I have a repeated measures data set where each variable is measured at several time points.That is say I have 20 subjects in my sample. Each subject is measured for three time points. (There are some missing data for some time points). I want to analyze how data is spread for a particular variable.
If there is no repeated measures I know I can use a method like histogram. But when there is repeated measures can I still use a histogram?
If I am plotting a variable for a particular subject should I use the mean value of the three time points for that subject or can I think of as the three time points as individual values for 3 subjects and draw a histogram as I normally would?
$\begingroup$
$\endgroup$
3
-
$\begingroup$ What is the motive of the analysis? Just to show how people move along? Or for detecting cases that are out of trend? Or as to gather some insight for your model's residual? etc. $\endgroup$– Penguin_KnightCommented Jun 10, 2015 at 12:05
-
$\begingroup$ @Penguin_Knight I want to see if the variable has a pattern/trend? $\endgroup$– sam_roxCommented Jun 10, 2015 at 12:13
-
1$\begingroup$ Plot three, one for each time point, and make the range of y and x axes the same across the graphs. Then by paneling them horizontally you can observe frequency change, by paneling them vertically you can observe if the whole sample's central tendency has moved and has the dispersion changed in each time point. This should give you a good start. $\endgroup$– Penguin_KnightCommented Jun 10, 2015 at 12:19
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
If I interpret your question correctly, you want to ignore the dependence between individual subjects, and look at the distributions on each time point. With only 20 points it would be best to display the data as side-by-side dotplots, rather than histograms. For really large data sets you would use side-by-side boxplots. If you really insist, you can stack the histograms for the three time points. All of these are easy to do with ggplot2 in R.
answered Jul 15, 2015 at 12:06