Log in
RegisterGeometric means ± SD [🇷 for BE/BA]
Dear Helmut,
We should bear in mind that when we back transform to the original scale we end up with asymmetry with respect to +/-. Therefore, it is IMHO not appropriate to do +/- back transformed standard deviation. Symmetry is however obtained with respect to * and /. So the error bars should be such that lower end = gMean / gSD and upper end = gMean * gSD, where gSD = exp(standard deviation from log-transformed values).
[Mathematical rationale: exp(m - s) = exp(m)/exp(s) = gMean / gSD and exp(m + s) = exp(m) * exp(s) = gMean * gSD), where m and s are mean and sd on log scale, respectively]
Best regards,
Ben.
❝ I suggest to plot geometric means* ± backtransformed standard deviations. Concentrations likely follow a log-normal distribution.
We should bear in mind that when we back transform to the original scale we end up with asymmetry with respect to +/-. Therefore, it is IMHO not appropriate to do +/- back transformed standard deviation. Symmetry is however obtained with respect to * and /. So the error bars should be such that lower end = gMean / gSD and upper end = gMean * gSD, where gSD = exp(standard deviation from log-transformed values).
[Mathematical rationale: exp(m - s) = exp(m)/exp(s) = gMean / gSD and exp(m + s) = exp(m) * exp(s) = gMean * gSD), where m and s are mean and sd on log scale, respectively]
Best regards,
Ben.
Complete thread:
- R code for analyzing classical 2X2 crossover designed bioequivalence data lizhao 2016-01-22 00:13 [🇷 for BE/BA]
![[Mix]](https://static.bebac.at/img/mix.png "open in Mix view")
- R code for analyzing classical 2X2 crossover designed bioequivalence data ElMaestro 2016-01-22 00:59
- R code for analyzing classical 2X2 crossover designed bioequivalence data lizhao 2016-01-22 01:26
- On random effects and bogus statements ElMaestro 2016-01-22 12:13
- FDA 'own' SAS code for 2x2 d_labes 2016-01-27 09:57
- FDA 'own' SAS code for 2x2 nobody 2016-01-27 10:09
- OT: PM to ElMaestro Helmut 2016-01-27 13:39
- On coffee and such... ElMaestro 2016-01-27 13:47
- FDA 'own' SAS code for 2x2 nobody 2016-01-27 10:09
- Geometric means ± SD Helmut 2016-01-27 13:51
- Geometric means ± SD nobody 2016-01-27 15:20
- Money makes the world go ’round Helmut 2016-01-27 16:11
- Money makes the world go ’round nobody 2016-01-27 16:31
- Fixed or random effect kumarnaidu 2016-11-04 14:05
- PROC GLM fixes Subject mittyri 2016-11-04 23:00
- PROC GLM fixes Subject kumarnaidu 2016-11-05 05:22
- PROC GLM fixes Subject mittyri 2016-11-05 10:13
- PROC GLM fixes Subject kumarnaidu 2016-11-07 06:24
- PROC GLM fixes Subject mittyri 2016-11-07 11:57
- PROC GLM fixes Subject kumarnaidu 2016-11-07 12:42
- PROC GLM fixes Subject mittyri 2016-11-07 11:57
- PROC GLM fixes Subject kumarnaidu 2016-11-07 06:24
- PROC GLM fixes Subject mittyri 2016-11-05 10:13
- PROC GLM fixes Subject kumarnaidu 2016-11-05 05:22
- PROC GLM fixes Subject mittyri 2016-11-04 23:00
- Fixed or random effect kumarnaidu 2016-11-04 14:05
- Money makes the world go ’round nobody 2016-01-27 16:31
- Money makes the world go ’round Helmut 2016-01-27 16:11
- Geometric means ± SDBen 2016-11-08 21:20
- Misunderstanding Helmut 2016-11-09 15:29
- Misunderstanding Ben 2016-11-14 19:40
- RTFM Helmut 2016-11-14 21:50
- Misunderstanding Ben 2016-11-14 19:40
- Misunderstanding Helmut 2016-11-09 15:29
- Geometric means ± SD nobody 2016-01-27 15:20
- FDA 'own' SAS code for 2x2 d_labes 2016-01-27 09:57
- On random effects and bogus statements ElMaestro 2016-01-22 12:13
- R code for analyzing classical 2X2 crossover designed bioequivalence data lizhao 2016-01-22 01:26
- R code for analyzing classical 2X2 crossover designed bioequivalence data ElMaestro 2016-01-22 00:59
Ing. Helmut Schütz