There is a misconception with statistics that we need to clear up. As statisticians, we do a lot of ‘number crunching’. I am not ignoring the fact that most of my time is spent getting the average of things and evaluating the the behavior of the data. But some of the work that we do is beyond the actual calculations. There are statistics concepts in your research question from the start. When you develop your quantitative methods section, you must know what is statistical feasible. If your dependent variable is collected at two different time points from the same person- that is a paired t test. If you plan on analyzing the score of a survey, you need to know how the survey will be scored in the first place. So you really need to identify the statistics as you formulate the research questions. And that is why some advisers will tell you to hire a statistician early the process. Because a statistician will help you identify how you will collect the data; who will collect it; and how you will answer the research questions. This is before you actually crunch numbers. Now there are professions where your focus is the calculator. Statistics for a dissertation proposal or the method section of a publication must be articulated in every day language. If I tell you that my groups comparison will be conducted on distinct individuals – I am also meeting the assumptions of my statistics tests. If your statistics requirements are interwoven in the non-statistical sections of your report- it ensures that the experiment will go smoothly. It is not easy to identify what language will have an affect on your stats- that’s why you hire an expert. Your statistician should ensure that your research question and methods section is solid. And if your pre-analysis work is legit- you should have NO problem from your committee.
-Moore to follow Amy