Data: structure(list(record_id = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1 - Pastebin.com
I wrote the following mixed model:
lmer(cumulative.wk.dose ~ poly(date,2,raw=TRUE)*bin + (1|record_id), data=df)
There are 52 days (date) and 3 bins (high, medium, and low). I want to compare the cumulative.wk.dose at date 52 for the 3 bins with emmeans but cannot figure out how to code it. I want to get confidence intervals and p-values for the high, med, and low bins at week 52.
The closest I got was
emmeans::emmeans(lmm.2, ~ bin|date,
cov.reduce = FALSE)
Which returns the following for every single date. How can I just return date 52?
date = 1:
bin emmean SE df asymp.LCL asymp.UCL
high 767 303 Inf 174.3 1360
med 428 206 Inf 25.5 831
low 133 316 Inf -486.6 752
date = 2:
bin emmean SE df asymp.LCL asymp.UCL
high 1078 301 Inf 487.6 1668
med 631 205 Inf 230.3 1032
low 274 314 Inf -341.8 891
...
date = 52:
bin emmean SE df asymp.LCL asymp.UCL
high 9924 303 Inf 9330.6 10517
med 6628 206 Inf 6225.5 7031
low 5451 316 Inf 4831.5 6070
and then using emmeans:pairs() I can return
date = 1:
contrast estimate SE df t.ratio p.value
high - med 339 366 107.8 0.927 0.6247
high - low 635 438 107.8 1.450 0.3190
med - low 296 377 107.8 0.784 0.7136
date = 2:
contrast estimate SE df t.ratio p.value
high - med 447 364 105.6 1.227 0.4401
high - low 803 435 105.6 1.845 0.1601
med - low 357 375 105.6 0.951 0.6094
...
date = 52:
contrast estimate SE df t.ratio p.value
high - med 3295 366 107.8 9.008 <.0001
high - low 4473 438 107.8 10.223 <.0001
med - low 1178 377 107.8 3.123 0.0064