I have long-term survey data for trees infected with a virus. The data was measured in 2000 and again in 2010. I'd like to compare the number of uninfected vs. infected trees in 2000 and 2010. However, the total number of trees at each timepoint is different. It seems I need to use McNemar's test for paired proportions, but the total number of my "paired" data for each year is not the same, because there were trees measured in 2010 that did not exist in 2000.

Here are the totals for 2000 and 2010:

Total # trees in 2000    Total # trees in 2010
        3640                     3148

Here are the #s for infection:

# Infected 2000    # Infected 2010
      1095               1331

Subtracted from total gives uninfected numbers:

# Uninfected 2000    #Uninfected 2010
     2545                 1817

I could not figure how to do McNemar's test when the "paired" data doesn't actually pair up due to the different total # of trees in each year. The null hypothesis for this comparison would be that there is no significant difference between the two years. Any advice would be greatly appreciated!


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.