A2-ArielRokem
From CS294-10 Visualization Fa07
[edit] Assignment 2 - creating visualizations with existing software
I have chosen to look at the effects of yoga practice on the well-being of teenagers in juvwenile detention. I obtained the data from Niroga. The data was collected in the form of questionnaires answered by participants after they had finished their participation in the program. The questionnaires had questions about the participant, such as age, gender, the duration of their stay in juvenile detention and the frequency of their participation in yoga practice during that time. Additionally, the participants replied to two standard psychological tests, designed to measure self awarness, general health and stress. The data contained the score of each participant on each of the questions in both of these tests (40 questions in total).
The first question I wanted to answer is whether yoga practice has a dose-dependent effect on well-being as measured by these tests. I started by plotting the average on each question in each of the duration and frequency conditions in the data set:
What a jungle! There seem to be some questions that systematically change with the frequency of the training, within some cells of duration of the stay in juvenile hall, but it is very hard to make any sense of the data with so many small multiples. In the next stage I tried collapsing the many questions into several different scores - scores for each of the subscales on the IHS and one average score for the PSS. I replotted the above plots for that transformation:
It is still very hard to make any sense out of this data and hard to say whether there is any dose-dependence of the well-being on yoga training. In the next stage, I proceeded to calculate a dose measure which was frequency*duration. This gives a measure of the actual number of yoga sessions each participant had. This still produces a picture that doesn't quite answer my initial question, or gives the rather unsatisfying answer that the answer is that these measures of well-being don't seem to covary with the amount of yoga training an individual receives:
However, in this image I also noticed that the number of participants in each group varies very significantly. Also - the largest group are the participants who didn't receive any yoga training at all. Thus - I decided to aggregate all the results from particpants who had received any training vs. participants who had not received any yoga training during their stay in juvenile hall:
There seems to be no apparent difference. Does this hold when the group is divided according to gender?
It turns out that it does:
However, Tableau does not provide any tools to evaluate these differences in any sytematic way. Are these differences statistically significant? Can we say whether yoga helps? Not using Tableau. Back to SPSS...
