r - 数据框上的 icc，每个评分者都有行

Question

首先让我说我是 R 的新手，并试图弄清楚如何在我的特定数据集上运行 icc，这可能与正常情况下有点不同。

数据集如下所示

+------------+------------------+--------------+--------------+--------------+
|    date    | measurement_type | measurement1 | measurement2 | measurement3 |
+------------+------------------+--------------+--------------+--------------+
| 25-04-2020 |                1 |         15.5 |         34.3 |         43.2 |
| 25-04-2020 |                2 |         21.2 |         12.3 |          2.2 |
| 25-04-2020 |                3 |         16.2 |          9.6 |         43.3 |
| 25-04-2020 |                4 |           27 |            1 |            6 |
+------------+------------------+--------------+--------------+--------------+

现在我想对所有这些行进行 icc，因为每一行代表不同的评分者。它应该保留 date 和 measurement_type 列。

有人能给我指出正确的方向吗，我完全不知道该怎么做。

-------- 编辑 --------
我导出了将带有一些测试数据的实际数据集。哪个可用here

这里有两张重要的纸是第一张和第三张。第一个包含研究的所有参与者，第三个包含每个参与者的所有 4 份不同报告。到目前为止，我所拥有的代码只是将每个报告与正确的参与者联系起来；

library("XLConnect")
library("sqldf")
library("irr")
library("dplyr")
library("tidyr")

# Load in Workbook
wb = loadWorkbook("Measuring.xlsx")
# Load in Worksheet
# Sheet 1 = Study Results
# Sheet 3 = Meetpunten
records = readWorksheet(wb, sheet=1)
reports = readWorksheet(wb, sheet=3)

for (record in 1:nrow(records)) {
  recordId = records[record, 'Record.Id']
  participantReports = sqldf(sprintf("select * from reports where `Record.Id` = '%s'", recordId))

  baselineReport = sqldf("select * from participantReports where measurement_type = '1'")
  drinkReport = sqldf("select * from participantReports where measurement_type = '2'")
  regularReport = sqldf("select * from participantReports where measurement_type = '3'")
  exerciseReport = sqldf("select * from participantReports where measurement_type = '4'")
}

Answer 1

由于在您的数据中，每一行代表不同的评分者，但 irr 包中的 icc 函数需要评分者为列，您可以忽略表的前两列，转置它并运行 icc。

所以，假设这个表:

+------------+------------------+--------------+--------------+--------------+
|    date    | measurement_type | measurement1 | measurement2 | measurement3 |
+------------+------------------+--------------+--------------+--------------+
| 25-04-2020 |                1 |         15.5 |         34.3 |         43.2 |
| 25-04-2020 |                2 |         21.2 |         12.3 |          2.2 |
| 25-04-2020 |                3 |         16.2 |          9.6 |         43.3 |
| 25-04-2020 |                4 |           27 |            1 |            6 |
+------------+------------------+--------------+--------------+--------------+

存储在一个名为data的变量中，我会这样做:

data2 = data.matrix(data[,-c(1,2)]) # generates the dataset without the first two columns

data2是这张表:

+--------------+--------------+--------------+
| measurement1 | measurement2 | measurement3 |
+--------------+--------------+--------------+
|         15.5 |         34.3 |         43.2 |
|         21.2 |         12.3 |          2.2 |
|         16.2 |          9.6 |         43.3 |
|           27 |            1 |            6 |
+--------------+--------------+--------------+

然后:

data2 = t(data2) # transpose data2 so as to have raters in the columns and their ratings in each line
icc(data2) # here i'm not bothering with the parameters, but you should explore the appropriate icc parameters for your needs.

应该生成正确的运行。