What do the weights argument in glmer refer to? I used sample sizes as weights with glm, but here I am not sure. The variance of sample sizes is quite low, but including it or not in glmer gives me a huge difference. For example, in the dataset below, using only one independent variable, the difference in results is huge (estimate, BIC, p.value). Does anyone have experience using weights in glmer and confirm if it works as expected or if I am doing it right?
A thread on r-sig-mixed models and in github pages, there seems to be an issue with weights argument in glmer, but since my knowledge of mixed-models is only weeks old, I am not able to follow it.
My data:
cv <- structure(list(name = c("AlfF", "AndH", "AntH", "BerG",
"BerR", "FreZ", "GerB","GerT", "GueS", "GueV", "HanN",
"HeiW", "JakW","KarN", "KerG", "KlaS", "ManS", "MarS",
"SilN", "TheG", "UweP","WerT", "AlfF", "AndH", "AntH",
"BerG", "BerR", "FreZ", "GerB","GerT", "GueS", "GueV",
"HanN", "HeiW", "JakW", "KarN", "KerG", "KlaS",
"ManS", "MarS", "SilN", "TheG", "UweP", "WerT",
"AlfF", "AndH", "AntH", "BerG", "BerR", "FreZ",
"GerB", "GerT", "GueS", "GueV", "HanN", "HeiW",
"JakW", "KarN", "KerG", "KlaS", "ManS", "SilN",
"TheG", "UweP", "WerT", "AlfF", "AndH", "AntH",
"BerG", "BerR", "FreZ", "GerB", "GerT", "GueS",
"GueV", "HanN", "HeiW", "JakW", "KarN", "KerG",
"KlaS", "ManS", "SilN", "TheG", "UweP", "WerT",
"AlfF", "AndH", "AntH", "BerK", "BerR", "ChrG",
"FraR", "FreZ", "GerB", "GerB", "GerT", "GueS",
"GueV", "HanN", "HeiW", "JakW", "KlaS", "ManS",
"MarH", "PetS", "SilN", "TheG", "UweP", "WerT",
"AlfF", "AndH", "BerK", "BerR", "ChrG", "FraR",
"FreZ", "GerB", "GerB", "GerT", "GueV", "HanN",
"HeiW", "JakW", "KlaS", "ManS", "MarH", "PetS",
"PetW", "SilN", "SveR", "UweP", "WerT", "AlfF",
"AndH", "AntH", "BerK", "BerR", "ChrG", "FraR",
"FreZ", "GerB", "GerB", "GerT", "GueS", "GueV",
"HanN", "HeiW", "JakW", "KlaS", "ManS", "MarH",
"MicH", "PetS", "SilN", "SveR", "UweP", "WerT"),
prop_yes = c(0, 0.2, 0.6, 0.1, 0, 0, 0.1, 0, 0.3, 0,
0, 0, 0, 0.1, 0.8, 0.1, 0.1, 0, 0, 0, 0, 0, 0, 0, 0.3, 0,
0.3, 0.3, 0.5, 0.4, 0.778, 0, 0.05, 0.5, 0, 0, 0.4, 0.2, 0,
0, 0, 0.2, 0.2, 0.3, 0.2, 0.6, 0.2, 0.2, 0.1, 0.1, 0.1, 0,
0.1, 0.3, 0.4, 0.1, 0.111, 0, 0.2, 0.1, 0.2, 0.8, 0, 0.111,
0, 0.1, 0, 0.2, 0.3, 0.1, 0.4, 0.333, 0.2, 0.1, 0.2, 0, 0.2,
0.182, 0, 0.1, 0.364, 0.1, 0.3, 0.375, 0, 0, 0, 0.2, 0, 0.1,
0, 0, 0, 0, 0.1, 0.1, 0, 0.3, 0, 0, 0.3, 0, 0.333, 0, 0,
0.667, 0.2, 0.571, 0.2, 0, 0.2, 0.6, 0.2, 0, 0, 0, 0, 0, 0.2,
0, 0, 0, 0, 0.2, 0.3, 0, 0.7, 0.3, 0, 0.2, 0.75, 0.2, 0.1,
0.1, 0.4, 0.1, 0.4, 0.3, 0.222, 0.2, 0.1, 0.1, 0.5, 0.2, 0.6,
0, 0, 0.1, 0.167, 0.333, 0, 0.222, 0.4, 0.5, 0, 0.3, 0.1, 0),
size = c(10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L,
19L, 10L, 6L, 10L, 10L, 10L, 10L, 10L, 10L, 10L,
10L, 9L, 10L, 6L, 10L, 10L, 10L, 10L, 10L, 10L, 9L,
10L, 20L, 10L, 8L, 10L, 10L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 10L, 10L, 9L, 10L, 10L, 10L, 10L,
10L, 10L, 9L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 9L,
10L, 10L, 10L, 10L, 10L, 11L, 10L, 10L, 11L, 10L,
10L, 8L, 7L, 10L, 10L, 10L, 9L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 10L, 10L, 9L, 10L, 10L, 4L, 6L, 10L,
9L, 9L, 10L, 7L, 10L, 10L, 10L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 10L, 10L, 8L, 10L, 10L, 10L, 10L,
10L, 10L, 10L, 9L, 10L, 10L, 10L, 10L, 5L, 10L, 10L,
10L, 10L, 12L, 12L, 10L, 9L, 10L, 10L, 9L, 10L, 10L,
10L), tmean_winter = c(-3.83, -4.31, -3.97, -5.21,
-4.6, -4.09, -4.05, -4.09, -4.85, -4.48, -4.77,
-6.66, -4.16, -4.68, -4.48, -5.07, -3.83, -4.28,
-4.79, -4.83, -4.09, -4.43, 2.36, 1.47, 2.13, 1.09,
1.93, 2.26, 2.28, 1.98, 1.66, 1.3, 1.69, -1.01,
2.22, 1.89, 2, 1.23, 2.33, 2.1, 1.68, 1.66, 1.95,
1.38, 1.61, 0.86, 1.82, 0.48, 1.45, 1.74, 1.5, 1.78,
1.14, 0.65, 1.17, -1.59, 1.69, 1.55, 1.44, 0.65,
1.59, 1.16, 1.14, 1.23, 0.81, -1.53, -2.61, -1.52,
-2.7, -1.77, -1.54, -1.68, -1.32, -2.16, -2.82,
-1.95, -4.56, -1.57, -1.77, -1.76, -2.55, -1.51,
-1.98, -2.05, -1.97, -2.62, -4.48, -5.25, -4.04,
-4.92, -4.59, -5.34, -5.09, -4.12, -4.36, -5.23,
-4.94, -4.7, -5.28, -4.55, -7.07, -4.18, -5.17,
-4.56, -4.56, -4.74, -4.58, -4.62, -5.08, -5.25,
-1.87, -2.67, -2.84, -2.47, -3.11, -2.3, -2.01,
-2.05, -2.96, -2.57, -2.75, -2.54, -4.18, -2.07,
-3.04, -1.81, -2.39, -2.24, -2.75, -2.75, -2.79,
-2.44, -2.85, -0.35, -1.47, -0.43, -1.02, -0.76,
-1.23, -1.57, -0.48, -0.65, -1.18, -0.58, -0.92,
-1.58, -1.07, -4.05, -0.52, -2.1, -0.36, -0.75,
-1.04, -0.67, -1.05, -1.81, -0.61, -1.64)), .Names =
c("name", "prop_yes", "size", "tmean_winter"),
row.names = c(NA, -158L), class = "data.frame")
glmer(prop_yes~tmean_winter+(1|name), family='binomial',
data=cv)
glmer(prop_yes~tmean_winter+(1|name), family='binomial',
data=cv, weights=size)
