Are you a statistics whiz? What is the population for this? I mean, how many Pacificas were made? And can you tell me the statistical significance goal you are shooting for? What sample size represents a 90% significance for the population of Pacifica production, scientifically speaking of course? And what significance does your current sample size represent? And the margin of error for each of the above?
Throwing our words like "Sizes remain marginal" sounds very English-y/Layman to me. Throw some science at us, man!
The size of the population is not part of the formula when calculating confidence intervals or required sample sizes
What does matter: the variance.
Also, "90% significance for the population" is, well, gibberish. I think you mean 90% confidence intervals, which means that if the same survey were conducted ten times, then nine times out of those ten the result would fall within the specified range. There's more than one variable here. You need to specify both the desired confidence interval, and the maximum width of the interval. For example, you might want 90% confidence that the actual number is no more than 10% different than the reported result.
I used to include confidence intervals in the results, but hardly anyone knows enough statistics to understand what they mean. Many people suggested I drop that information because it was too confusing to the typical site visitor, so I did.
The intervals are very wide for sample sizes I call "marginal," meaning in the high teens or low 20s. I don't consider a result "official" unless the sample size is at least 25, because that's where the results start to stabilize. Those with smaller sample sizes are asterisked, are not visible to non-members, and should not be used as anything but rough indicators.
Where would I like to be? I'd like a minimum sample size of 100, but we're not there yet in many cases. The way to get there is to start where we are, and keep building. For one model we already have a sample size of 220, and a number of others are already over 100. The Pacifica just isn't one of them. Would I like it to be? You bet.
Here's what we do have so far:
In terms of successful repair trips per 100 cars per year:
2004: 105, about average, but close to "worse than average"
2005: 47, better than average
2007: 109, worse than average, largely thanks to the new transmission
Sample sizes were small, even very small, for all three years. So, as noted above, these are just "ballpark" results.
A big thanks to everyone who has been helping. We'll have further updates in May, August, and November. With more participants, we could provide more precise information and cover all model years.
We now have a separate results page for each model that includes the results for competitors. The page for the Chrysler Pacifica:
Chrysler Pacifica reliability comparisons
