The main takeaway from the Laszlo Strategies poll article (“Just how connected are young people really?” WJW, Aug. 29) — younger Jews care less about Israel than older Jews, despite having had more Jewish education and summer camp — is hardly news. Much prior research (by Steven M. Cohen and others) arrives at similar conclusions about the diminished commitment to Israel among younger American Jews.
Beyond the ho-hum story line, however, what struck me most about the article was the way the poll was conducted. It’s one thing for Laszlo Strategies to present their results as based on a sample from Jerusalem U and other Jewish databases, and thus “not a random poll.” This, in itself, casts doubt on the survey’s validity but might be defensible had external evidence been presented to support the soundness of the sample. For one thing, having one’s email address on a Jewish database suggests at least some minimal level of Jewish involvement and would not capture many on the periphery of American Jewry.
In the very next sentence, after honestly admitting the poll’s questionable representativeness, they try to discount this by stating: “However, when the survey became large enough to include a range of ages and denominations, it became statistically valid.” I’ve worked on polls and surveys during the three decades of my professional life. Every serious statistician I know would agree that such a claim is utter nonsense. Statistical validity is purely dependent on how well the sample reflects the target population. This is largely a function of the quality of the starting sample and the survey response rate (not reported). It has little to do with the number of respondents.
The statistics cited in Laszlo Strategies’ research are called into question by this dubious methodology.
SID GROENEMAN, Bethesda
