Quote:
Originally Posted by Brian Carlton
While the demographics of CR subscribers are clearly different than the full population of owners, I don't see how the demographics differences between the two groups would affect the outcome.
The individuals merely report issues with the vehicle. Their results may vary as compared to the general population, but, the relative percentages between vehicles will probably be similar.
I value the statistics as provided by CR. What I question is their subjective evaluations of the vehicles. Here is where the bias of the driver will definitely come to bear. The driver should be escorted into the vehicle without being able to see it's moniker. All interior badging should be removed. Let the driver evaluate the vehicle without knowing, in advance, who the manufacturer is. Bet the results might be significantly different.
|
For stats to be valid, the sample drawn from a population is assumed to have the same characteristics of the population as a whole. It maybe that CR has the same distribution as the larger population. Or it maybe so wildly different in some key category that the sample is biased. Without having demographics to compare we don't know anything about the validity of the comparison.
Let's imagine an example. Assume that CR subscribers are more likely to be tree-hugging, compost-mixing, goat-raising, pot-smoking back-to-the-land unreformed hippies than is found in the general population.
Would their answers be the same as the larger population? I don't know.
Why don't I know? Because I have no demographics to compare.
Or there's a huge vat of millions and millions of marbles composed of thousands of different colors. Somebody hands me a jar of marbles and tells me, "These marbles came from the vat. The greatest number of marbles in the jar is yellow." I publish an article about marble colors and say, "Most of the marbles in the vat are yellow."