Tuesday, July 05, 2011

Why Gallup's Analysis of Jewish Voters is Dishonest



Yid with Lid

There it was in bold Gallup headlines. Solid Majority of Jewish Americans Still Approve of Obama Changes in approval among Jews continues to reflect broader U.S. patterns. That was the big Gallup news this morning but it's not true.

I assume the overall Jewish Americans numbers were correct, they gave President Barack Obama a 60% job approval rating in June, down from 68% in May, but statistically unchanged from 64% in April and drastically below the 83% at the beginning of 2009, but its the comparison to the general population that is disingenuous.

Gallup's monthly trend in Jewish approval of Obama continues to roughly follow the path of all Americans' approval of the president, more generally, as it has since Obama took office in January 2009. The 14-percentage-point difference in the two groups' approval ratings in June -- 60% among U.S. Jews vs. 46% among all U.S. adults -- is identical to the average gap seen over the past two and a half years. However, the monthly graph is somewhat variable due to the lower monthly sample size of Jewish respondents (around 350).But the latest Gallup figure I could find for Jewish party affiliation (Sept 2009) shows that American Jews are almost 50% more likely to be Democrats than the average American


Why is this important? Because Obama's approval rating has not fallen of as much for Democrats as it has for Republicans. Using numbers from April 2009 and May 2011 Gallup reports we find that the Democrat approval fall off is less than a third of the total US.


The Chart below is for illustrative purposes only, and not does not necessarily reflect the actual Gallup numbers which at this point are not available. If one were to adjust the total US numbers to reflect a party split similar to that of the Jewish voters above (66% Dem, 27% GOP, 7% Independents) and compare that number to the fall off of the Jewish vote we find



Using the May numbers (which are available but different than the ones Gallup reported today) we find that the adjusted fall off of Obama approval is much less than that of the "total US figures (21% to 34%). We also find that the fall off of Jewish approval of President Obama is more than twice that of adjusted numbers. While I wouldn't use the exact numbers I am confident in saying that the fall off of the President's approval, is much larger than that of the equivalent general population. In other words, Gallup's main proposition is wrong.

With a major organization such as Gallup when there is a "bogus" survey, it usually is in the sample mix (i.e. weighted to heavy toward one group), or the analysis.

Without having the sample mix for Gallup, I cannot comment on that, but the entire set-up of the analysis is misleading at best, intentionally dishonest at worst. First of all, taking a look at the approval point spread difference as Gallup does at the start of its analysis doesn't make sense because the Total US and Jewish Vote fall offs are working off of different bases. Ten percentage points off of 60% is a bigger hit than 10% off of 80% something that Gallup does not acknowledge.

More important is the fact that comparing the comparing President Obama's loss of approval in the Jewish community, which is 66% Democratic, to the loss of approval of the entire population which is only 43% Democratic is not an honest comparison because the President has maintained his approval much better amongst Democratic voters.

The old proverb says there are three kinds of lies: lies, damned lies, and statistics. Today Gallup's Jewish vote survey has proved that proverb very true.

No comments: