Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL Do 03.10.2024 00:00:03 190.334 -2 9.878.838 51,9 Mastodon 4.3.0... 500 Mi 02.10.2024 00:00:06 190.336 -2 9.870.374 51,9 Mastodon 4.3.0... 500 Di 01.10.2024 00:00:07 190.338 -3 9.863.625 51,8 Mastodon 4.3.0... 500 Mo 30.09.2024 00:00:08 190.341 -3 9.853.351 51,8 Mastodon 4.3.0... 500 So 29.09.2024 00:00:03 190.344 +1 9.843.604 51,7 Mastodon 4.3.0... 500 Sa 28.09.2024 00:00:03 190.343 -4 9.834.146 51,7 Mastodon 4.3.0... 500 Fr 27.09.2024 00:00:02 190.347 -2 9.824.358 51,6 Mastodon 4.3.0... 500 Do 26.09.2024 00:00:00 190.349 -4 9.813.984 51,6 Mastodon 4.3.0... 500 Mi 25.09.2024 00:00:00 190.353 -2 9.804.415 51,5 Mastodon 4.3.0... 500 Di 24.09.2024 00:00:00 190.355 0 9.794.803 51,5 Mastodon 4.3.0... 500
Paul Harrison (@pfh) · 04/2022 · Tröts: 172 · Folger: 149
Do 03.10.2024 06:08
Adding some speculation: For a model with one factor, a GLM will estimate means for each group on a linear scale, even when the model is then expressed on a different scale such as log. Large positive outliers can therefore have a large effect. Maybe it is better to be estimating means on a log(ish) scale, as the voom-limma method does.
[Öffentlich] Antw.: 0 Wtrl.: 0 Fav.: 0 · via Web