STAT 152
FALL 2006
SAMPLING SURVEY
TEAM II
GROUP V
FINAL WRITE UP REPORT
Analysis of ENGINEERING and Art&Humanity
Upper-class Students Exercise Habbit
By : Nicki Wells
David Tsay
Wing Yin Leung
CONTENT:
Purpose of the survey.
Summary of findings and conclusions.
Brief description of the survey methodology.
Detailed description of findings
Discussion of problems encountered
Conclusions and suggestions
Appendixes: group reports
1. Introduction
A topic at the forefront of both social and political discussion in 2006 is that of America’s rising obesity levels. Experts attribute this serious health risk to a number of causes, with low activity levels being one of the primary reasons for America’s weight gain in the twenty first century.Not only does regular exercise assist in the maintenance of a healthy weight, but research at TuftsUniversity revealed that students who were exercising at least three days a week reported feeling better physically and were also happier than those who didn’t maintain regular physical activity1[1]. The college years are imperative for developing healthy habits during the transition from youth to adulthood.
College students, with classes, homework, clubs, internships, jobs and social activities don’t have a lot of time for exercise. However, the United States Department of Health and Human Services has recommended that individuals get at least an hour of moderate to vigorous physical exercise every day.
For this reason we wanted to investigate the amount of exercise upperclassman students belonging in either but not both of engineering and arts and humanities programs at the University of California–Berkeley were getting each week. For our purposes we classified upperclassmen as those students of junior or senior standing. In particular we were interested in comparing the differences seen between those individuals who had declared themselves as Engineering majors and those who were Arts or Humanities majors. A common stereotype of engineering students is that they are so busy and involved in their studies, that they have little time for anything else. In fact a survey conducted by the National Science Foundation found that 29% of adults thought that scientists had “few interests outside of their work” and Gibbons et al. concluded that this figure would be even higher if the respondents were asked specifically about engineers[2].
Our data showed that the mean level of exercise for all students surveyed was below the daily sixty minutes recommended. Evidently the reason for this discrepancy is not unawareness, as most students also reported that their exercise levels should be higher, and that if they had more time, they would exercise more. Another important factor which appeared to influence the amount of time that students spent on physical activity was the distance that the student lived from campus.
Our study also served to dispel the myth that engineering students think about nothing more than their classes, as the overall difference between the physical activity levels of the UCB juniors and seniors engineering students and their arts & humanities counterparts were minimal. It was however of interest to note that differences were evident when comparing males and females within the majors. Female engineering students appeared to take part in more than their male counterparts, while female arts and humanities majors on average did less than males in this major.
2. Methodology
The sampling method
Stratification (Engineering, A&H) , 1 Stage Cluster Sampling (SRS of classes with equal probability, unequal class size )
We had stratified our population between students declared for Arts/Humanities and Engineering majors. The number of students in College of Engineering is and =1802 for College of Arts & Humanities.We used information from the 2005 UC Undergraduates ExperienceSurvey[3] to calculate the sample size with error=0.1.We had sample size for the stratum of engineering majors is
And the sample size for the stratum of arts & humanities is .
The questionnaire design.
As we mentioned above, there are several majors that we were interested to compare and find out the difference. Our major focus in this study was engineering and arts & humanities students in Junior or Senior Standing. Also, we want to avoid for the duplicate record and make sure the dataset is clean. In order to avoid for that issue, the first question in the questionnaire is asking if the student has filled out the survey before. We would exclude the student who filled it out before; also we want to exclude students who are double majoring in Arts & Humanities and Engineering since it would be hard to categorize them. In order to find out the student’s average exercise hours, we would ask an additional question if the amount they stated the same with their average weekly level. We set up questions about gender, age and type of exercise to find out the correlation between Gender and Exercise, Age and Exercise. For example, the percentage of male would skew higher than female in Engineering majors. It would be interesting to know if this factor would strong enough to offset the average in all engineering students. Also we included few more auxiliary variables to find out the correlation between some life factors and exercise.
The analysis method.
In analysis methodology, we apply the Bootstrap method to estimate the mean and standard error for both groups of students with the minimization of the variances. We also applied some statistical hypothesis test to find out the results, as T-test, ANOVA, Chi-Square Test and Non-Parametric test. In order find out the significant factors among all the variables, we perform a linear regression analysis and distinguish the correlation between all variables and the significant factor.
3.Findings
3.1 Background of the sample population:
3.2 Difference between Engineering upper class students and ArtsHumanities upperclass students
Estimates / Engineering / Arts & Humanities(SE)
[95% CI] / 3.269523
(0.279297)
[2.722 , 3.817] / 3.128855
(0.1475469)
[2.840 , 3.418]
Based on the common stereotypes of engineering students that they are so busy and involved in their studies and that they have little time for anything else, we actually have the findings which are against these common stereotypes.
The estimated Exercise mean for Engineering student is 3.27 and 3.13 hours a week in Arts & Humanities majors. The means for both majors is relatively similar.
Chi-square Test:
From the Chi-square Test(Independence of hours and majors), based on the p-value =0.89 we would say that the number of hours students spending on exercise is independent of the majors.
From the Chi-square Test(Independence of hours and gender), based on the p-value =0.7292 we would say that the number of hours students spending on exercise is independent of the gender when we combine the gender together from Engineering and Arts Humanities.
There is no relationship between the number of hours spends on exercises with majors and the general gender.
However when we look at the gender within the strata, we would suspect that there exist variability for exercise hours between males and females within the strata.
Cross Tabulation / Engineering / Arts & HumanitiesMale / 3.103287
(0.3956734)
[2.33 , 3.88] / 3.457267
(0.5406976)
[2.39 , 4.51]
Female / 3.968828 (1.752258)
[0.53 , 7.40] / 2.747019
(0.3064672)
[2.15 , 3.35]
ANOVA Table
Source / df / Sum of Squares / Mean SquareBetween Strata / 1 / SSB= 1.56 / MSB = 1.56
Within Strata / 427 / SSW= 4933.8 / MSW = 11.55
Total / 428 / SSTO= 4935.36 / MSTO = 11.53
From the ANOVA table, we observe that most of our variation is within each stratum.
3.3 Most students are aware that they should be doing more exercise
From the graphs below, we can see that most of the students notice the problem of lacking exercise; they claim that they had less exercise in the week they took the survey than the previous week. Students who are willing spend more time on exercise is the majority in this survey. In the next section, we would see the factors which affect the amount time they spend for exercise.
3.4 Factors that affect the amount of time on exercise
This table is to find out which factor (midterm, family matter, injury, others) affected the hours of exercise the most.
Distance between fitness facilities and home is also a factor affecting the amount of hours student spend on exercise.
People living close to the RSF tend to exercise more. Art & Humanities students tend to live further from the RSF. We can speculate that A&H students tend to live further away from school. This is consistent with our finding earlier that Engineering students has a little higher average exercise hours than A&H students.
Finally,personal preference and habit are also factors. People who enjoy exercise tend to exercise more.
4. Discussion
There were definitely some problems and imperfections encountered throughout the entire surveying process that could potentially affect the data and alter our findings. It is not surprising that a lot of problems arose during the data collection process; it is often quite difficult and challenging to actually carry out a sample survey in real life, especially for our team made up of students in a college statistics class with little or no real life survey sampling experience.
First of all, our team decided to split up the work of surveying in classes so that each student of Team 2 would go to a classor two with questionnaires and do the surveying. As a result, our team had about twenty members who did the surveying and each member might and probably have done something differently than the others that could affect the consistency of our final data. Some members might have made and persuaded the students of their surveyed classes to take the questionnaires more seriously than what the other members might have done. Also, some members tried harder than others to get more complete sets of filled-out questionnaires for their classes and even went back for a second time to get the students whom they missed the first time to do the questionnaires. Not every member of our team who surveyed did this, so this distorts the consistency of our data collecting process. Also, a potential problem during the data collection process is that some students might be absent for their classes and so the samples that were surveyed might all be slightly biased toward the students who do go to class and willing to respond to our questionnaires.In addition, another potential problem during the data collection process is that the surveys might be done in a very short amount of time since either the professors probably don’t want to use up too much of their class time or that the students themselves might not want to spend too much of their time doing them; consequently, the surveys might not have been done completely and well. As a result of this, some items of the surveys might be left blank and some students might only answer the first few questions of the surveys. This results in what is called an item nonresponse in which some measurements are present for the observation unit but at least one item is missing; the data collection team tried to overcome this by filling in the missing items with other similar surveys with similar items. Similarly, another problem during the data collection process is that a few professors simply just refused to let the surveys be done for their classes. Therefore, this results in what is called an unit nonresponse in which the entire observation unit is missing; the data collection team tried to overcome this by using data and surveys from similar classes such as classes from the same department to fill in for the missing data and surveys. In general, it would be unwise to simply throw out missing observation units and items because that would change the weights of the sampling and could potentially create bias.It is good that the data collection group used imputations such as deductive imputation and hot-deck imputation to fill in the missing data for the item and unit nonresponses. Overall, the data collection group did their job well; also, the entire process serves as a real-life learning experience for each member of our team who participated in the surveying of classes.
Furthermore, there were also some problems and difficulties that other groups of our team encountered that should be addressed here. First of all, for the data analysis group, they simply just had a lot of data to analyze but it was somewhat difficult for them to analyze the data for some specific characteristics in some particular directions because the data analysis group just simply did not know who the target audience this survey sample will be for; it is up to the write-up group to decide who the target audience is. As a result, the data analysis group just did a lot of general analysis and a lot of them might not turn out to be useful or meaningful for the write-up group after they decided who the target audience would be. Nevertheless, the data analysis group did their job well for the information they knew and the data they were given. In addition, for the questionnaire design group, they have only a few features of the questionnaire such as using a scale from 0 to 6 for the students to show how much they enjoy exercising that could be inherent areas for inconsistency of responses because a student’s interpretation of a “4” on a scale might actually be another student’s interpretation of a “5” on a scale. Nevertheless, overall they designed the questionnaire quite well for the questions are all specific and generally ask for clear-cut answers. Moreover, for the sampling plan design group, ideally what they would have probably liked to do was to have a larger target population including all registered students at UC Berkeley and not just the UCB junior or senior undergraduates declared either for arts & humanities or engineering but not both programs. Having a larger target population might enable the survey sampling to draw stronger conclusions at the end that apply to more people. However, there exist difficulties making all UC Berkeley students the target population for it would be very time consuming; also, a lot of freshmen and sophomores do not even have majors so it would be hard to use their responses to compare the various kinds of data and analysis between engineering and arts & humanities students. Thus, perhaps it is better and more efficient to design the survey for a smaller target population but to do it very well and draw solid conclusions from it. The sampling plan design group did a good job for the task that they were given at the time they came up with the sampling plan.
5. Conclusion
After discussing the various problems that the groups and our team encountered, there are definitely room for improvement and suggestions for the future when these kinds of survey samples are done. First, for the survey sampling design group, it would be better if they could be the ones that know or come up with the target audience and the specific purposes of the survey sampling so that they can perhaps come up with an even better survey sampling design that would result in having fewer problems later on with the other groups. In addition, the data analysis group would also benefit from the survey sampling design group knowing or deciding the target audience early on so that the data analysis group would know specifically what to look for and perform specific analysis to test out the issues concerned by the target audience. As for the questionnaire design group, they could improve upon their questionnaire by perhaps shortening it a bit and yet keeping all the most relevant questions because it is just not realistic to expect the students to fill it out completely given a short amount of time. As for the data collection group, they would do better if they have fewer people surveying to make the process more organized and also having those few specific people well trained so that they collect the data in very similar ways to ensure the consistency of data collection.
Overall, there are variousconclusions that can be made for the UCB junior and senior undergraduates declared either for Arts & Humanities or Engineering but not both programs.First of all, the sampling survey has data supporting that the UCB juniors and seniors female Engineering students appeared to exercise more than their male counterparts while the female Arts & Humanities majors on average exercise less than their male counterparts. Next, having midterms might have an association with the UCB junior and senior Engineering and Arts & Humanities students exercising less during those exam times. Also, the students who live closer to the RSF tend to exercise more and the students who simply enjoy exercising do exercise more as well. In addition, it can be seen that UCB junior and senior undergraduates declared either for ArtsHumanities or Engineering but not both programs just do not exercise enough even though they know this and know that they should be exercising more. Moreover, despite the stereotype that Engineering students are more caught up with their academics and have less time and exercise less compared to Arts & Humanities students, it has been found that for the UCB juniors and seniors Engineering and Arts & Humanities students thatthey actually exercise similar number of hours per week. Thus, these UCB juniors and seniors Engineering students should not use their academics as excuses for not exercising enough for if they really want to exercise, they would be able to find and make some time for it.In fact, despite that these conclusions are only made for the UCB juniors and seniors Engineering and Arts & Humanities students, other people such as all UC Berkeley students in general and even students from other colleges can also take a look at the data and the conclusions of this survey sample and perhaps learn some things for themselves even if they are not part of the target audience and even if the conclusions are not drawn for them.