For a SIP with publicly available data, one suggestion is to look at college tuition rates across the country. I would look at schools and compare different tuition rates based on their location, size, division I II or III, private or public, etc. I would use collegeboard.com to find this information. I would then try to randomly pick regions throughout the United States and look at a certain number of colleges in each area. Next I would explore some hypotheses such as: private schools are x amount more expensive than public schools, or a certain region is more expensive than another region.
For a SIP with original data I would use the topic of alcohol. I would ask students how much money they spend on alcohol in a given week, if they are a male or female, how much they study per week, what their GPA is, what their family income is, if they have a job or not, etc. Some possible hypotheses to test would be: Men spend more money on alcohol than women, those who study more spend less money on alcohol, students who have a job spend x amount more on alcohol than students who don’t have a job. A possible sample would just be randomly selected students here on campus.