Talk:Race, Evolution, and Behavior/2006-11-01 Archive

methodology of aggregation
is the "methodology of aggregation" simply meta-analysis or is it something different? anything published on that question? --Rikurzhen 04:58, 30 October 2006 (UTC)


 * FWIW, it seems like poorly controlled meta-analysis, with a bit of smoke and mirrors alluding to Fermi equations. Lieberman's paper talks extensively about Rushton's "principle of aggregation - (quoting Rushton) "Focusing on correlations between just two items or situations can lead to major errors of interpretation.  The more accurate assessment is to use a principle of aggregation and average across a number of measures ....this is because the randomness in any one measure...is averaged out over several measures, leaving a clearer view of underlying relationships".  The critical flaw here (as exposed by Lieberman), is that Fermi equations only work if you have a number of chained equations where the variation is random - taking what are obviously terribly biased measures, no matter how many of them you take, is NOT the same thing as a chain of equations, and do not lead to a "more accurate assessment". --JereKrischel 07:34, 30 October 2006 (UTC)

Yes, a meta analysis is exactly what Rushton does. One of the weaknesses of most conclusions in science (including Rushton's) is that they depend too much on meta analysis, which is simply an average of all studies. Becaus sometimes its much better to look at one single well controlled study than to average hundreds of poor ones as is usually done in academia. For example, imagine if during the next presidential election, academics wanted to predict who would win, but couldn't afford to do the kind of well controlled nationally representative statistically significant poll that Gallup does, and thus instead just decided to look at all the various internet polls, magazine polls, etc, and averaged them. If you averaged enough of these unscientific polls you would probably get the right answer (since sampling error in different polls would probably cancel out) but you could never state with any kind of statistical confidence how accurate your results are.

This is exactly what Rushton does, but he's not alone. Virtually all major conclusions in the social sciences are based on what the bulk of the studies show (each individual study flawed in some major way), and it's only later, when someone discovers a flaw that all the studies had in common, do conclusions get overturned. But this isn't because of incompentence on the part of scientists, it's because of a lack of funding. Can you imagine how much money Rushton would need to get a large, statistically representative sample of the ENTIRE negroid race (not just African Americans, but Nigerians, pygmies, watusi, bushmen), the ENTIRE caucasoid race (Europeans, Arabs, East Indians), the ENTIRE mongoloid race (Chinese, Japanese, Koreans and are Native Americans and South East Asians even mongoloid on the genetic level?), all the different countries he would need to survey, and than perform reliable measures of all 60 traits? So realistically, that's not how science works. Instead he'll take some condom size data from the world health organization, some sexual activity data from a shopping mall survey, some head size measurments from the U.S. army, and all he can do is hope that the various pieces of data triangulate towards truth. And if the pattern he claims to have discovered gets repeated and replicated enough times (especially by people who disagree with him because people who agree may not publish their studies if they don't confirm Rushton), by enough independent sources, using a variety of methods, he can claim to have proven it.

My biggest problem with Rushton is that the theory seems too conveniently perfect to be true. How could all 60 of these traits all perfectly followed the pattern he predicts? If he's honestly just aggregating all the studies you would expect at least some of them to go in the opposite direction just by chance error. Even if Rushton is 100% right, when you're dealing with 60 variables, a few of the aggregates from samples should go in the opposite direction just by sampling error and methodological error, especially considering the imperfections in the studies. Maybe there were r/K traits that didn't confirm his theory which he excluded from the study. In fact he is honest enough to mention some anomalies in his book, so we should add a section about excluded data as not mislead the reader into thinking all r/K variables perfectly follows Rushton's theory if you simply average enough studies. So while as far as I can tell, Rushton is taking an honest and objective average of all known studies to show how the races score on r/K variables, it's kind of arbitrary which variables he defines as r/K, and there could be other variables that he chooses not to define as r/K because whne he aggregated the studies, they didn't reinfoce his rankings (excluding those that don't fit his pattern) Minorcorrections 17:50, 30 October 2006 (UTC)


 * Interesting. Surely Rushton picked the phenotypes that would demonstrate his point and ignored those that don't. Insomuch as most of these phenotypes are probably reflections of just a few fundamental causes -- developmental timing and intelligence -- I'm not so surprised that he found so many examples. Has anyone published these critiques? --Rikurzhen 22:15, 30 October 2006 (UTC)


 * Yes, Lieberman speaks at length about it here. His critique regarding Rushton goes beyond simply picking and ignoring phenotypes - his completely unsupportable interpretations of data are strongly criticized. --JereKrischel 02:08, 31 October 2006 (UTC)
 * I think Lieberman's critque is a little over the top and a bit off topic when he talks about mixed race people (Rushton never claimed everyone was 100% pure) and his arguments about race science serving to justify the social order is a rip-off of Stephen Jay Gould. Lieberman is being very POV when he claims that Rushton's categories are artificial and that Rushton is wrong to average out the inconsistencies in the data.  What's wrong with trying to simplify complexity? Organizing and simplifying a complex world is the point of science.  Rushton does need a lot more controlled and modern studies to substantiate his theories, but I do feel the rough and ready aggregation of existing studies provides a good starting point for defending a hypothesis. Just my opinion of course. Minorcorrections 04:32, 31 October 2006 (UTC)
 * Well, I think one of the problems (IMHO), is that you can't point to anyone who is 100% pure anything - there is no standard (that I've heard of) to determine who is more "pure" than someone else. Without being able to identify who belongs in what group, or who is the prototype for a given group, you're not simplifying complexity, you're disregarding and/or arbitrarily using data. --JereKrischel 17:31, 31 October 2006 (UTC)
 * Well everything about his theory is rough. It's difficult to get exact brain size measures so most of the time he relies on head size. It's difficult to get DNA tests showing exactly what race you belong to so most of the time he relies on how people self-identify.  But if he's found all these differences using such crude methodology, his pattern should only become stronger as data gets more precise. Generally poor data creates a lot of random noise, which makes it HARDER to find patterns, not easier.  So the fact that Rushton found a pattern IN SPITE of all the imperfections of the data tells me there might be something to his theory. 134.117.83.67 19:25, 1 November 2006 (UTC)
 * Poor data creates noise which can be mistaken for a pattern, which is obviously what Rushton has done. A difficult question to answer regarding DNA is that we don't have any prototypical definitions of races - you cannot identify a standard by which you can claim someone 100% Mongoloid, and someone else 98% Mongoloid, even using DNA.  Without standard definitions, all you have is imagination. --JereKrischel 00:43, 2 November 2006 (UTC)
 * Poor data usually has a lot of random error and random by definition lacks a pattern. And Rushton does define his races.  He says a black is anyone most of whose "recent" ancestors were born in sub-Saharan Africa.  A pure negroid would be anyone with vrtually 100% of their ancestors born in sub-Saharan Africa. And similarly for caucasoids and mongoloids. Minorcorrections 06:43, 2 November 2006 (UTC)
 * Also, Lieberman implies that the data must be unreliable and biased if it used to show caucasoids were smartest and now in a new social climate shows mongoloids are. But if you read pg 87 of The Mismeasure of Man, Gould criticises Paul Broca for claiming that whites were smartest, even though his data at the time clearly showed that mongoloids had big brains.  Broca tried to claim that small brain size was relevant in evaluating black intelligence, but large brain size obviously meant nothing since mongoloids had such big brains to which Gould replied "An unbeatable argument. Deny it at one end where conclusions are uncongenial; affirm it by the same criterion at the other.  Broca did not fudge numbers; he merely selected among them or interpreted his away around them to favored conclusions".  It would have been fascinating to have seen Gould's take on Rushton.  Too bad his efforts were wasted on The Bell Curve which unlike Rushton really didn't have any new ideas. Minorcorrections 14:22, 31 October 2006 (UTC)


 * I think it goes further as well, since Rushton asserts that Cacuasoid primacy today and historically is function of genetics, it is hard for him to explain why we don't live in a Mongoloid dominated world (even though he asserts that is on the horizon). I believe somewhere in the Lieberman paper it mentions this internal contradiction of Rushton, although it may have been some other source that made that argument. --JereKrischel 17:31, 31 October 2006 (UTC)
 * But if we did live in a mongoloid dominated world Rushton's theory would be boring because he would just be describing the current state of things. What makes Rushton' theory interesting to me, is that he predicts what's coming. I believe China was the richest and most powerful nation in the world for over a 1000 years and was capable of discovering and colonizing Europe which was still living as barbarians, but you could argue (if you like Rushton's theory) that mongoloids were too genetically advanced to resort to criminal acts of aggression.  Rushton would probably argue that caucasoids are on top right now because they have primitive traits like agression and criminality that have allowed them to dominate, plus the Black death gave a temporary boost to the caucasoid gene pool by whiping out the poorest caucasoids.  Rushton once claimed that this gave rise to a renaissance in Europe and only now are caucasoids reverting back to their natural level, and we can expect the natural order of things to return with mongoloids back on top. 134.117.83.67 19:14, 1 November 2006 (UTC)
 * Again, you're arguing that the pattern was broken because of environmental factors, which puts into question whether or not genetic factors really matter that much. What excuse will be given when Negroids rise to power? --JereKrischel 00:43, 2 November 2006 (UTC)
 * He didn't give an environmental explanation, however genes and environment certainly interact, and when genetic differences are small, as they are between the races,  environment can sometimes overpower them.  And there's nothing in Rushton's model that precludes negroids from rising to power during certain eras,  however Rushton would expect such eras to be the exception, not the rule. Minorcorrections 06:56, 2 November 2006 (UTC)

see Needham's Grand Question. it's apparently a longstanding question as to why China fell behind Europe when they were otherwise poised to have been the ones invading Europe rather than the other way around. --Rikurzhen 00:15, 2 November 2006 (UTC)


 * Well, if given the choice between environment, and genetics, which one would you argue? :) --JereKrischel 00:43, 2 November 2006 (UTC)


 * I'm very skeptical about Rushton's suggestion than an *ancient* evolutionary difference underlies the effect. It seems much more likely to be a recent evolutionary effect, perhaps a feedback on the genome from the transition from nomadic to settled lifestyle. It's unfortunate that his last update was 2000, because a lot has happened since then to expand our understanding of human evolution in the last 100k years. --Rikurzhen 00:50, 2 November 2006 (UTC)
 * Yes there have been a lot of advances, but nothing seems to contradict Rushton's model. Some of the dates should probably moved forward, because humans may have left Africa as recently as 50,0000 years ago, instead of the 110,000 years ago figure Rushton got from Cavali-Sforza.  However I believe Rushton's sequence of negroids branching off the evolutionary tree first, followed by Caucasoids, and mongoloids branching off last, is still correct. Minorcorrections 06:56, 2 November 2006 (UTC)