Keep detailed records of weather conditions along with ET slip data, and and evaluate how the combo behaves. I happen to use a weather station that spits out barometer, temperature, humidity, etc., and it calculates density altitude from these parameters. Then compare runs and see how they vary with density altitude (or whatever specific measurements you can make), taking into account any differences in 60', traction, wind, sign of the zodiac, etc. in these comparisons.
I think my data come out to about 0.05-0.07 sec per 1000' d.a. for a 9-sec car... been broken all year so it's not fresh in my mind. I would imagine proportionally (?) lesser and greater effects on ET would be observed with faster and slower cars, respectively.
Art