2. Background
When people go off to college their habits in life change. Students begin to mature
and take part in adult activities such as drinking. Many people have noticed that
for some reason Dominoes and other take out franchises are on campus on the
weekend. Could this be because people need extra food to suppress the effects of
alcohol?
3. Initial & Precise Questions
Initial Question:
Whether drinking is related to ordering take out?
Precise Question:
Whether there is an association between the number of nights students go out
and drink, and the number of nights students order take out?
4. Variables
Explanatory Variable:
Whether or not college students drink alcohol
-Measured by amount of nights per week college students drink
Response Variable:
Whether or not college students order takeout
-Measured by amount of nights per week college students order take out
5. Data Collection
-Four Randomly selected dorms throughout campus, randomly selected floors
-Anonymous survey was on Iphone and Ipad using Google Form. (eliminates
response bias, not answering us directly)
-Two questions: How much do you drink/go out in a week? How many times do
you order takeout a week?
-Possible answers for both questions include: (1) Zero nights/week, (2) 1-2
Nights/Week, (3) 2 or More Nights/Week.
8. Sampling Test
- Our data tells us that 41 of the 67 students (61.19%) we surveyed were
female. According to a 2012 Forbes Article, the proportion of female students
for all colleges in the United States was 57.85%.
- Using a one-sample proportion test, we did not find significant evidence at the
5% level (p-value = 0.57936) that the population proportion of females (from
our sample) differs from 57.85% (reputable figure).
- Therefore, there is not significant evidence that our study is biased. The test
confirms that our study is representative of our population and that we
sampled randomly.
9. Hypothesis Testing
-Alternate Hypothesis: Drinking and ordering take out have a correlation.
-p-value = 0.0029171
0.0029171 < .05 so we found significant evidence at the 5% level that drinking
alcohol and ordering take-out are related in all college students. Therefore, there
is evidence of dependence between the frequency to which students order take-
out and their consumption of alcohol.
10. Assumptions
(1)The sample we chose is indeed a simple random sample as determined by
our sampling design. Each dorm, floor, and student surveyed was chosen at
random using a hat. Therefore, assumption 1 is met.
(2)Being that our population is all college students, the large population
assumption is met. Our sample consisted of 67 students*20 = 1,340. Clearly
there are at least 1,340 college students in the world.
(3)The normality assumption was not met for our study. The Chi-Squared
template indicated that only 44.44% of expected cells had a count of at least
5. This is much less than the required 80% criterion.
11. Sampling Bias
- Students surveyed were mostly freshmen and sophomores
- People surveyed from other dorm buildings
12. Measurement Bias
- Non-response bias
- Some students not present for survey
- Some did not answer door
- Students response around friends
13. Possible Lurking Variables
- Student athletes (less likely to drink, more health conscious)
- Wealth of students
● More money=More alcohol
=More take out
- Jobs
14. Ways to Improve Study
- Could be a larger sample size
- Get all grades involved, see if the fact that Juniors and Seniors being above 21
might skew or affect data a certain way
- Also Juniors and Senior might have jobs so the fact they have more money than
Freshman or sophomores could skew data a certain way
- Change the day of the week that data was collected, Friday isn’t the best day
possible to collect data about drinking
15. Conclusion
We found significant evidence that a college student’s alcohol consumption is
associated with a college student's frequency to order take out.
Editor's Notes
Since most students in these two groups are under the drinking age, they are likely to underestimate their drinking either out of caution or simply because they are rule abiding. In addition to this, underclassmen are probably more comfortable going to Barone and other on-campus locations for food rather than ordering take out. This would result in data that would favor those who order less take out (underestimates the response variable).
People surveyed that were visiting from other dorm buildings is a sampling bias. This is due to the fact that they could be upperclassmen. If they are upperclassmen they are more likely to drink and more likely to order take out because they do not visit barone as often. Overall this could overestimate the explanatory and response variables.
many students were not present. Being that it was a Friday afternoon, it is possible that those who were not at their dorms were instead out with friends getting ready to party that night. Hence, the students we did survey left out a group of students who party more often. This would indicate that the data we collected would favor students who party less/drink less (underestimates explanatory variable).
4. Another possible measurement bias that we identified is how people responded when they were with their friends. Even though the survey was on an ipad/iphone, some people spoke out loud when answering the questions when they were in a group. Therefore they may overestimate their drinking in order to impress their friends.