thanks jan. Your answer helps me a lot. An additional question. I can put that data into a matrix. I try to simulate data as people responding to a psychological test.
Hi. Alexander. It is for personality test or attitudes. I've seen some studies where the researcher changes the number of response categories. For example,
Lozano et al (2008) leave it rather unclear what they actually did. They claim they used normally distributed data. However, questionnaire data creates discrete (not continuous) responses. I think you have two options:
You could either use the link to stackoverflow I gave you. There you'll find code that uses the binomial distribution. The binomial distribution could be considered the discrete sister of the normal distribution (i.e. because if you increase the number of trials it converges to a normal distribution).
Another possibility would be to use "rnorm(n, mean, sd)" to create the vectors and then reduce the numbers into bins. However, there is the question on how wide the bins are supposed to be. You could probably use the hist function to do the binning.