It seems your model is built to generalize to other participants. This is not possible when there is no information about the variability between participants (as you have data for only one participant!).

It seems your model is built to generalize to other participants. This is not possible when there is no information about the variability between participants (as you have data for only one participant!).

you might have to increase your sample size if possible. My experience with repeated measures is that is very dependant and sensitive to your n. Also, did you check for assumptions? 

Here's another way to think about why you are not getting a significance test in your output.  The error term for the F-test in a one-way repeated measures ANOVA is the Mean Square for the Treatment x Subjects interaction.  Like any other interaction, it measures how much the effect of one variable varies across levels of the other variable.  In this case, it asks how much the Treatment effect (i.e., the pattern of differences among the means for your 3 time points) varies by subject.  In order to answer that question, you need at least two subjects.  You have only one subject.  Therefore, it breaks down.  HTH.

I am looking to see whether changes are significant within one participant for example, if at time 1 the person scores 50 in a test, second time they score 55 and third they score 65, then how do I test whether these changes are statistically significant?

This is a case study so larger samples are not possible.

Thanks!

There are 2 options:

either you have an assumption about the course of the changes (linear, exponential, quadratic, etc), then this is a regression problem, or you don't have any idea about this course, then you will need information about the variance of the measurements at the individual time-points (that is, you will need several independent measurements at each time point for this patient).

Thank you, I have decided to use a method more suitable for small sample sizes.