***** FRIDAY (FUN)1 FACT ****** 1 Fun to follow in coming weeks. It’s a common question that gets asked: What measure should I use in my error bars? If you also struggle with that, here’s a recap of two of the most common options, and when to use them. Standard deviation (SD) is a characteristic of a distribution – it is specific to the observations that you have in the data, and is used to describe or summarise their qualities. It tells you how far your observations are from the mean value. There is a population standard deviation (σ), and a sample standard deviation (s), where the former assumes you know all the observations there can be, and the other assumes you have a sample of them – both are still measures of dispersion in the observations. If you are running experiments, you quite certainly deal with the latter. * It is an acceptable measure of dispersion when showing the descriptive statistics of your raw data. Standard error (of the mean) (SE or SEM) is a characteristic of an estimate. It is the variation of the sample means, and tells you how far the “true mean” of the trait is likely be from the one you measured, and within what range you would expect the mean to be if you repeated the sampling. SE tells you what is the potential error in the results. * It is an acceptable measure when you want to infer beyond your sample, and most definitely when you show something that comes from a statistical model. Curiosity: Danish Statistics text books do not have a Danish term for standard error. Sometimes “spredning” is used, which refers to all dispersion parameters (SD, variance, SE, etc.), and as such is imprecise. When specifically detailing SE, Danish text seem to refer to it with the English term. Danish speakers, if you own a text book that says otherwise – please do correct me! [cid:image002.png@01DBE763.5535CF50] Maarit Mäenpää Statistician Dept. of Agroecology Aarhus University Blichers Allé 20, Postboks 50 DK-8830 Tjele Web: agro.au.dk Phone: 51 335 104 E-mail: m.maenpaa@agro.au.dk<mailto:m.maenpaa@agro.au.dk> AGRO employee? Book a statistics consultation: https://tinyurl.com/4tvesj3u [AU Logo] ”Please note that this email may contain personal data. You must ensure that this data cannot be accessed by anyone else without good reason, and that it is deleted immediately when it is no longer required in relation to the purpose for which it was sent”.
