The first collumn is what you get when you multiply engine rpm in thousands with the numbber of teeth on the primary ,i.e 13*9.37or 121.8 (rounded off to 122),this number is proprtional to the road speed.Now if you repeat this for the 14t test run,you get a different # also proportional to speed in the exactly same manner since nothing else has changed.If you take the ratio of these numbers that is for the 14t and 13t run, you get 1.058 this ought to be the same as the measured speed ratios 36.5/34.8 or (1.049).
In the 16t row I did the same thing and compared the 16t run with the 14 t run,there the ratios ,1.14 and 1.022 don't agree at all.I was trying to check on the integrity of your data.I trust your speed data (time measurement) more than your rpm readings ,which could well be on the flaky side.Is it clearer now?