December 6, 2007

Putting in a good word for the Pew Hispanic Center

In Slate, Mickey Kaus says he's dubious about some argument on Hispanic voting and the GOP advanced by the Pew Hispanic Center:

The Pew Hispanic Center reports that between July 2006 and October 2006 Hispanic voters went from 49/28 Dem-Republican to 57/23--a net Democratic gain of 13 points. In an excellent bit of 'comes-at-a-time'-ism, Pew attributes the shift to Republican anti-comprehensivism:

This U-turn in Hispanic partisan allegiance trends comes at a time when the issue of illegal immigration has become an intense focus of national attention and debate

HuffPo's normally sophisticated Thomas Edsall makes the argument less 501-c-3-ishly: "GOP Driving Hispanics Away with Anti-Immigrant Push." The problem, of course, is that the Pew Center doesn't tell us how many points the Democrats gained among non-Hispanic voters, or all voters generally. These were not good months for the GOP.

I'm perfectly willing to believe that the immigration debate has hurt the GOP among Hispanics, but without any sort of control group it's impossible to tell how much. (Gallup, for example, has the Republicans losing about 5 points among all voters over the same period--suggesting that the real Hispanic immigration backlash amounted to 8 points net, not 13.) ...

P.S.: What are the chances that the Pew Hispanic Center is going to conclude that Hispanics are not important or distinctive--they're really just like everyone else and really not worth studying much? I'd say close to zero. The study would be more credible if it came from the Pew Hellenic Center. ... 3:32 P.M. link

Personally, I've found the Pew Hispanic Center to be one of the more objective and useful sources on immigration, a topic where ignorance, lies, and wishful thinking are the norm. Robert Suro and the others at the Pew Hispanic Center are willing to publicly state, for example, that the Hispanic vote isn't as big or powerful as the media typically assume. This is "testimony against interest," because clearly the "Pew Hispanic Center" would become wealthier and more influential the more people believe Hispanics are electorally powerful, so it's particularly credible. The Pew people, to their credit, have always resisted the unspoken assumption common in the media that one Hispanic's vote somehow counts for more than one non-Hispanic's vote.

I often cite Pew Hispanic Center research. For example, here's something I wrote in last August:

And now the Pew Hispanic Center has crunched the Census Bureau's numbers for 2006 and discovered that the total Latino share of the vote last year fell—to 5.8 percent [from 6.0 percent in 2004]. According to Pew:

"while Latinos represented nearly half the total population growth in the U.S. between 2002 and 2006, the Latino share among all new eligible voters was just 20%. By comparison, whites accounted for 24% of the population growth and 47% of all eligible new voters."

Overall, whites cast almost twelve ballots for every one ballot cast by a Hispanic.

So, let's give some respect to the Pew Hispanic Center. They're not perfect but they're a lot fairer than they would have to be. Indeed, they'd probably get more media attention if they catered more to the conventional wisdom about the purported Latino Electoral Tidal Wave, rather than pop balloons about it as did their report last summer entitled "The Latino Electorate: A Widening Gap between Voters and the Larger Hispanic Population in the U.S."

So, the answer to this question is that, yes, the GOP may have lost 3 or 4 points among Hispanics on 2006 due to resistance to amnesty, but the size of the Hispanic vote is so small, that it's insignificant -- 5.8% times 4% = 0.23% -- compared to picking up votes (or at least not losing them) among the other 94% of the population.

My published articles are archived at -- Steve Sailer


seph said...

Their choice of July 2006 and October 2006 as their endpoints has more than a faint whiff of "cherry-picked data" about it. The large illegal marches were in early May and most of the Congressional debate was also earlier in the year; if they wanted to measure the flux in Hispanic support for the Republicans as a result of the amnesty debate then it would have been better to have selected an interval that enveloped those events, say from March 2006 to June 2006.

np said...

Rasmussen has the data for overall party affiliation here for comparison. It looks like the Republicans lost support during that time period.

Anonymous said...

The problem with all of this is that there are no control groups. Based on this data, someone could argue that the resistance against amnesty actually attracted voting Hispanics toward the GOP, but that other more potent forces caused the overall numbers of pro-GOP Hispanics to drop.

If we're going to be scientific about it, let's actually be scientific about it.