" The Reform Movement is shrinking rapidly because it's been too tolerant of interfaith marriages."

Yes, that would explain why it went from being the second-largest movement in the 1990 National Jewish Population Survey (NJPS) to being the largest movement in the 2000-1 NJPS. 

For the record, I'm affiliated with the movement that used to be the largest and has fallen to #2 (1990 percentages: 35% Reform, 43% Conservative; 2000-1 percentages: 39% Reform, 33% Conservative.) --a change which many have attributed, not without reason, to Conservative Judaism's perceived unfriendliness to intermarried families. I discuss a roundup of statistics dealing with intermarriage and Conservative Judaism here , for those who like the details--but the basic deal is: the data from the 1990 NJPS on those who switched their affiliation from Conservative Judaism to another movement or another religion showed that 58.9% went to Reform Judaism. Why? "Our findings thus suggest that switching is very often related to intermarriage, and quite likely is the direct result of intermarriage," write Sidney Goldstein & Alice Goldstein, in their analysis of the data. (And if you like stats on adult children of intermarriage, to have a better sense of whether folks like me and Laurel are anomalies or not, you might like to have a look here.)

 Let's remember, of course, 2 important caveats about the use and abuse of statistics:

1) correlation is not causation

--for example, if your data shows that those who have attended Jewish day school have lower levels of intermarriage than those who don't, you can't conclude that day school "makes you less likely to intermarry," so we can end intermarriage by sending everyone to day school: it's likely that the factor(s) that led a family to enroll their child in day school might be the same one(s) responsible for the relatively lower rate of intermarriage--so to know what the cause is, you'd have to either run an experiment or have other ways to do comparisons or gather survey data that would help you determine causation.

2) description is not prediction

--if your data shows that about 1/3 of intermarried households surveyed in 2000 were raising their children exclusively Jewish (a number often cited from the 2000-1 NJPS, though others would say it should instead be nearly half (42%), since the data showed: 32% Jewish (religion), 10% Jewish (secular)--vs. 11% Jewish + other, 35% Christian, 8% no religion, 4% other-- I refer you once more here), you can't necessarily say that "only 1/3 of the children from mixed marriages will end up Jewish," as some sort of rule for predicting future affiliation. What we know about a population from one data snapshot in time gives us a basis for extrapolation and prediction--which we'll do better or worse at, depending on how well we consider all the relevant factors that might affect the situation over the timespan involved--but it doesn't predict destiny. So you can't confidently say that "90% of children who have one Jewish grandparent will end up losing their Jewish identity forever" (never mind the fact that the main issue is whether such children are given a Jewish identity in the first place!)