我有一个简单的数据集，其中包含两次治疗（trt），3次重复以及在不同日期重复进行的测量。响应变量是yield。我们正在寻找治疗上的差异。 我知道我的重复测量嵌套在rep中，像aov（yield〜trt）这样的简单方差分析不合适，因为它将每个重复测量视为独立的rep。 我认为我必须指定一个错误字词，并将日期嵌套在rep中，但是我对语法感到困惑。
这是我的代码
summary(aov(data=yi,yield~trt + Error(rep/date))) #this means date nested in rep
谢谢您的帮助！
这是我的数据集
structure(list(date = structure(c(17676, 17680, 17683, 17687,
17690, 17695, 17698, 17702, 17705, 17709, 17712, 17716, 17719,
17723, 17726, 17730, 17733, 17737, 17740, 17744, 17747, 17751,
17754, 17759, 17761, 17765, 17768, 17772, 17775, 17779, 17782,
17676, 17680, 17683, 17687, 17690, 17695, 17698, 17702, 17705,
17709, 17712, 17716, 17719, 17723, 17726, 17730, 17733, 17737,
17740, 17744, 17747, 17751, 17754, 17759, 17761, 17765, 17768,
17772, 17775, 17779, 17782, 17676, 17680, 17683, 17687, 17690,
17695, 17698, 17702, 17705, 17709, 17712, 17716, 17719, 17723,
17726, 17730, 17733, 17737, 17740, 17744, 17747, 17751, 17754,
17759, 17761, 17765, 17768, 17772, 17775, 17779, 17782, 17676,
17680, 17683, 17687, 17690, 17695, 17698, 17702, 17705, 17709,
17712, 17716, 17719, 17723, 17726, 17730, 17733, 17737, 17740,
17744, 17747, 17751, 17754, 17759, 17761, 17765, 17768, 17772,
17775, 17779, 17782, 17676, 17680, 17683, 17687, 17690, 17695,
17698, 17702, 17705, 17709, 17712, 17716, 17719, 17723, 17726,
17730, 17733, 17737, 17740, 17744, 17747, 17751, 17754, 17759,
17761, 17765, 17768, 17772, 17775, 17779, 17782, 17676, 17680,
17683, 17687, 17690, 17695, 17698, 17702, 17705, 17709, 17712,
17716, 17719, 17723, 17726, 17730, 17733, 17737, 17740, 17744,
17747, 17751, 17754, 17759, 17761, 17765, 17768, 17772, 17775,
17779, 17782), class = "Date"), yield = c(990, 1560, 1520, 1845,
1820, 2260, 1175, 1630, 1305, 2065, 1060, 2060, 1565, 1755, 1380,
1875, 1590, 1640, 1185, 1585, 830, 2055, 1285, 2555, 1230, 2095,
1565, 1935, 1235, 1510, 1570, 660, 1010, 720, 1370, 1305, 1670,
1120, 1780, 1210, 1675, 1600, 1490, 1010, 1060, 985, 1075, 915,
1640, 895, 1535, 1330, 1990, 980, 2270, 1135, 1525, 1180, 1020,
1010, 1050, 710, 975, 1290, 1115, 785, 1625, 1705, 1235, 1700,
1430, 1175, 1115, 1275, 1035, 1085, 1185, 1475, 870, 1580, 845,
1560, 635, 2315, 800, 2365, 945, 1515, 1435, 1855, 1355, 1585,
1315, 745, 1125, 655, 1085, 1430, 1980, 985, 1675, 1090, 1510,
1085, 1730, 1065, 1460, 795, 1345, 1165, 1145, 655, 1310, 645,
1780, 945, 2375, 950, 1805, 1760, 1510, 1590, 1515, 1030, 645,
1681, 750, 985, 1065, 1350, 875, 1520, 1045, 1800, 1165, 2070,
1410, 1435, 1490, 1705, 1230, 1865, 1015, 1740, 745, 1275, 1180,
2030, 755, 1540, 1410, 1340, 1155, 1495, 1160, 1335, 1030, 690,
1330, 1005, 1240, 1190, 1625, 965, 1565, 1195, 1305, 1180, 1255,
1090, 1675, 1250, 1615, 595, 1395, 695, 1600, 1165, 2215, 675,
1300, 1190, 1035, 985, 985, 935), trt = c("117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"117 lb/ac Preplant", "117 lb/ac Preplant", "117 lb/ac Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant", "No Preplant", "No Preplant", "No Preplant", "No Preplant",
"No Preplant"), rep = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
3, 3, 3, 3, 3)), row.names = c(NA, -186L), class = "data.frame")
答案 0 :(得分：1)
如果我正确理解了您的问题，则您对yield ~ trt感兴趣，但是数据不是独立的，因为每个rep都有多个测量值。
据检查数据可知，唯一发生的嵌套是trt中rep的嵌套。如果您按照以下方式建模，则会自动考虑：
require(lme4)
LMM <- lmer(yield ~ trt + (1 | rep), data = yi)
summary(LMM)
# or:
LMM <- aov(data = yi, yield ~ trt + Error(rep))
summary(LMM)
如果您担心收益取决于日期，则可以将其建模为固定效应，因为(1 | rep)已经考虑了重复措施。此外，考虑一下date产生随机影响的含义：如果您要重复此实验，是否会随机选择新日期？这些日期是否来自某个较大的日期分布，从而导致收益率呈正态分布偏移？ date *似乎有季节性的影响，您可以将其建模为例如像this。
*尝试：plot(yield ~ date, yi, pch = paste(rep), col = factor(trt))