One potential SIP study would be to examine the factors that go into voter turnout of presidential elections. An initial regression could be set up with voter turnout being a factor of exit poll results regarding major issues, overall campaign spending, number of voting age populace registered to vote and voting age populace demographics (age, race, income, gender etc). Within these categories, there are a number of interesting hypothesis tests that could be done. For instance, one could ask if there is a significant increase in voter turnout when greater amounts of money is spent campaigning; if having the economy be the top issue in exit polls drives more people to the polls than national security; and if there is a statistically significant difference between the proportion of college age students that turned out to vote in the 2008 election as opposed to other years.
Here are some sites that could prove useful:
The American Presidency Project: http://www.presidency.ucsb.edu/data/turnout.php
The Roper Center's U.S. Election page: http://www.ropercenter.uconn.edu/elections/presidential/presidential_election.html
U.S. Census Bureau: http://www.census.gov
Pew Center Election Turnout: http://www.pewcenteronthestates.org/ttw/trends_map_data_table.aspx?trendID=19
A possible international spin on this idea would be to find these same data sets for another industrialized nation such as Great Britain or France, and see how the information compares.
My second proposal looks at the effect of the Kalamazoo Promise on teachers and curriculum in middle and high school in comparison to other school districts in communities with similar demographics. Although the Promise has been in effect for a number of years now, there is still relatively little knowledge about its effects. Of the information available, much attention is paid to its effects on students. However, it seems as though little research has been done to see how a program such as the Promise might affect teachers. Therefore, this proposal provides the chance to make original findings that have relevance both locally and nationally.
Methodologically, a stratified random sample of teachers in Kalamazoo Public Schools at the middle and high school levels would be administered a survey that would cover a range of issues including, but not limited to, the extent to which teachers talk about college with their students, personal fulfillment with regards to occupation, number of hours spent preparing outside of the classroom, amount of interaction with parents, and amount of classroom time spent doing college prep. This same survey would then be randomly administered in two or three other school districts within cities with demographics fairly similar to Kalamazoo. A range of hypothesis tests could then be done to see if the Kalamazoo Promise substantially increases the results for teachers in KPS with regards to each of these questions. One could also see if those changes are more likely to occur within groups of teachers that teach certain subjects or have taught for a certain number of years.