Lou,
The point I wanted to make is that how you treat the data is mostly determined by what you want to accomplish. If the mean value of the ratings is used, that makes sense under the assumption that there exists one true value for the course and the reason for many ratings is so that rater characteristics or errors will "average out". Eliminating disparate values can avoid a mean that is incorrect under that sort of assumption.
But if you were to assume that a given course will appeal differently to various types of individuals, the ratings that would be "outliers" under the first set of assumptions take on a particular benefit specifically because they are very different. Under this type of assumption you would consider a large number of ratings falling all around the same value to be redundant (or maybe you'd consider that clustering a separate parameter that's informative about the distribution of our posited "types" or species of individual preferences). This type of analysis most certainly does not eliminate disparate values.
From what little I know about the magazine panels, it seems clear that the former set of assumptions are in play and not the latter. Which is certainly fine by me. I just always take care to point out the Cheshire Cat principle that ones proper direction "...depends a good deal on where you want to get to...", not to score any points in favor of one destination or another. Or put less generously, I'm being pedantic.
I should know better than to engage a professional on a subject that I am, at best, a mildly interested layman (I'm sorry to confess that I spent much of my time in quant and stats imagining instead that I was in the superior learning environment of a Doak 5 course a few miles away).
Pedantic? Nah. Making a distinction without a difference?
Of course, you are 100% right about polls, data, and manipulation of information to achieve the desired results. We are frequently provided links on this site as wise revelation of gospel supporting our most cherished beliefs. We could engage in endless duels of expert analysis citing fairly common, mechanically-correct statistical methodologies without reaching agreement. Among the biggest fallacies in research, particularly outside the physical sciences, is the underlying caveat- ceteris paribus.
The thesis of this thread is rather mundane, surely falling under central tendency. I was unaware that "Golf" attempted to come up with a composite based on disparate preferences or criteria. Does that mean that it provides no guidance to its raters and somehow aggregates the data? Seems like this would take some serious manipulation of the information. How many outliers would it take for any semblance of statistical significance? What type of subjective analyis of the outliers (and what definition of what is an outlier) would be necessary to draw useful inferences?
Ratings are not rocket science. I don't have the statistical horsepower to "prove" that offseting errors and variances are sufficient to make them useful, but for my purposes, that explanation is (sufficient). Is Cypress Point or Pine Valley the #1 course in the U.S.? Who knows? I like to think that they are in the relevant range of, say, 1-10, and that's close enough for me.
JK,
Just because I recognize that tee times, like electricity, if unused are gone and produce no revenues, it doesn't mean that I relish the reality. Actually, beyond the agronomical benefits of light traffic, perhaps it also adds considerable intrinsic value to the members in terms of convenience and exclusivity. My only observation is that clubs which choose to host raters do so voluntarily with little negative effect on the bottom line- or perhaps you know some members as well who dropped out of their clubs because of too many comp rounds.
Many people drop memberships for various reasons- the economy, family responsibilities, declining skills, changes at their club, etc. I suspect that the number of raters who have dropped out from private clubs in favor of free golf is as infinitisimally small as the percentage of my tax dollars ending up in your pocket for working in Michigan City. Of course, I could be wrong.
BTW, your disgust is as genuine as the Washington elites running against Washington. Precious!