The Rutgers-Eagleton/Fairleigh Dickinson University Garden State Polling Partnership


Screen Shot 2019-04-01 at 12.41.39 AM

 

Media Contacts:
Ashley Koning, akoning@rutgers.edu, 848-932-8940
Krista Jenkins, kjenkins@fdu.edu, 973-443-8390

 

As response rates to telephone surveys continue to drop and cell phones continue to rise, the polling industry is at a crossroads in terms of how best to assess public opinion moving forward. Therefore, the Eagleton Center for Public Interest Polling (home of the Rutgers-Eagleton Poll) and Fairleigh Dickinson University (home of the FDU Poll) have partnered to explore the future of public opinion polling in the Garden State in an in-depth experiment involving tests of survey mode and sampling.

 

For almost 50 years, the Rutgers-Eagleton Poll – established in 1971 at Rutgers University’s Eagleton Institute of Politics – has been conducted by telephone, using what is known as a probability-based sample to survey New Jersey residents. That methodology has since been used by all other academic organizations that have conducted surveys in New Jersey – including Fairleigh Dickinson University (established in 2001), Monmouth University (established in 2005), and Quinnipiac University.

The polling landscape has dramatically transformed within the last decade, however. Due to technological changes (like cell phones and caller ID), behavioral changes (like fewer people answering their phones and responding to surveys), and an increased number of unsolicited calls (like telemarketing and spam), telephone surveys have become far more difficult and far more expensive. Response rates are now in the single digits, meaning more call attempts have to be made than ever before to achieve a single completed interview – which, in turn, means more time and more money. It now costs almost three times as much to complete a telephone interview than it did just five years ago, with fielding costs reaching over $100 per completed interview at some of the most well-known and respected telephone survey call centers.

The polling profession has started to adapt by moving online but has faced a major hurdle – the current inability to take a probability-based sample of Internet users. The industry has attempted to tackle this problem in two ways:

  • By conducting a probability sample by mail or phone and recruiting those respondents to join an online panel (with those not online being given that capacity by the survey organization). This has been the approach of organizations like the Pew Research Center and Ipsos’ KnowledgePanel, the latter of which was used for this current study.
  • By conducting a non-probability sample, where respondents volunteer to be surveyed rather than the probability sample where they are selected to be surveyed. The New York Times/CBS News Poll took this approach in 2014, for example.

A number of research studies have found that the results of probability and non-probability samples are similar, if weighted correctly at the end. But probability samples are still slightly more accurate, may have better reliability over time, and allow for the computation of sampling error – a statement of the probabilities of how likely the poll is to be accurate.

Because of the need to move away from telephone surveys, the Rutgers-Eagleton Poll at Rutgers-New Brunswick’s Eagleton Institute of Politics and the FDU Poll at Farleigh Dickinson University have combined their resources to conduct one of the first ever in-depth experiments testing the effects of both survey mode and type of sample on statewide public opinion polling. The extensive study involves testing an identical questionnaire on three different samples:

  1. A probability-based sample of 621 respondents from a traditional dual-frame telephone survey conducted by live callers on both landline and cellular phone between March 7 and March 12, 2019. The telephone survey was fielded by Braun Research, Inc with sample provided by Dynata.
  2. A probability-based sample of 629 respondents from Ipsos’ online probability-based KnowledgePanel® conducted online between March 13 and March 22, 2019.
  3. A non-probability sample of 643 respondents from Ipsos’ opt-in panel conducted online between March 17 and March 28, 2019.

The results reported on in this series of releases by Rutgers-Eagleton and FDU will report results only from the combined samples of the telephone survey and online probability-based panel. The questionnaire was developed and all data analyses were completed in house by Dr. Ashley Koning and Dr. Cliff Zukin at the Eagleton Center for Public Interest Polling (ECPIP) at Rutgers University-New Brunswick and Dr. Krista Jenkins at Fairleigh Dickinson University. William Young and Kyle Morgan assisted with preparation of the questionnaire and analysis and preparation of this release. This poll is paid for and sponsored by both the Eagleton Institute of Politics at Rutgers University-New Brunswick and Fairleigh Dickinson University.

 

POLITICAL LEADERS NOT REALLY DOING IT FOR GARDEN STATERS

When it comes to likeability, New Jersey’s current and former elected officials leave a lot to be desired, Garden State residents say. The inaugural joint survey from the polling units at Rutgers University’s Eagleton Institute of Politics and Fairleigh Dickinson University finds that not a single politician scores a favorable majority with Garden State residents.

 

NEW JERSEYANS SUPPORT MILLIONAIRES TAX; GOV. MURPHY GARNERS LACKLUSTER RATINGS ENTERING SECOND YEAR IN OFFICE

New Jerseyans largely support Gov. Phil Murphy’s proposed millionaires tax, but expressed mixed views about the governor’s overall performance, according to the first poll in a partnership between the polling bodies at Rutgers University’s Eagleton Institute of Politics and Fairleigh Dickinson University.