Q: I’m comparing platelet count in obese women with and without Type 2 diabetes. I have run t-tests, so I’m reporting means in my research paper. Should I report standard deviation (SD) or standard error (SE) along with means?

SD reflects sample variability, while SE is an estimate of how close your mean is to the true population mean. Both are useful measures, but your readers will need to know how much your data is dispersed from the mean. High variability suggests that individual data points deviate widely from the mean, indicating potential heterogeneity in the sample. This information helps readers gauge the reliability and generalizability of study findings. If a sample has low variability, the mean is a more robust descriptive statistic of the dataset. On the other hand, high variability may signal that the mean might not be as reliable, and that tests using means (e.g., t-tests) may not be appropriate. Hence, it is necessary to report SD along with means in your paper.

Barde and Barde (2012) have explained the use of SD vs SE in detail. The SAMPL (Statistical Analyses and Methods in the Published Literature) guidelines explicitly recommend against using SE to indicate variability in a dataset.

Would you like personalized guidance when using descriptive or inferential statistics? Talk to an experienced biostatistician under Editage’s Statistical Analysis & Review Services.