I'm trying to complete a PCA on the datasets that I have in R. Each participant in the RData file was asked to fill out a survey four times a day for a couple of months. However, the survey issued in the morning contained three extra questions (e.g. number of hours slept) so the remaining three survey results for that day would be missing three scores. I've thought of two possible options but I'm not really sure which would the best route to take to preparing the datasets prior to PCA.

Option A: Removing the three variables all together

Option B: Fill in the missing values with the recorded values from the morning survey. If a survey wasn't completed in the morning (i.e. there isn't a score for the morning survey), then the mean is used.

At the moment, these are the only options I can think of and I'm not sure if either of them are suitable. Please let me know your thoughts about this. Thank you in advance. Also, this is my first time making a post so I hope I provided enough information.

  • $\begingroup$ Motivated by a similar problem, I asked a closely related question at stats.stackexchange.com/questions/1781. Although it has not been satisfactorily answered, several people have provided interesting suggestions and some of those might be relevant to your case. I have experimented with multiple imputation and made some progress with that. $\endgroup$ – whuber Jan 2 at 14:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.