Daisy Fang, Entry #5, It's A Wrap!

It is still hard to believe that I finished my last day in the Arts and Mind Lab. I have been very busy the last two weeks of my time here as we dived into data analysis. While we continued to recruit and test the Bridge Boston control group, data collection for cohort one after a year of music training was officially concluded. In order to prepare for data analysis, I spent a few days watching Khan Academy videos to familiarize myself with the basics of statistics. Furthermore, Alessandra and I cleaned and organized the cohort one data into an excel spreadsheet which compares the first and second testing scores of each measures so that the data was ready to be entered into SPSS.

On Monday, Jill, Alessandra and I met with Dr. Winner to begin the long-awaited process of data analysis. During the meeting, Dr. Winner patiently explained to me that we were going to run a repeated measure ANOVA analysis, since each participant was tested with the same measures between two time points. We were interested in seeing the interaction of time point and group, as we hypothesized the control group score to stay the same while the treatment group score improves over time. It was really nice of Dr. Winner to spend time and walk us through the detailed the process of using SPSS using one of the six measures as an example. After the meeting, I was able to run the complete analyses on all measures and create a summary of the results. 

Summary of tests of within-subject contrasts of the Dot Counting task. Timepoint*GroupCvT is the interaction we are interested in, but unfortunately the significance level is not high enough.

Mean score graph of the Dot Counting task

In the repeated measure ANOVA that we performed, group (treatment and control) was identified as the between subject factor, and time point (K2 and 1^st grade) the within subject factor. Although the mean of each group changed in the expected direction for some of the measures, some changed in direction contrary to the hypothesis. No interaction between group and time point significant enough was shown, as the significance level of all measures were above 0.05 which indicates a 95% confidence interval. The graph shown above is the analysis of the Dot Counting task, testing working memory as a measure of executive functioning.

Initially, I was frustrated that after all the hard work, analyses showed that the data turned out to support the null hypothesis. But then I realized that we only analyzed the data for one cohort (half of the sample size) and after only one year of treatment. Dr. Winner pointed out that it is difficult for transfer studies to show results. She also taught me that null findings are also extremely important because it defies what is assumed to be true. Although many people claim that music education helps children improve executive functioning, the argument is false as concluded by the study so far. 

Looking back on the six weeks I spent at the Arts and Mind Lab, I have learned so much more than just how to collect and analyze data. The long data collection period may get tedious, but it is crucial in ensuring the validity of the results. I also realized that the essence of science research is not trying to prove the hypothesis correct, but to thoroughly explore the question with an open mind. I will miss working with the undergrads and grad students every day, and I hope to come back and visit in the future!