ACCURACY OF RECENT POLLS IN QUESTION

A recent round of political polls commissioned by the website CTCapitolReport and conducted by the Merriman River Group of Hamden, uses a method of contacting “likely voters” that is being questioned by some polling experts.  Add the fact that the pollster’s work has been challenged before and you have serious questions about the polls’ accuracy

The polls have gotten a lot of media attention and have raised some political eyebrows in the state as they have shown some races surprisingly close and even one that shows the Republican challenger leading the Democratic incumbent when even the challenger himself hasn’t claimed his own internal polls have him leading.

Merriman River uses automated calling for its polling which can be unreliable and have a high degree of uncertainty leading to skepticism about the results says Jeffrey Stone-Cash, professor of political science at the Maxwell School of Citizenship at Syracuse University. Stone-Cash was a pollster for 25 years before joining the faculty at Syracuse.

“The problem is automated calls have a very high refusal [hang-up] rate and a very high degree of self-selection,” Prof. Stone-Cash said. “You don’t know if the person answering the phone and taking the survey is a 10-year old boy or the person targeted. I’m skeptical [of polls using automated calls],” he said. The Quinnipiac University polls use live callers.

Merriman River’s polling executive director Matthew Fitch defends his polls saying, “Most people mistrust what they are not familiar with.” In the written news releases accompanying the polls, Merriman River Group says its technology is new. “In 2010, Merriman River Group launched their NexGen IVR polling system, which offers professional grade polling and analysis with reduced costs and faster completion time,” the narrative says.

Fitch also minimizes the number of “wrong responders” the polls get. “Political polls are not that interesting and not really something a child or a house guest would answer,” he said.

The two CTCapitolReport polls that have generated the most talk are the ones in the 1st and 5th congressional districts. The poll in the CT-1st has incumbent Democrat John Larson leading Republican challenger Ann Brickley, a hard-working but largely unknown candidate, by just seven points. Larson, along with Rosa Delauro , is considered one of the safest incumbent in the state.

The other CTCapitolReport poll in dispute is the CT-5th where it has incumbent Democrat Chris Murphy losing to Republican state Senator Sam Caligiuri, 49.7-percent to 44.3-percent. A poll done by the Democratic Congressional Campaign Committee shows Murphy ahead by 13-points. One poll or the other is dead wrong in this race.

This isn’t the first time a poll done by the Merriman River Group has been challenged. A year ago, Merriman River did a poll in Bridgeport about the gubernatorial race (and it was then) and the Bridgeport mayoral race. Experts said in no uncertain terms that it was statistically impossible for the poll to have the margin of error it claimed given the sample size it claimed.