The Need for Flexibility and Adaptability in Research

When I first posted on this blog, I was expecting to start working with our survey data in mid-June, but as I have now found out again, things don’t always go as planned. With Professor Rapoport, as well as with Justin DBK and Ani-Rae Lovell, we have created a survey that will be sent out to Tea Party Movement members and supporters of the multiple TP organizations. Rapoport was able to obtain a good contact within the TP, which we were going to use to send out the survey from the Tea Party Express headquarters, likely increasing our response rate of the survey due to the legitimacy of this TP organization. However, this connection proved difficult to obtain due to the long chain of people involved, and the necessary timing of our survey. We will now send out the survey as soon as possible through another TP organization, Tea Party Nation. In the meantime, Rapoport has received the data set for a survey in which he participated, which included over 45,000 respondents. This survey included a couple of TP questions, from which I will analyze the data and draw conclusions about political activity in the TP Movement based off of past political involvement in the Republican party, for example, with Perot in 1992.

To analyze this data, Rapoport has reduced the data set to 10,000 cases and has dropped some of the less useful variables. I am now in the process of coding the variables, and trying different tactics to obtain the information I need, such as weighting the data by different variables so as to account for interaction effects or pre-existing bias amongst identifying groups (gender, race, income level, etc) so that I will have the most accurate account of political activity at the national level. Weighting the variables allows me to generalize my conclusions and apply and compare them to other studies.

I have been using SPSS for all of my statistical analysis, a program which I have previously used for Rapoport’s research. As this data set is so large, it has proven more difficult than I expected to use SPSS because I have more variables to account for, to weight, and to analyze. I plan to isolate the TP and identifying questions by themselves and create a new data set that is simpler to work with and therefore will hopefully result in less error on my part. In the next week, I am going to continue going through this data, and in the meantime, the original TP survey that we have created will be sent out (we hope)!