Читать книгу Statistics in Nutrition and Dietetics - Michael Nelson - Страница 51

It is your job as a scientist to summarize data in a coherent form (in tables, graphs, and figures), tell an interesting story about the relationships between the variables you have measured, and interpret the results intelligently for your reader, using appropriate statistical analyses.

Of course, you need to keep accurate records of observations, and make sure that your data set (spreadsheet) is stored securely and that you have backup copies of everything. Bulking up a report with tables of raw data is bad practice, however. No one will read them.

Chapter 15 provides lots of examples about how to summarize data to make the presentation of results interesting. It also shows how to present results according to the type of audience you are talking to. If I am presenting new results about the impact of school food on attainment to scientific colleagues, I will include lots of information about the methods that I used to identify my samples, make observations, and analyze the data, as well as details about the findings themselves. My scientific colleagues will need enough information to be confident that my data are unbiased, that I have used the right analytical approaches, and that the results are statistically significant. This is the same basic approach that I will take when I am writing a paper for submission to a peer‐reviewed journal. In contrast, if I am presenting results on the same topic to a group of teachers, I will use lots of graphs and charts to summarize the results in a way that tells a good story. The underlying data will be the same, of course. But the teachers are likely to be bored by too much detail about methods – they probably just want to know the headline about whether better school food has a positive impact on attainment, and how big that impact is likely to be.