Humberto Dutra
2008-Jul-03 07:38 UTC
[R] post hoc comparisons on NLME for longitudinal data
I am trying to fit a non linear mixed effect model but I also want to do a post hoc comparison. My data is binary and consist of recording mice track prints on plates plates in plots that submited to one of 4 different treatments (fruits and vegetation complexity manipulated for two levels each. The design is random blocks repeated measures with presence or absence of track prints as a response variable. Treatment is fixeed, time is fixed, and block is a random effect. I would like to know how I could do a post hoc comparison to see which months had a significant effect of an specific treatment. I searche on Bates and Pinheiro book but they don't have an example of it.> names (kirk)[1] "Obs" "track" "time" "block" "treat" ##Obs is the track plate identity, track is response variable, time is the month in which it was recorded and treat is one of the four treatments.##> mice <- lmer(track ~treat*time+(1|block), family=binomial, method = "PQL",data = kirk) > miceGeneralized linear mixed model fit using PQL Formula: track ~ treat * time + (1 | block) Data: kirk Family: binomial(logit link) AIC BIC logLik deviance 956.9 1002 -469.4 938.9 Random effects: Groups Name Variance Std.Dev. block (Intercept) 0.58182 0.76277 number of obs: 1168, groups: block, 3 Estimated scale (compare to 1 ) 1.000414 Fixed effects: Estimate Std. Error z value Pr(>|z|) (Intercept) -1.39258 0.54124 -2.573 0.010084 * treatFIVR -1.69691 0.53920 -3.147 0.001649 ** treatFRVI -0.86542 0.49603 -1.745 0.081039 . treatFRVR -1.54753 0.59766 -2.589 0.009617 ** time -0.05365 0.08459 -0.634 0.525957 treatFIVR:time 0.43627 0.13169 3.313 0.000924 *** treatFRVI:time 0.23028 0.12637 1.822 0.068418 . treatFRVR:time 0.16126 0.15067 1.070 0.284502 --- Signif. codes: 0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1 Correlation of Fixed Effects: (Intr) trFIVR trFRVI trFRVR time tFIVR: tFRVI: treatFIVR -0.327 treatFRVI -0.359 0.359 treatFRVR -0.297 0.298 0.324 time -0.506 0.503 0.548 0.454 tretFIVR:tm 0.319 -0.906 -0.351 -0.292 -0.640 tretFRVI:tm 0.336 -0.337 -0.896 -0.304 -0.668 0.429 tretFRVR:tm 0.283 -0.282 -0.307 -0.899 -0.561 0.359 0.375 thank you for your help Humberto Dutra ========================================================='Discipline - Success doesn't just happen. You have to be intentional about it, and that takes discipline.' - John Maxwell Humberto Dutra ========================================================='Discipline - Success doesn't just happen. You have to be intentional about it, and that takes discipline.' - John Maxwell