Non-Greeks and Greeks.

Evenings Out

The test statistic for this hypothesis test was -2.353163182. Therefore, since this value is in the rejection range, the conclusion at a 95% confidence level is to reject the H_o. This means we are 95% confident that there is a difference in the number of evenings Non-Greeks and Greeks go out per week.

After this series of hypothesis tests we conclude that there are significant differences in some of the habits and lifestyles of Greeks and Non-Greeks. The statistics show that there is not a significant difference in sleeping habits, studying habits, dating habits, quality of life, or grade point average. However, there is a significant difference in smoking habits, drinking habits, time spent on extra-curricular activities, and evenings out per week for Greeks and Non-Greeks.

Differences in Greeks and Non-Greeks in those individuals who do smoke or drink

Since a large number of Hanover College students do not drink or smoke, we created a database which included only those individuals who drink or smoke. This caused our sample size to be less than 30, causing us to use a t-test. We tested at a 95% confidence level for the same reasons as in the z-tests. In all cases we will be questioning if drinkers and smokers who are Greek or Non-Greek differ in their drinking and smoking habits. The same hypothesis test is used for both categories.

H_o: (μ_Non-Greek – μ_Greek) = 0

H_a: (μ_Non-Greek – μ_Greek) g 0

The process was done through Excel using “t-Test: Two-Sample Assuming Unequal Variances.” If the critical t-score is closer to the expected value of the true mean than the absolute value of the test statistic t-score then the conclusion is to “Reject the H_o.” If the critical t-score is not closer to the expected value of the true mean than the absolute value of the test statistic t-score then the conclusion is to “Fail to Reject the H_o.” The conclusions follow below.

Smokers

The t-score for the hypothesis test was -1.904848897. The critical t-score for this hypothesis test was 2.100923666 at 19 degrees of freedom. Therefore, since this value is not in the rejection range, the conclusion at 95% confidence level is to fail to reject the H_o. This means that we are 95% confident that there is not a difference in smoking habits of Non-Greeks and Greeks among those who do smoke.

Drinkers

The t-score for the hypothesis test was -3.33543. The critical t-score for this hypothesis test was 1.993944 at 88 degrees of freedom. Therefore, since this value is in the rejection range, the conclusion at 95% confidence level is to reject the H_o. This means that we are 95% confident that there is a difference in the drinking habits of Non-Greeks and Greeks among those who do drink alcohol.

Factors that Affect the Quality of Life of Hanover College Students

In order to determine if there are any significant factors affecting the quality of life of Hanover College students, a regression analysis was performed. The following equation represents our model.

Quality of Life = ß₀ + ß₁ (Current Class Standing) + ß₂ (Cumulative GPA)

+ ß₃ (Dating Someone) + ß₄ (How Many Drinks a Week) + ß₅ (Greek)

Excel was used to perform the regression analysis; the dependent variable was quality of life and the independent variables were from the categories of current class standing, cumulative GPA, dating someone, how many drinks per week, and Greek. The output from Excel produced the following information.

	Coefficients	P-Values
Intercept	5.912153996	1.14543 E-08* (99% C.L.)
Current Class Standing	-.03002359	.814550404
Cumulative GPA	.531377209	.080645344* (90% C.L.)
Dating Someone *	-.78224982	.007119382* (99% C.L.)
How Many Drinks a Week	-.023795839	.16171321
Greek *	.526026709	.105876953

* notes dummy variables * notes statistical significance

In order to determine if there is statistical significance in any of the independent variables, we must conduct a hypothesis test.

H_o: ß₁ = 0

H_a: ß₁ g 0

The same hypothesis test is used for all of the independent variables. Before comparing the p-values it is helpful to note what results are expected. It is expected that current class standing will have a negative relationship on quality of life. The rest of the independent variables are believed to have a positive relationship. As noted in the above chart, the only independent variables with statistical significance were cumulative GPA at .080645344 and dating someone at .007119382. To determine that they were statistically significant the absolute value of the p-value had to be less than the alpha. The confidence levels where this statistical significance occurs are noted on the chart. Also, the intercept is statistically significant, but this is of little importance to our conclusions.

When examining the data, the following interpretations of estimated coefficients may be made. If current class standing increases one year, then quality of life will decrease .0300. If cumulative GPA increases one point, then the quality of life will increase .5314. If you are dating someone, then the quality of life will decrease .7822. If you drink one more alcoholic beverage per week, your quality of life will decrease .0238. If you are Greek, your quality of life will increase .5260. Notice dating someone is the only independent variable that is statistically significant at 99% confidence level. Surprisingly to us, there was a negative relationship between dating someone and the quality of life of a Hanover College student.

In order to determine if we have a good model, the R² must be analyzed. The R² in our model is 0.108596072. This means that our model explains 10.86% of the variance in the determination of quality of life. Note that this R² would not be categorized as an acceptable percentage of variation explained. The next regression statistic which we examined was R_a². In our model this was 0.075334731. It is important to note that this is a few hundredths away from R². This tells us that one of the independent variables does not have a strong relationship with the quality of life. One of the statistics which tells us the significance of the model is the significance of F. In order to determine if the model is significant, we use a hypothesis test to examine the significance of F.

H_o: ß₁ = ß₂ = ß₃ = ß₄ = ß₅ = 0

H_a: ß₁ g ß₂ g ß₃ g ß₄ g ß₅ g 0

In our model the significance of F was 0.008170431. Testing at a 99% confidence level, we see that .008170431 is less than .01. We therefore, reject the H_o and conclude that our model has statistical significance.

To make sure that there was not a multicollinearity issue within the model, we developed a correlation matrix to see if any variables had a correlation of above a .7. None of our variables had a correlation of above .7. Therefore, multicollinearity is not an issue within our model.

Conclusion

With Greek life being such a controversial topic in today’s society and media, it is important to see the true difference between students who are Greek or Non-Greek. Students who are coming to the decision of joining or not joining a Greek organization should realize the possible differences between the two lifestyles. In a random survey distributed to 248 students there was a return rate of 58%. Then, we tested to see if there was a difference between the habits and lifestyles of Greeks and Non-Greeks. We also looked at the affect certain habits played into the student’s perceived quality of life.

Our results show that there is a significant difference in smoking habits, drinking habits, time spent on extra-curricular activities, and evenings spent out per week for Greeks and Non-Greeks. It also shows that dating someone has a statistically significant negative affect on a student’s quality of life. In addition, we found that if a student’s GPA increases one point, then the student’s quality of life will increase .5314 on a 10-point scale.

As stated, Greek life is a very controversial topic in society and Greek organizations have recently had large amounts of bad publicity. Our analysis could be used by anti-Greek organizations as well as pro-Greek organizations to show that there are differences between Greeks and Non-Greeks. The media could also use this same information when discussing current Greek issues.

There are also many uses of the information on college campuses. Faculty members may be interested in the information so that they could better understand their students by knowing there are differences in some habits and lifestyles of Greeks and Non-Greeks. Similarly, administration could use the information when determining whether or not to promote Greek Life on the campus or in determining whether or not to allow another Greek organization on campus.

Based upon the factors that affect the quality of life for college students, a campus counselor could use the information when dealing with a depressed student that doesn’t enjoy campus life. For example, she could advise the student that a poor quality of life could be attributed to dating someone. Also, that a better quality of life can be reached by earning a higher GPA.

The above are only a few ways that the information from our study can be used. However, in reality, any person who comes in contact with Greek and Non-Greek students could use the information to better understand the differences in their habits and lifestyles.