I have a dataframe with one factor column with two levels, and many numeric columns. I want to split the dataframe by the factor column and do t-test on the colunm pairs.
Using the example dataset Puromycin I want the result to look something like this:
Variable Treated Untreated p-value Test-statistic CI of difference****
Conc 0.3450 0.2763 XXX T XX - XX
Rate 141.58 110.7272 xxx T XX - XX
I think I am looking for a solution using PLYR that can an output the above results in a nice dataframe.
(The Puromycin only contains two numeric variables, but the solution I am looking for would work on a dataframe with many numeric variables)
UPDATE - I will try to clarify what i mean.
I would like to go from data that look like this:
Grouping variable var1 var2 var3 var4 var5
1 3 5 7 3 7
1 3 7 5 9 6
1 5 2 6 7 6
1 9 5 7 0 8
1 2 4 5 7 8
1 2 3 1 6 4
2 4 2 7 6 5
2 0 8 3 7 5
2 1 2 3 5 9
2 1 5 3 8 0
2 2 6 9 0 7
2 3 6 7 8 8
2 10 6 3 8 0
To a result dataframe that look like this:
"Mean in group 1" "Mean in group 2" "P-value of difference" "N"
var1 ## ## ## ##
var2 ## ## ## ##
var3 ## ## ## ##
var4 ## ## ## ##
var5 ## ## ## ##
Maybe it is something with mapply I am looking for because i want to split up my dataframe into dataframe1 and dataframe2 by a two-level factor, and apply a function( t-test) to the first parts of dataframe1 and dataframe2, and then a t-test on the second parts of dataframe1 and dataframe2, and then a t-test to the third parts of dataframe1 and dataframe2, and so on on all the column-pairs generated by the split by factor.
Maybe this produces the result you are looking for:
df <- read.table(text="Group var1 var2 var3 var4 var5
1 3 5 7 3 7
1 3 7 5 9 6
1 5 2 6 7 6
1 9 5 7 0 8
1 2 4 5 7 8
1 2 3 1 6 4
2 4 2 7 6 5
2 0 8 3 7 5
2 1 2 3 5 9
2 1 5 3 8 0
2 2 6 9 0 7
2 3 6 7 8 8
2 10 6 3 8 0", header = TRUE)
t(sapply(df[-1], function(x)
unlist(t.test(x~df$Group)[c("estimate","p.value","statistic","conf.int")])))
The result:
estimate.mean in group 1 estimate.mean in group 2 p.value statistic.t conf.int1 conf.int2
var1 4.000000 3.000000 0.5635410 0.5955919 -2.696975 4.696975
var2 4.333333 5.000000 0.5592911 -0.6022411 -3.104788 1.771454
var3 5.166667 5.000000 0.9028444 0.1249164 -2.770103 3.103436
var4 5.333333 6.000000 0.7067827 -0.3869530 -4.497927 3.164593
var5 6.500000 4.857143 0.3053172 1.0925986 -1.803808 5.089522
Maybe you can find this useful
res <- sapply(split(Puromycin[,-3], Puromycin$state), t.test)[c(1:3,5),]
conf.level <- sapply(sapply(split(Puromycin[,-3], Puromycin$state), t.test)[4, ], '[', 1:2)
res <- rbind(res, conf.level.lower=conf.level[1,], conf.level.upper=conf.level[2,])
res
treated untreated
statistic 4.297025 4.206221
parameter 23 21
p.value 0.00026856 0.0003968191
estimate 70.96417 55.50182
conf.level.lower 36.80086 28.06095
conf.level.upper 105.1275 82.94268
You can also use a custom made package matrixTests for this. Example using the data.frame prepared by @Sven below:
df <- read.table(text="Group var1 var2 var3 var4 var5
1 3 5 7 3 7
1 3 7 5 9 6
1 5 2 6 7 6
1 9 5 7 0 8
1 2 4 5 7 8
1 2 3 1 6 4
2 4 2 7 6 5
2 0 8 3 7 5
2 1 2 3 5 9
2 1 5 3 8 0
2 2 6 9 0 7
2 3 6 7 8 8
2 10 6 3 8 0", header = TRUE)
library(matrixTests)
col_t_welch(df[df$Group==1,-1], df[df$Group==2,-1])
obs.x obs.y obs.tot mean.x mean.y mean.diff var.x var.y stderr df statistic pvalue conf.low conf.high alternative mean.null conf.level
var1 6 7 13 4.000000 3.000000 1.0000000 7.200000 11.333333 1.679002 10.963146 0.5955919 0.5635410 -2.696975 4.696975 two.sided 0 0.95
var2 6 7 13 4.333333 5.000000 -0.6666667 3.066667 5.000000 1.106976 10.938135 -0.6022411 0.5592911 -3.104788 1.771454 two.sided 0 0.95
var3 6 7 13 5.166667 5.000000 0.1666667 4.966667 6.666667 1.334226 10.995151 0.1249164 0.9028444 -2.770103 3.103436 two.sided 0 0.95
var4 6 7 13 5.333333 6.000000 -0.6666667 10.666667 8.333333 1.722862 10.146824 -0.3869530 0.7067827 -4.497927 3.164593 two.sided 0 0.95
var5 6 7 13 6.500000 4.857143 1.6428571 2.300000 13.142857 1.503624 8.285649 1.0925986 0.3053172 -1.803808 5.089522 two.sided 0 0.95
来源：https://stackoverflow.com/questions/13790611/apply-t-test-on-many-columns-in-a-dataframe-split-by-factor