Triangulating Differential Nonresponse by Race in a Telephone Survey
2007

Assessing Racial Nonresponse in a Health Survey

Sample size: 3543 publication 10 minutes Evidence: moderate

Author Information

Author(s): Jessica T DeFrank, J. Michael Bowling, Barbara K Rimer, Jennifer M Gierisch, Celette Sugg Skinner

Primary Institution: The University of North Carolina at Chapel Hill

Hypothesis

Does differential nonresponse by race occur in the PRISM study when race data is incomplete?

Conclusion

The study found slight differential nonresponse by race, indicating that black women were less likely to be reached for participation compared to white women.

Supporting Evidence

  • The E-Tech method had moderate sensitivity (48%) but high specificity (97%) in estimating race.
  • Black women comprised only 10.7% of study participants, significantly lower than the expected 23%.
  • Interviewer estimates had high sensitivity (100%) and specificity (95%) for identifying race.

Takeaway

The study looked at how many black and white women participated in a health survey and found that fewer black women were involved. This means researchers need to find better ways to include everyone.

Methodology

The study used two methods to estimate race: E-Tech, which analyzed names and zip codes, and interviews with a subsample of study refusals.

Potential Biases

There is a risk of nonresponse bias due to the underrepresentation of black women in the study.

Limitations

The E-Tech method misclassified many black participants as white, and the sample size for validating the interviewer method was small.

Participant Demographics

Participants were primarily insured women aged 40 to 75, with 10.7% identifying as black.

Statistical Information

P-Value

p<0.05

Confidence Interval

95% CI, 0.48–0.58

Statistical Significance

p<0.05

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication