Functions with R The first step is to invoke the boot library >library(boot) The boot function need the following arguments: boot(data,statistic,R,arguments used by statistic) Data is the data you want to apply your statistic to. Statistic is a function that computes a specific TS, it must return a single value R is the number of bootstrap replication If the function computing the TS needs additional arguments, they can be provided as the last arguments

Creating the appropriate function to pass to the statistic argument The function boot uses the function provided in the statistic argument to sample “something” with replacement. In this course, we will only see the case when the “thing” sampled is subject. When your data are a single vector (see below), you sample elements boot.mean <- function(data,d) {data2<-data[d] bootmean<-mean(data2,na.rm=T) return(bootmean) }

Create a statistics function for SD

Applying boot on the function myData<-c(1,2,3,4,5,6,7,8,9,10) Res.boot<-boot(myData,boot.mean,R=200) Try this with the data on the first call on Day 2

Obtain the CI Boot.ci(Res.boot) Create a histogram of the TS in each bootstrap sample. The initial statistic will appear as a red line (col=“red”), with a width 3 times larger a normal line (lwd=3) >hist(Res.boot$t) Use print.default on the Res.boot to find the TS value in the initial sample, then: >abline(value of the initial TS,col=“red”,lwd=3)

Exercise: add vertical blue lines for the bca confidence interval

Obtaining a CI around a SD difference: interindividual variability in mood on Tuesday, call 1 vs call 6 Modify your function so that it indicates what to sample (hint: the rows since they represent your subjects) Then compute the SD across subjects for call 1 Tuesday and for call 6 Tuesday. Subtract one SD from the other. This gives you your TS that your new function should return. Apply boot

Exercise: draw a similar graph of the difference in intraindividual variability between Tuesday and Sunday Step 1: create the function that returns the TS Step 2: bootstrap the appropriate data Step 3: use boot.ci to obtain the CI Step 4: draw the graph