the following is the command I used and the results I got for my question, whether the visitation frequency of my bee is different in different experiment types in different locations. I used the lme function of R. I used experiment type as the fixed effect and location as the random effect. I used ANOVA after this to get the F value. Is this right? WIthout ANOVA how I can interpret the results? I had 5 experiment types, but in results it is showing only 4 experiment types. The top most one (expt.antless) is missing.
> names(acera.freq1)
[1] "expt.type" "visit.freq" "location"
> model<-lme(visit.freq~expt.type,random=~1|location,method="ML")
> summary(model)
Linear mixed-effects model fit by maximum likelihood
Data: NULL
AIC BIC logLik
1065.928 1087.919 -525.9638
Random effects:
Formula: ~1 | location
(Intercept) Residual
StdDev: 4.617241 4.682169
Fixed effects: visit.freq ~ expt.type
Value Std.Error DF t-value p-value
(Intercept) 10.192564 1.216852 148 8.376177 0.0000
expt.typeblack.ant -3.579023 1.074261 148 -3.331615 0.0011
expt.typecrazy.ant -5.804671 1.740132 148 -3.335765 0.0011
expt.typeother.ants -5.352438 1.936756 148 -2.763610 0.0064
expt.typered.biting.ant -2.680195 2.048081 148 -1.308637 0.1927
Correlation:
(Intr) expt.typb. expt.typc. expt.typt.
expt.typeblack.ant -0.173
expt.typecrazy.ant -0.128 0.126
expt.typeother.ants -0.081 0.146 0.027
expt.typered.biting.ant -0.132 0.117 0.143 0.046
Standardized Within-Group Residuals:
Min Q1 Med Q3 Max
-1.65727163 -0.70522770 -0.02959964 0.53356588 3.28792891
Number of Observations: 171
Number of Groups: 19
> anova(model)
numDF denDF F-value p-value
(Intercept) 1 148 56.51744 <.0001
expt.type 4 148 6.28705 1e-04
Best Answer
In short, "expt.antless" condition was used as a reference.
For example, based on the summary table, you can expect that (1) visiting frequency (visit.freq) to decrease by an average of 3.58 from "expt.antless" condition to "expt.typeblack.ant" condition, (2) visiting frequency to decrease by an average of 5.8 from "expt.antless" condition to "expt.typecrazy.ant" condition, and so on.