A closer look at the methodology of the Brown University poll

October 10th, 2012 at 12:36 pm by under Nesi's Notes, On the Main Site

The 496 interviews used for Brown University’s new survey were conducted in two sets over 10 days.

The survey results combine findings from two groups of interviews conducted almost a week apart, the first done from Sept. 26 to 29 and the second done from Oct. 4 to 5, according to Marion Orr, the professor who oversees Brown’s polling. “We did two consecutive sets of days,” Orr told WPRI.com.

Asked whether the poll included cell phones as well as landlines, Orr said the call list was created using phone numbers that some Rhode Islanders voluntarily write down on the official voter registration forms they file with the secretary of state’s office.

A breakdown provided by Brown shows the 496 survey respondents identified themselves as 45% independents, 37% Democrats and 10% Republicans, with an additional 6% of voters classified as “other” and 3% who didn’t know or didn’t say. As a comparison, the 501 voters surveyed in last week’s WPRI 12 poll were 40% independents, 41% Democrats and 17% Republicans, with another 2% who refused to say.

The Brown poll’s gender breakdown was 50.2% female and 49.8% male, and its age breakdown was 10% under 30; 18% ages 30 to 44; 31% ages 45 to 60; 28% ages 61 to 74; and 11% ages 75 and older. Orr said the university trains and pays undergraduate and graduate students to conduct the survey interviews.

• Related: Cicilline leads by 6 points in new Brown poll; Republicans trail (Oct. 10)

This post has been expanded.

Tags: , , , ,

3 Responses to “A closer look at the methodology of the Brown University poll”

  1. [...] said the telephone survey of 496 registered Rhode Island voters was conducted over 10 days, from Sept. 26 to Oct. 5, and has a margin of error of plus or minus 4.4 percentage points on [...]

  2. [...] Nesi has more on the methodology of the Brown poll. Like this:LikeBe the first to like this. from → Uncategorized ← [...]

  3. Vic Profughi says:

    Congratulations to Ted and WPRI for getting this methodological info out of Brown. That said, the methodology is even worse then I could have imagined. Numbers drawn from the voting lists–my goodness, no reputable pollster in the nation would ever do that! As many as 1/4th or more on the lists do not provide their information, and there is no way of knowing if the numbers actually match real voters or the community’s they identifiy with. Polling over ten days–a farce. Reputable firms do them in two to four, tops. Why? Because responses are time sensative; at best reflecting the time snap shot when the question is asked. Wish the media would stop reporting this junk as legitimate surveys.