We've all seen it, and frankly the first 3 cars make pretty good sense -except I have no idea how the Mustang didn't end up on that list- but we all wondered how the Mercury Topaz or Suzuki Reno even made it, having been discontinued for years. Insurance.com claimed to have used a sample size over over 500,000 cars.

The real kicker comes when you also factor in the number of different cars on the road today, so for the purpose of science *let's say there's 500 different models.*

Okay so with 500 different cars, that leaves us with a *sample size of 1000 cars per model.* But this raises a few questions:

- How are they coming about the sample size?
- Are there even 1000 of those cars on the road today?
- Are they using equal random sample sizes?

So, let's say they do use equal sample sizes; does that really make sense? A Ford F150, for example, has millions of units sold each year, where a WRX is a fraction of that. This makes comparing the two vehicles difficult without selecting a fair value for both. This value can be subjective though, as insurance.com might have not gone about picking the vehicles they were using in a fair manner.

Now, let's say that they didn't use equal sample sizes, but instead focused more on the percentage of each car. This makes an even bigger mess of everything. *What if there are 300,000 Cadillac Escalades used in this experiment?* That would make for a smaller sample size of each other vehicle. Let's say for the Mercury Topaz *they used a sample size of 37 vehicles. *So in order to reach their ticketed percentage of 28.8%, that would require 10.656 vehicles, *or for the sake of this argument we'll say 11. *

Now I don't know about everyone else, but 11 drivers out of 37 vehicles seems pretty average if you ask me. And by just a few more drivers getting tickets the percentage could easily be skewed.

I'm sure I'll have some opponents on my position, but I refuse to blindly follow these numbers they gave without actually knowing how the "experiment" was conducted.

**Photo via hooniverse.com**