Is error rate data normally distributed?

Hi All,
I have performed an experiment where I collect the number of 'typing errors' that has occurred. For instance, a user entered 51 different numbers out of which he/she mis-typed 8 numbers. Thus error rate will be (8/51)*100 = 15.68%. Now, I have collected 14 participants' error rate data. All I need to know is that if the error rate data is normally distributed or not?

I performed the 'Shapiro-Wilk' test of normality which resulted in p = .69

With this information can I conclude that the data is normally distributed? From a theoretical standpoint, is the 'error-rate' normally distributed?

Please share your views.
I'm so not a stats-head, but I would suggest you plot the data. Plot the observed against the expected to see the shape of the residuals (QQ plot). Or maybe plot the residuals against the expected values and look at the shape (I think this is a NPP or PP plot?) I don't know if Shapiro-Wilk tests normality of the population?? (But it well may - as I said I'm not a pro....)