Underpinnings of Policy Reseach Satisfaction


New York – In a recent ResearchFocus report of policy research, Integrity Research conducted analysis and a survey of buy-side clients of policy research. In designing the survey, we made judgments as to what the potential characteristics were that affected the overall client satisfaction of the services.  These characteristics included:          

  1. Research quality
  2. Value for money spent
  3. Experience of the analysts
  4. Access to the analysts
  5. Accuracy of forecasts or recommendations
  6. Ability of the research provider to access key government officials
  7. Ability to generate actionable ideas
  8. Performance of the ideas
  9. Compliance Practices
  10. Institutional Sales coverage

Our correlation analysis found that several of these factors were unimportant to the overall satisfaction of the user of policy research, including: ability to access to key government officials; accuracy of forecasts; access to analyst, and; compliance practices.  For compliance practices the correlation coefficient was -0.1514.  This raised the issue of whether compliance is really an unimportant factor in overall satisfaction, or whether it is actually a negative factor. For example, buy-side clients could believe that the quality of the information is eroded by excessive compliance practices.

To test this, we set out the conduct a regression analysis of these factors. The regression allows us to test the statistical significance of the regression coefficients. If, for example, compliance is negatively correlated to overall satisfaction of policy research, then the t-score of compliance coefficient should be below 2.0 (95% level of confidence). In fact the t-score was -0.61 indicating that compliance is not an important factor in the overall satisfaction of clients with policy research providers.

The top three factors in customer satisfaction from our regression analysis were: research quality, experience of the analysts and the ability to generate actionable ideas.  All of these factors were significant at the 90% level of confidence.


About Author

Leave A Reply