Statistically speaking, correlation doesn't necessarily mean causality.
If Respect's appeal were strongest amonst the poorest people, then in some areas of the country that would mean a correlation beteen voting and the proportion of the population who are muslim, as there is a strong link between deprivation and poverty and the proportion of muslims. Certainly the correlation between muslim population and Respect vote is very strong in London - no-one denies that, and certainly those of us in Respect have no problem with it, as it a reflection of Respect's appeal to both the most anti-war and the poorest populations. Statistically, muslim millionaires are not a significant part of the population!
Elsewhere the correlation is not as great. In the North West for example, for the Euro-elections in 2004, there was a stronger correlation between Muslim population and the Liberal Democrat vote than the Respect vote. No-one would suggest that the LibDems are a muslim dominated party, though it is also clear that as the loyalties of muslims to Labour has been shattered by Iraq, some of it has been picked up the LibDems (eg LibDems easily won Rochdale from Labour on the back of an anti-war muslim vote).