Calculating P-value comparing two percentages of accuracy?

[phys-lexic]

I have been out of unversity statistics and biometry for a few years do date. I am working in a clical research facility and we are comparing two methods of microbial identification and am now working on the results / charts / statistics. Forgive me for my ignorance, but I am forgetting which method would I use to calculate the p-value for comparing percentages; i.e. out of 100 samples method 1 was 90% accurate and method 2 was 60% accurate (random values). I was going through my old notes and tried to apply chi-squared, etc. It seems I didn't retain much.

Thanks in advance.

 

[Stephen Tashi]

You haven't described a specific situation. For example, were both methods applied to each of 100 samples? Or was one method applied to 100 samples and the other method applied to 100 different samples?

Perhaps your old notes deal with "paired" and "un-paired" observations. Is the accuracy of an application of a method to a single sample a result like 0.93? For un-paired observations, look up the "t-test for equality of means".