Former Employees: Frank Luntz's Research Is a Scam

His job is to put together a focus group, which is supposedly semi-representative of public opinion, and ask them questions.

His former employees say he's very pushy about making the focus group give the answers he wants.

Influential GOP pollster Frank Luntz's research is a 'total schtick and scam' that relies on manipulated focus groups, former employees say

...

Former employees of the influential Republican pollster and strategist Frank Luntz say that his methods are a 'scam', according to a new report.

Three former Luntz employees, including Chris Ingram, a former senior vice president at the Luntz Research Company, spoke out about the pollster's practices in interviews with Salon.

Ingram, who worked for Luntz from 1997 through the early 2000s, said that Luntz's claim to deliver objective data is a 'total shtick and a scam' and called his methods 'quite frankly bulls**t.'

Luntz did not immediately respond to a request for comment from DailyMail.com on Saturday.

...

In recent days, his focus group work has been featured on The Daily Show and the New York Times opinion section, and he makes regular appearances on MSNBC.

'The clowns at MSNBC didn't have a clue about how the focus groups or panels worked, or what Frank what was doing,' Ingram told Salon.

'The actions were basically contrived: He screened out anybody that isn't going to give the viewers the opinion that Frank, on behalf of his client, is looking for. Somehow, he is able to bulls**t people,' he added.

Ingram said that Luntz tried to manipulate his focus groups using 'dial testing,' in which participants turn knobs on a dial to indicate their reactions in real time.

'Frank, when he would be hired by clients, whether they would be corporate or political, would sit in that room yelling, 'Keep turning the dials! Keep turning the dials!'' Ingram said.

He said that Luntz was mainly concerned with results that would yield more 'compelling' data for the client, rather than accuracy.

I know a bit about this subject. I completely believe the part about looking for "compelling" data over "accurate" data, because pollsters I know say that political pollsters do this all the time.

Here's an example: Suppose you poll 450 people. Scanning down your break-outs by race, age, and sex, you see that support for a candidate has surged 20 points in the last month among Hispanic women.

So obviously you write up the report noting this huge increase in support from Hispanic women, and speculating about possible reasons for it.

The thing is, though: You absolutely shouldn't. The 450 total respondents are enough to say with 95% confidence that your numbers are within 5% of the real number. Which isn't very accurate already.

But how many Hispanic women did you poll? 30? 40? Maybe 60, tops, if you deliberately oversampled Hispanic women?

The margin of error for such a small sample is... I don't even know if they'd bother figuring it. It's a huge margin of error.

These numbers are worthless, and no one should base anything on them.

But pollsters do routinely write reports highlighting such meaningless statistical noise, because clients would rather have "grabby" findings and confident recommendations rather than accurate ones. An accurate reading of this hypothetical would be to say "Although the data show an increase of support with Hispanic women, this is statistically meaningless and should be discounted."

Clients are paying for boring but accurate stuff like that.

I once interviewed for a position with a firm that did consulting, and the guy told me straight-up, "Be decisive, be bold, clients aren't paying for 'maybes.'"

So I totally believe Luntz is doing the same thing. Most of them do this.

And I believe it's worth with Luntz, because he's not doing polls. He's doing mere focus groups with 6, 8, 10, or 12 people per panel. There is not even a calculable margin of error with such tiny samplings.

And on top of that, he seems to be a liberal partisan taking money from the Republicans and then using his position to steer the Republican Party into a direction that this liberal would prefer it to take, rather than what Republican voters would prefer.

Ever wonder how it can be that the Republican voter base is so completely, one-sidedly on the side of border security, and yet the Republican establishment keeps presenting us with an endless series of amnesty proposals and no border wall, no deportations?

Partly that's due to the Republican Party being nothing more than a paid political wing of the US Chamber of Commerce, but part of that must surely be due to liberal ringers in the DC Establishment who are forever telling the GOP that the only path to victory is through Corporate Liberalism.

These corrupt courtiers need to all be turned out.

posted by Ace at 05:36 PM

| Access Comments