I have been a researcher and biostatistician for over a decade. I have written grants, papers and many statistical reports, yet I never thought about starting my own blog and talk about what I am good at. But we are in the internet era and it seems that you are nobody unless you have your own blog, so here I am . I apologize in advance if I am not the best writer, after all numbers are my thing 🙂
I thought a lot about what the title for my first post should be and I came up with one that I believe reflects pretty well why I became a statistics tutor. During my many years tutoring students one of the things that I’ve enjoyed the most has been to teach them to love statistics. Well, maybe for some of them, if not the vast majority, if I couldn’t get them to love statistics I got them to ‘hate less’ or at least to ‘fear less’ the stats.
My purpose with these series of posts is to bring statistics closer to people like you, who come from another discipline and don’t necessarily have a strong background in statistics. I hope that my advice will help you to lose this fear and be able to run your own analysis or at least know which are the appropriate methods given the nature of your data and your research questions.
I will start with my first piece of advice.
When do you need to start working with a statistician? The answer is as soon as possible. The optimal time to get the statistician involved is at the start of the development of your research proposal. The statistician will help you decide your study design, sampling strategy, make sure your data collection instruments are valid and finally put together a data analysis plan with the appropriate statistical methods given your study hypotheses.
I’ve had many clients come to me with a research proposal with the wrong statistical methods and incorrect sample size calculations (or what is worse, a sample size that has been arbitrarily chosen). Sometimes a client has come to me after data has been collected following a flawed study design and when it’s too late to fix it. Even though in these cases the statistician can still help you mitigate its the negative effects I suggest not to wait to get the services of a statistician until you are in such a desperate situation.