The job requires an experienced statistician (at least two years experience), preferably someone who has worked in the social science field. Candidate must speak English as a first language.
I'm conducing a experiment with two groups of people. The field of study is adult education. One group will learn using a standard tutorial, and one group will learn using a tutoral that includes simulation-game elements. The general material for both groups is the same. Both groups will take a pretest and posttest. Thus, the purpose of the experiment is to see whether one group outperforms the other group on pretest/posttest differences. I'll also collect demographics like age and gender to see if there are important differences. Preferably, the statistics will be done using Excel.
In short, the contractor will essentially determine ALL statistical techniques to be used on this study. This will include indicating the desired sample size, the particulars of implementing double blind implementation, the particulars behind getting validation by experts that the tutorials and test questions have subject matter validity. It's critical that the contractor be able to reference a recent university-level textbook (within the last ten years) and corresponding page numbers in that book that specifically indicate WHY the recommended techniques are valid for the study. Thus, a Web site recommendation would not be acceptable. The contractor must also produce a spreadsheet of mockup data showing the outcomes of the statitical techniques.
I see the general work taking place as follows:
1. Initial conversation to brief the contractor
2. Contract makes statistical recommendations
3. I write the report documenting these choices.
4. Contractor verifies these choices in my document.
5. Contractor creates an Excel spreadsheet including mock data to show the outcomes of the techniques.