Sleep Study of College Students at North Carolina State University Allen, Jessup and Moreno

Executive Summary In conducting this study, the researchers hypothesized that students with heavier course loads (i.e. more credit hours per semester) obtain less hours of sleep each night than those with lighter course loads. That being said, students in more math-oriented concentrations (Engineering, Physical Sciences, etc.) would also receive low levels of sleep. This is contrasted by those in Humanities and Social Sciences, who are thought to receive more hours of sleep per night due to a less extensive course load. It was also believed that the reasons for these students’ lack of sleep would be due to extracurricular activities, as compared to those in math-based majors losing sleep due to attending to their studies. While the gender of each subject was recorded, it was not initially thought to have a significant impact on the hours of sleep per night; it was only used to ensure an even sample of the student population. Major findings from the study showed that the majority of the sampled student body had some sort of sleep deprivation, and linked homework to being the foremost cause for this. This research also showed that gender and major had no correlation with sleep.

1

Contents Description of data

3

Statistical analysis

4

Major Findings

8

Discussion

9

Appendix

10

2

Description of Data The intention of this study is to determine the sleep habits of college undergraduates and the underlying causes that prevent them from receiving an adequate amount of sleep per night. A survey was given over the first week of October, 2009, to a sample size of 257 North Carolina State University students. In this survey, students of all grade levels were questioned on their average sleep patterns during the current semester. The survey retrieved data on gender, main attributed cause of sleep loss, hours of sleep received on average, major, and their rank. The main attributed causes of sleep were defined as homework, going out with friends, watching television, working late, doing stuff on a computer, trouble falling asleep, or other. In addition to gender and the average amount of sleep of the survey takers, they were also asked about their course of study, grade level, number of credit hours taken this semester, and reason for not obtaining as much sleep as the subject would like. For the last question, students were given seven options to choose from: Doing homework, going out at night with friends, watching TV, spending time on the computer, working late at a job, not being able to fall asleep, or other. A copy of the survey can be found in the appendix. By conducting this survey during the midpoint of the fall ’09 semester, we hoped to get a general idea of the average number of hours of sleep of the student population, at a time in the semester when class workload was evenly distributed. Also, while the hours of sleep needed per night varied from person to person, eight hours was considered to be the norm of the average adult.

Statistical analysis Data collected from the surveys was analyzed through the use of the R statistical computing program. Once the data was entered, several plots were made. These included plots of sleep vs. gender, sleep vs. class rank, sleep vs. credit hours per semester, and sleep vs. major. In addition to the plots, summaries were formed for each of the aforementioned plots. These summaries were linear regression models that show the correlation between different aspects of the data. All code used to form summaries can be located in the appendix. 3

The R2 value of a given set of data is used as a predictor for future outcomes of the experiment, and is something that can be determined from a linear regression model. R2 values for this given set of data proved to be low for all parameters of the data, including sleep based on gender, sleep based on major, and the causes for a lack of sleep.

The above graph displays the weekly hours of sleep per night dependent upon the major of the survey participants. The graph displays an even distribution amongst majors, with engineering having the lowest and highest sleep values. No major achieved an average of higher than seven hours of sleep. The few outliers below five hours were probably due to the several reports of students taking more than twenty hours as a course load. The below summary was generated by a linear regression line of sleep versus major. The low R squared value shows that there isn’t a correlation between sleep and major. This can be attributed to each major having an even course load. Residuals: Min

1Q Median

3Q

Max

-3.5833 -0.6304 0.2288 0.6271 2.9167

4

Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 6.83000 0.24187 28.238 |t|) |t|) (Intercept) 6.54592 0.12189 53.705 mydata=read.table("clipboard", header=T) > gender=as.factor(mydata$Gender) > major=as.factor(mydata$Major) > rank=as.factor(mydata$Rank) > cause=as.factor(mydata$Cause) > sleep=mydata$Sleep > credit=mydata$Credit > plot(sleep ~ gender) > plot(sleep~major) > plot(sleep~rank) > plot(sleep~cause) > plot(sleep~credit) > fit=lm(sleep ~ gender+major+rank+credit+cause) > summary(fit) > fit2=lm(sleep~cause) > summary(fit2) > fit3=lm(sleep~credit) > summary(fit3) > fit4=lm(sleep~gender) > summary(fit4) 10

> fit5=lm(sleep~rank) > summary(fit5) > fit6=lm(sleep~major) > summary(fit6)

11