Improving your results section

I realise that result sections are very different depending on your specialty and your personal style. But I thought I would give tips as to what I have found makes for a “good” results section in any research piece that I have read/written. I currently write about research within a medical context. The people who read my research are generally clinicians with little to no statistical knowledge or who do not frequently read statistical methods, but who have a lot of interest in the findings and potential implementation of the results.   

Within psychology, I have experienced the tendency to “fish” around in the data to report the exciting and juicy significant findings, but this is often at the detriment to the overall quality of the research.

1. This is probably the single most important point. Always refer back to your research aim(s). This may sound like an obvious thing to do. But often, when your neck deep in figures and tables, you can get side tracked and forget exactly what it is you set out to do. The best results sections I have read are those that stick to what they planned to do. Not only is the study designed around this (i.e. appropriate sample size / measures/ time scale), but the final message is a lot clearer. A colleague of mine was asked to structure her thesis results in line with her research aims to make it clearer how they were related.
2. Be transparent and succinct in what you have done and what you are reporting. Ask someone (I often ask the other half who has no knowledge in my field) to read through and see if they can get the jist of the findings. Clarity is key.
3. Really look at the data you have. View it as much as you can, in as many ways as you can to really get a feel for what is happening. 
4. The rule of parsimony is often one best applied to the analysis section. Really think about whether or not you need to transform your data and what, in turn, that will mean the analysis means. Do you need to do all of those complex analyses when a simple one will suffice?The best results I have read are those that more often simply describe the data (with means, ratios, confidence intervals).
5. Think about who will read this. If like me, people who read your work are not frequently exposed to statistics, keep the results appropriate. The whole point of writing, is so that other people read it. Don’t put people off. Make it as easy as possible to read, reflect, and retweet!

  You got this! Mum, PhD