Not sure I'm getting your point. I think your saying it's a factor because it's there. My point is the study states very clearly if you remove the negative factors all races are exactly equal. That tells me it has nothing to do with race.
On Wed, Jul 21, 2010 at 3:48 PM, Larry C. Lyons <larrycly...@gmail.com> wrote: > > Not arguing with you about that, what you're saying is very standard. > The data across the literature shows similar results. The problem is > that race is so intertwined with SES in the United States that any > causal model has to use race as a correlated predictor when including > socio-economic status as a predictor. > > Run your model through something like Lisrel or AMOS (both are > statistical programs that do causal modelling and latent variable > analyses). One of the critical analyses to look at is the goodness of > fit indicator or GFI. This chi square value ought not to statistically > significant if your candidate model is a good fit to the data. In > other words if the GFI is statistically significant there is at least > one other hypothetical model that provides a better fit to the data > than the current candidate. > > Taking your model, I would expect that you'd get at best a marginal > value for the GFI, ( .05 < p < .20). Including race as a correlated > independent variable will substantially increase the GFI, (.10 < p < > .50). > > The real difficulty with correlated models is determining the direct > and indirect influences of any one predictor on the dependent > variables, and the other independent variables downstream from it. At > best you can say that in conjunction with X, Y has a direct influence > of .zz on the dependent variables, and an indirect influence of .www. > But you cannot really determine the sole influences of Y. > > In this case we can say that Race and SES has a substantial effect on > later school achievement based on the NORC HS and Beyond dataset. But > we cannot statistically separate the effects of both variables on the > dependent variables at this time. > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Order the Adobe Coldfusion Anthology now! http://www.amazon.com/Adobe-Coldfusion-Anthology-Michael-Dinowitz/dp/1430272155/?tag=houseoffusion Archive: http://www.houseoffusion.com/groups/cf-community/message.cfm/messageid:323414 Subscription: http://www.houseoffusion.com/groups/cf-community/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/groups/cf-community/unsubscribe.cfm
