学習に取り掛かりtidyr、問題を抱えていspread()ます。
偽の実験データセットは次のとおりです。
library(tidyr)
df <- structure(list(mood = c(0.855, -0.103, 0.421, -0.222, 0.772, -0.027, -1.088, 0.923, -1.516, -1.503, -0.358, -0.357, -0.344, 0.294, 0.348, -0.174, 0.872, -1.188, 0.842, -0.246, -0.758, 0.674, 0.045, 0.72, -1.253, 0.00599999999999995, -0.0749999999999999,1.623, -1.754, -0.44, -0.607, -0.083, -0.827, -0.337, -0.6, 0.429, -0.383, -1.755, 0.894, 0.146, -0.658, -0.409, -0.531, 1.388, -0.688, 0.521, -0.662, 0.852, -1.363, 0.18, -0.775, 0.393, -0.926, 0.809, -0.857, 0.889, 0.0969999999999999, -1.553, -0.21,1.769, -0.114, -0.203, 0.805, 0.186, 0.286, -0.076, 0.137, 1.208, 0.33, 0.34, 0.832, 0.815, -0.427, 0.444, -0.838, 1.45, 1.701, -2.265, 0.531, 0.808),
subj = structure(c(1L, 12L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L,11L, 13L, 1L, 12L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 13L, 1L, 12L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 13L, 1L, 12L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 2L, 3L, 4L, 5L,6L, 7L, 8L, 9L, 10L, 11L, 13L), .Label = c("s1", "s10", "s11", "s12", "s13", "s14", "s15", "s16", "s17", "s18", "s19", "s2", "s20", "s3", "s4", "s5", "s6", "s7", "s8", "s9"), class = "factor"),
depressed = structure(c(2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L,1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L), .Label = c("no", "yes"), class = "factor"),
activity = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,2L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("exercize", "relaxation"), class = "factor"),
drug = structure(c(1L, 1L,1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L,2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("placebo", "SSRI"), class = "factor")), .Names = c("mood", "subj", "depressed", "activity", "drug"), row.names = c(NA, -80L), class = "data.frame")
をのぞいてdfみると、activityとdrugは被験者内因子であるのに対し、depressedは被験者間因子であることがわかります。
head(arrange(df, subj))
mood subj depressed activity drug
1 0.855 s1 yes relaxation placebo
2 -0.758 s1 yes relaxation SSRI
3 -0.658 s1 yes exercize placebo
4 -0.114 s1 yes exercize SSRI
5 -1.503 s10 no relaxation placebo
6 -0.440 s10 no relaxation SSRI
df各主題moodが別々の列で表され、各行が と の組み合わせに対応するactivityワイド形式に変換したいと思いdrugます。残念ながら、各レベルの行のセットを作成しないようにする方法がわかりませんdepressed。これが私がこれまでに持っているものです：
df %>% spread(subj, mood) %>% `[`(1:5)
depressed activity drug s1 s10
1 no exercize placebo NA 0.18
2 no exercize SSRI NA 0.34
3 no relaxation placebo NA -1.50
4 no relaxation SSRI NA -0.44
5 yes exercize placebo -0.658 NA
6 yes exercize SSRI -0.114 NA
7 yes relaxation placebo 0.855 NA
8 yes relaxation SSRI -0.758 NA
depressed結果の出力が4行しかないように、拡張を避けたいと思います。