Week 7

With the meeting last week with John Charlton, we had found out what tests to perform, but not really the main reasons behind them. On Tuesday we had a meeting with Stephen Manning, who explain to us the kinds of tests you usually do with statistical analysis. One of these mentioned was the Anderson-Darling normality test usually done to ensure the T-test data for the HR data was normal, which we had not yet done.

I ended up using software called Minitab to produce the normality tests, and that showed us that the 20+ hardcore demographic was slightly skewed on their base HR, which after discussing it with Steve on the Thursday afternoon, we decided we would take him out due to the student admitting to having a heart-condition causing a very high heart rate.

After discussing the content of our team-report, it was decided that I would be doing the results section, as me and Nick are probably the two who best understand the statistics side of the report in the team. Matt and Nick went through the basics of this section with me, where we discussed the layout of the section to be written over the second week of the Easter break.

Week 6

This week was the final phase of testing, and additionally began to analyse some of the results we had obtained. During the afternoon I and Nick discussed with Dr. Roger Jackson some of the statistics we were using to analyse the data. Consequently we were informed that we should be using a slightly different formula than we had been using. The following day we focused on trying to work out the statistics of our data, and splitting people up into their separate demographic, although we didn’t make that much progress due to me being unsure of how to properly analyse the data. As a result we had arranged a time to meet with John Charlton, where we could discuss the statistical analysis.

On Thursday we had a meeting with John Charlton, where we explained our test method and John managed to take us through the process of using SPSS to help analyse the data from the SAM form and heart-rate, in which we use a non-parametric test (the Wilcoxon signed ranks test) to measure the Pleasure-Arousal-Dominance values of those who played the light and dark maps, with a paired T-test for the average base Heart Rate and those for the dark and light maps.

From these we managed to conclude that the data from the HR showed that it was significant to a 1/5000 chance between the dark and light map, and additionally the same for arousal and dominance on the PAD scale. The only one we can’t say is significant is the Pleasure, which could be due to several reasons, which we will focus on next week.