Students' Datasets

Ryan's question on dichotomization (01/13/2006)

Ryan asked this question: Here's something that sounds trivial, but I think it would be helpful to have a formal explanation of the proper procedure by which to do it. How do you dichotomize data that has a bimodal distribution? Here's the raw de-identified data and a histogram from STATA. I was just drawing a line at 0.6 and putting those on one side of the line into "language poor" and those on the other side into "language normal".

Bob's question on two groups of changes (01/27/2006)

Bob asked this question: Here is some very simple data looking at morphine analgesia (total duration of response) in control animals and opiate receptor lesioned animals (toxin). A simple question that can be asked is: does morphine attenuate pain differently or the same between the two groups?

• Chun's first journey to answering Bob's question.

Judith's survey data from the Emergency Department (02/06/2006)

Note: The research is ongoing and all handouts for this project are confidential.

Judith is doing a survey study: My project is to implement a reminder system in the Emergency Department (ED) at Vanderbilt. Prior to implementing the system, we surveyed the physicians and nurses to find their attitudes about vaccination in the ED and their opinions on how a reminder system should be implemented. All of the surveys had a unique identifier so we could distribute a second copy of the survey to people who did not turn it in. We submitted an abstract to the 2006 Society for Academic Emergency Medicine conference and we are working on a paper using the data.

I will hand out three items in the class. The first one is the survey we distributed to ED physicians and nurses. The survey was slightly different for nurses, but only on one or two questions. All of the survey information (including the actual questions), should be in the first tab of the Excel document. The second one is the submitted abstract for the conference. The third one is a data CD in Excel format. The Excel file has a total of 5 tabs in it:

1. Survey Info: This has the question scale, the number, and the actual question.
2. MD Summary: This links to the raw data, but it contains counts of what was answered for each question, how many surveys were returned, etc. This is all the data in a kind-of readable format.
3. RN Summary: Same as MD summary, but data for the RNs.
4. MD raw data: This contains all the data from the access database. Each physicians has a unique ID, and then I have everything they answered (or left blank). The bottom of the page has some descriptive data (averages, counts for each answer, etc). It's where I link to on the summary page.
5. RN raw data: Same as MD raw data, but for the nurses.

Sylvie's study on inflammatory process and oxidative damage (02/20/2006)

Sylvie wants to examine indicators of inflammatory process and markers of oxidative damage in steady-state HbSS adolescents and healthy controls, and their impact on resting energy expenditure. See here for more background and Sylvie's data analysis summary, and the data.