Monday, October 6, 2008

Does ethanol decrease mileage?

I've got a good amount of data now regarding how much ethanol relates to mileage. I've tracked ethanol content, average mileage, and average speed for almost every tank I've driven.

I threw this data into Excel and ran a multiple regression. I wanted to see how these two factors (average speed and tank ethanol content) influence mileage.

Now, I'm a little weak in statistics, but my reading is that these two variables can explain about 60% of the variation in mileage. Average speed is a very strong factor in controling mileage. Ethanol content is not. In fact, we can pretty much throw ethanol content out, it is not a significant factor (95% confidence). But it is borderline at about 6%.

Interesting stuff. If I had to guess, if I could collect data on, say, average temperature at start, to capture the decrease in mileage due to cold starting, or average temp while driving, to capture when I use the AC, I'd be able to explain a lot more of the mileage.

No comments: