Studying vs Extracurricular Time
Tanner Werderits
10 17, 2019
Abstract
In this paper, we set out to see if more extracurricular time for high school students would lead to an increase in time spent studying for those students. Most people would think that because students are spending more time at extracurriculars they would have less time to study, but we think being involved with more extracurriculars leads to a better daily attitude which would lead to an increase in the amount of time spent studying for high school students. The data used for this study comes from The American Time Use Survey done by the United States Department of Labor for the years 2010 and 2018 and only looks at high school students between the ages of 14 and 18. We weren able to find statistical evidence that more time spent at extracurriculars will increase the amount of time spent studying by students. From the regressions we ran there was some evidence to support our theory, but there didn’t seem to be very many people surveyed that were involved in extracurriculars.Empirical Strategy
Growing up we had always thought that students who were more involved with extracurriculars have better grades than those who don’t. If students are spending time at extracurriculars, they will have less time to study which will lead to a lower grade point average is the thought process that some many think. We think that even though students are spending more time at extracurriculars which could be spent studying, being involved in a group and having that experience of working as a team will boost their everyday mood and allow them to focus on their goals. This in turn will lead to them studying more to improve their grade point average, allow them to get into the school they desire, and find a job that fits their interests.
We initially chose to represent this relationship with the simple regression below:
\(study time = \beta{0} + extracurricularx_{1} + \epsilon_{1}\)
We chose to run this test first just to see if there was any evidence of a positive linear relationship between time spent studying and time spent at extracurricular activities. Because there are so many factors that go into the amount of time someone spends studying each day, we assumed that this relationship would not be perfectly linear and wouldn’t leave us with any meaningful interpretations of the data.
We added the respondents age into the equation in our second second regression below:
\(study time = \beta{0} + extracurricularx_{1} + agex_{2} + \epsilon_{1}\)
When adding another variable to our initial regression, we are able to see if a variable other than extracurricular time affects time spent studying. The thought behind adding age into the regression was because as students get older we expect them to be taking harder classes that should result in more time spent studying.
The first two regressions were mostly just to see if our initial thoughts had any truth to them and to see at the most basic level, (simple linear regression) if our variables were correlated in any way. From the start we knew that there is only a certain amount of time in a day. If extracurricular time continued to increase there had to be some threshold where too much time was spent at extracurriculars and there would be a decrease in time spent studying. For example, if a respondent recorded that they spent 60 minutes of their day doing something that is categorized as an extracurricular activity, they will still have plenty of time to study later that night before they go to bed. If a respondent recorded 180 minutes doing something that is categorized as an extracurricular activity, they most likely don’t have any significant amount of time to study. Altering the functional form to a polynomial allows us look at this relationship when we expect study time to eventually start to decrease. The only thing that would change from the multiple regression above is that there would be a squared term attached to the extracurricular variable.
Data
The data we used for this study came from the Bureau of Labor Statistics yearly American Time Use Survey. This survey is done yearly and measures the amount of time people spend doing various activities, such as paid work, childcare, volunteering, and socializing. The data analyzed was from the surveys done in 2010 and 2018. The way that the survey works is the person being surveyed will keep track of everything they do for an entire day, down to the minute, and report this data. We were able to condense that data to look only at respondents who said they were in high school and between 14 and 18 years old. This allows us to take out any bias of people who dropped out of high school and the potential difference in how college students spend their time compared to high school students. We chose not to control for day of the week the respondent was surveyed. Some may say that students are more likely to be involved in extracurriculars on weekdays, but might be more likely to spend more time studying on weekends. That statement does have some validity, but from our past experiences as high school students we thought that because students’ schedules vary so much and with extracurricular events taking place on the weekends; the day the respondent was surveyed wouldn’t matter. Each day should average out to similar results for the variables we are analyzing because students’ schedules vary greatly.
Key Measures
Our dependent variable, study time, is measured in minutes per day that the respondent was surveyed. Extracurricular time, an independent variable, only represents the time spent on extracurricular educational activities (excluding sports) in minutes per day which is important to keep in mind when looking at the results of the research. The other independent variable, age, is measured in years. Another thing to keep in mind when looking at the results of the research is that the respondents only kept a “time diary” for one day and not multiple days, so the time spent either studying or at extracurricular activities is the time the respondent had in minutes for that day. ## Descriptive Statistics
atus <- read.csv("atus_00002.csv")sample <- atus %>%
filter(SCHLCOLL == 3)%>%
filter(AGE >= 14 & AGE <= 18) %>%
select(STUDY_TIME, EXTRACURRIC_TIME, AGE)
rm(atus)
stargazer(sample,type="text")##
## ==============================================================
## Statistic N Mean St. Dev. Min Pctl(25) Pctl(75) Max
## --------------------------------------------------------------
## STUDY_TIME 762 50.808 96.018 0 0 65 886
## EXTRACURRIC_TIME 762 3.366 24.203 0 0 0 400
## AGE 762 16.290 0.956 15 16 17 18
## --------------------------------------------------------------Results
The initial simple regression that was ran only takes into account extracurricular time compared to study time. The issue with using this regression to describe a relationship is that there are many other factors need to be taken into account to try and predict the amount of study time for a high school student. We were able to come to the conclusion that of the people surveyed, that were in high school, and between the ages of 14 and 18, study time will increase by 0.21 minutes for every one minute increase in extracurricular time during the day. The tests we ran also told us that if extracurricular time is 0 minutes, we can expect study time to be 50 minutes for the high schooler. The data is very spread out and the results are not precise enough to say that these two variables have a linear relationship. We are unable to confirm that more time spent at extracurricular activities will result in an increase in time spent studying, holding all else constant for this data set.
When we controlled for age, we found that study time would increase by 0.214 minutes for every one minute increase in extracurricular time for high school students between the ages of 14 and 18 while holding age constant and that for a one year increase in age, study time is estimated to decrease by 0.84 minutes for high school students between the ages of 14 and 18 while holding time spent at extracurricular activities constant. Our initial assumptions that as a student gets older their study time will increase were wrong here. As you can see the new relationship we tested decreased the amount that study time would increase for a one minute increase in extracurricular activities. It was not a significant change, but you can see that controlling for age had a small effect on the outcome from the test. The data is still very much scattered with a lot of the respondents having 0 minutes of extracurricular activities on the day they were surveyed. Because of this we were still unable to say that these new predictions for the affect time spent doing extracurricular activities has on study time are correct.
simple <- lm( STUDY_TIME ~ EXTRACURRIC_TIME, data = sample)
summary(simple)##
## Call:
## lm(formula = STUDY_TIME ~ EXTRACURRIC_TIME, data = sample)
##
## Residuals:
## Min 1Q Median 3Q Max
## -136.62 -50.08 -50.08 9.92 835.92
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 50.0802 3.5090 14.272 <2e-16 ***
## EXTRACURRIC_TIME 0.2163 0.1437 1.506 0.133
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 95.94 on 760 degrees of freedom
## Multiple R-squared: 0.002974, Adjusted R-squared: 0.001662
## F-statistic: 2.267 on 1 and 760 DF, p-value: 0.1326sample %>%
ggplot(aes(x = EXTRACURRIC_TIME, y = STUDY_TIME)) +
geom_point() +
geom_smooth(method = "lm") +
ylim(0,400)
multiple1.age <- lm(STUDY_TIME ~ EXTRACURRIC_TIME + AGE, data = sample)
stargazer(simple, multiple1.age, type = "text")##
## ===========================================================
## Dependent variable:
## ---------------------------------------
## STUDY_TIME
## (1) (2)
## -----------------------------------------------------------
## EXTRACURRIC_TIME 0.216 0.214
## (0.144) (0.144)
##
## AGE -0.836
## (3.650)
##
## Constant 50.080*** 63.704
## (3.509) (59.600)
##
## -----------------------------------------------------------
## Observations 762 762
## R2 0.003 0.003
## Adjusted R2 0.002 0.0004
## Residual Std. Error 95.938 (df = 760) 95.998 (df = 759)
## F Statistic 2.267 (df = 1; 760) 1.158 (df = 2; 759)
## ===========================================================
## Note: *p<0.1; **p<0.05; ***p<0.01multiple1.age <- lm(STUDY_TIME ~ EXTRACURRIC_TIME + AGE + I(EXTRACURRIC_TIME^2), data = sample)
stargazer(simple, multiple1.age, type = "text")##
## =============================================================
## Dependent variable:
## ----------------------------------------
## STUDY_TIME
## (1) (2)
## -------------------------------------------------------------
## EXTRACURRIC_TIME 0.216 0.832***
## (0.144) (0.316)
##
## AGE -0.168
## (3.654)
##
## I(EXTRACURRIC_TIME2) -0.003**
## (0.001)
##
## Constant 50.080*** 52.258
## (3.509) (59.679)
##
## -------------------------------------------------------------
## Observations 762 762
## R2 0.003 0.009
## Adjusted R2 0.002 0.005
## Residual Std. Error 95.938 (df = 760) 95.757 (df = 758)
## F Statistic 2.267 (df = 1; 760) 2.380* (df = 3; 758)
## =============================================================
## Note: *p<0.1; **p<0.05; ***p<0.01Robustness Checks
After failure to accurately predict the relationship between time spent doing extracurricular activities and the amount of time a high school student between the ages of 14 and 18 studies during the day, we ran a new test to see if at some certain amount of minutes spent at extracurricular activities study time would start to decrease. The results from this test supplied the most precise estimates for this relationship. The new test suggests that a one minute increase in extracurricular time is associated with a change in minutes of study time of 0.832+2*-0,003*time spent doing extracurricular activities for high schoolers between the ages of 14 and 18. The data still does not fit this new predicted relationship very well, but it is by far the closest to the actual relationship between the variables. It is enough evidence to say that there is a relationship between the amount of time high school students spend at extracurricular activities and the amount of time they spend studying in a single day.
Conclusion
Throughout the research done we were able to conclude that there is a relationship between time spent at extracurricular activities and study time of high school students, but we don’t know the best equation to predict the amount of study time. The third test we ran gave us the closest relationship between the two variables while also holding age constant. The data is very spread out and of the people sampled, very few had any time spent doing an extracurricular activity which we think is making it difficult to find a good equation to predict study time compared to extracurricular time during a day.
When looking at the practical significance of our findings we assumed that most extracurricular activities lasted for roughly 60 minutes a day if the student was involved. If you take the data we received from our third test, we would expect study time to increase 0.832 minutes for one additional minute of an extracurricular activity, but we expect study time to only change by 0.832+2-.00360= 0.472 minutes for an additional minute of an extracurricular activity after 60 minutes have already occured. The relationship shows that more extracurricular time leads to more time spent studying, but eventually too much extracurricular time will lead to a decrease in study time like we hypothesized. If the student has no extracurricular time we expect them to spend roughly 52 minutes of their day studying.
When looking at this data and the results we found it is important to look at some of the limitations we faced with the actual sample. There was only 762 individuals that fit into the filters we applied to the Time Use Survey data. The mean for extracurricular time was also only three minutes for the sample. There was statistical significance that our data fit a polynomial regression model, but our sample size could defiently be larger and more respondents that actually engaged in an extracurricular activity the day they were studied. The reason the mean for time spent doing an extracurricular activity is so low is because most respondents reported 0 minutes in the category, and you can see this represented in the graph. A larger sample size would allow us to get an answer that is closer to what actually happens between the variables in the real world.
We think if the extracurricular time included when people were playing sports and not just education extracurricular activities, we would have been able to find a closer estimate. We may also need to take into account the day of the week the respondent was surveyed. There is a chance if we were able to get our hands on the amount of time spent playing sports as well we could add that into our predicted relationship and potentially get a good estimation for the relationship. Throughout the paper there are a couple assumptions that we made based on our personal experiences and observations being students. When looking for other sources to back our assumptions, we ran into trouble because most research done on study time involved college students and was compared to time spent working a job. We didn’t see any strong evidence here to support some of the assumptions made. This shouldn’t have a large effect on the results because the assumptions were reasons justifying why we ran the test we did.
Our paper only looks at the effect extracurricular time has on the amount of study time for a high school student. The data we used didn’t supply a grade point average for people in this age range. Instead of just looking at the amount of study time, there is a chance to research the effect time spent at extracurricular activities has on grade point average for the student. This has been researched before, but there is always new data available for people to analyze. The thought behind a study looking at grade point average instead of study time is that being involved in extracurriculars supplies students with better time management skills and not just a better attitude. These students may be more effective when studying, resulting in less time spent studying.
-I used “we” instead of “I” because most people write a research paper with multiple people.