My Land Rover LR3 HSE has a fairly extensive menu of data on display. One of these displays is "Average Speed." Like all the numbers on the display, it resets when the mileage on the trip odometer (actually one of the two trip odometers) is reset to zero. I do this at each fill up, so the average speed on the indicator shows the average for the current tank full.
The average speed should reflect, among other things, the amount of time I spend on the highway at 55 m.p.h. versus the time I spend on streets and in traffic jams. It crossed my mind eight fill ups ago to add the average speed data for the tank full to the myriad of other data I collect when I fill up. Since highway mileage should be higher, it's reasonable to expect that higher average speeds for a tank should correlate with higher miles per gallon for that tank.
To check this theory, I've plotted m.p.g. on the vertical axis versus average speed on the horizontal axis. As expected, higher speeds are accompanied by higher mileage numbers. The linear trendline, calculated by Excel, has a slope of about 0.32, meaning that each mile per hour increase in average speed over a tank full yields an increase in 0.32 m.p.g. for that tank full. The coefficient of determination ("R squared"), however, is low at 0.48. Thus, while there is a positive correlation between average speed and gas mileage, average speed is a weak predictor of gas mileage. More data will enable a deeper analysis.
For those who are curious about what the actual numbers are, the lowest average speed has been 31.7 m.p.h. and the highest has been 39.1. The latter number was for a tank full the bulk of which was expended on the interstate from Las Vegas to Los Angeles. That tank full produced a fuel economy of 23.41 m.p.g. The lowest average speed produced a fuel economy of 19.82 m.p.g.