Harris and KN Spar over Stanford Study

US agencies Knowledge Networks (KN) and Harris Interactive are arguing over KN's claims that Stanford University research demonstrates the superior accuracy of its online surveys. Harris believes KN's reporting of the study is inaccurate and selective, and claims that KN's founders were closely involved with the Stanford research.

In August, KN publicised the results of a study conducted by the Stanford Institute for Quantitative Studies in the Social Sciences, claiming it proved KN's online panels were more accurate than the competition. The study, Comparing the Results of Probability and Non-Probability Sample Surveys, involved a comparison of survey data collected by telephone and via the Internet - using both non-probability and probability samples. The firms compared were GoZing, Greenfield Online, Harris Interactive, Knowledge Networks, SPSS, SRBI, SSI and Survey Direct.

KN claims to maintain the only online panel based on statistically valid sampling rather than volunteering. It states that the study shows that KN surveys produced 'results that are dramatically more valid and reliable' than other methods. It says 'KN's overall error rate for key demographic, attitude and product usage points was 3.9%, outperforming the telephone study (6%), Greenfield (7.7%), and Harris (6.5%)'.

KN goes on to give examples of how its figures compare with benchmarks taken from official data. For example, it claims that, 'while the KN (49.6%) and telephone (45.7%) data show that nearly half of respondents do not use coupons in a typical week, the firms using volunteer samples came in with a much lower average estimate of 26.7%'.

However, according to Dan Hucko, Senior VP, Corporate Communications at HarrisInteractive, 'the results that KN has been communicating as final appear to deviate from the preliminary results that were presented by the Stanford researchers'. In fact, he says, the Stanford team's summary of the research, presented at the American Association for Public Opinion Research (AAPOR), stated that there was a 'striking similarity of distributions across nearly all questions'. And for some of the questions listed in the presentation, KN achieves higher error rates than either Harris or Greenfield.

Harris Interactive also points out that Douglas Rivers, a Stanford professor involved in the project, is the Co-Founder and Chairman of KN, and that Norman Nie, the Director of the Stanford Institute for Quantitative Studies in the Social Sciences, is the company's Co-founder and Vice-Chairman.

Knowledge Networks is online at www.knowledgenetworks.com . Stanford's presentation of the research is available at http://communication.stanford.edu/faculty/Krosnick/2005%20AAPOR%20Presentation.ppt.

All articles 2006-23 written and edited by Mel Crowther and/or Nick Thomas, 2024- by Nick Thomas, unless otherwise stated.

• Knowledge Networks
• All news
• MrWeb info