Flawed Poll

By Barry Rascovar

October 19, 2015 – True or false: Rep. Elijah Cummings of Baltimore would easily defeat the two most prominent contenders for the U.S. Senate seat being vacated by Maryland Democrat Barbara Mikulski next year.

A Flawed Poll

Rep. Elijah Cummings of Baltimore

If you believe the Washington Post poll published last Friday, the answer is “true.”

But don’t believe everything you see in polls, especially polling snapshots that contain serious and disturbingly invalid tabulations.

The Post poll showed Cummings with 33 percent of the vote against Rep. Chris Van Hollen (20 percent) and Rep. Donna Edwards (20 percent), the two declared main contenders for Mikulski’s seat in next April’s Democratic primary.

The results were in line with a private poll commissioned by Cummings last spring.

But if you delve deeper into the poll’s methodology, there is reason to question its reliability.

Only 550 people were surveyed on the Senate question, a small number. Most established polling organizations insist on a sample of 1,000 to 2,000 respondents to get accurate snapshots of voter sentiment.

Poor Decision

Far worse was the decision by the Post and its partner, the University of Maryland, to ask the Senate question to both registered Democrats and Democratic-leaning independents.

Note to the Post and UM: Registered independents cannot vote in Maryland primaries. So why in the world would you include them in a survey of voter sentiment on the Democratic Senate race?

The Post’s Senate poll results are tainted.

It may be that even when independents are removed from the tabulations, the numbers stay roughly the same – though the sample then might be too small to accurately gauge true Democratic sentiment.

Independent voters represented one-third of the people surveyed by the Post and UM for this extensive poll. That means the number of Democrats who were asked the Senate question might be quite small, perhaps only 350 or so individuals.

There’s also the problem of polling too early in the election cycle.

Governor Brown?

Well-known names always score best when balloting is far, far away.

In prior elections, Kathleen Kennedy Townsend and Anthony Brown looked like runaway victors in early polls. Both failed miserably on Election Day.

Cummings says he intends to make his decision on a Senate race this fall. He’s consumed right now by his role as chief defender of former Secretary of State Hillary Clinton against smears from House Republicans over the death of the U.S. ambassador in Benghazi, Libya, three years ago.

He’s also a national spokesman on African-American issues and a prime defender of the Obama administration against determined Republican attacks in the House of Representatives.

It could be a difficult choice for Cummings, whose reelection to his congressional seat is a slam-dunk. The Post’s fatally flawed Senate poll won’t be of much use in making that decision.

###

One thought on “Flawed Poll

  1. Patrick McGrath

    Barry…Excellent column that needed to be written. I ran into Rep. Van Hollen Sunday and asked him about the poll. He dismissed it as flawed too.
    Van Hollen says when he polls, his pollsters limit it to Democrats who have voted in past primaries…people who are actually likely to vote. He didn’t mention the Independents you said were stupidly included in the Post poll, but he said the Post poll, included registered Democrats…regardless of their past participation in primaries.
    He said those contacted, after being asked which candidate they favored, were asked a second question…did they plan to vote in next year’s primary. Van Hollen says when people are asked that kind of question, after already having given their opinion about whom they favor, they will almost always say yes they will vote, even if their past history indicates they seldom do.
    He also added that the election is so far away, that he is not yet widely known in Baltimore. He said he expects that to change when he begins to advertise and campaign more heavily in the Baltimore area.

Comments are closed.