Outlook: Newsletter of the Society of Behavorial Medicine

Fall 2020

Outlook main page »

We Don’t Know What They “Don’t Know”: Why Researchers Should Allow “Don’t Know” Responses to Perceived Risk Survey Items

Erika A. Waters, PhD, MPH✉; Jennifer L. Hay, PhD✉; Marc T. Kiviniemi, PhD✉; Heather Orom, PhD✉; Health Decision Making SIG


The premise that higher risk perceptions motivate protective behavior is central to many health behavior theories1 and is supported by meta-analyses2. But what about people who don’t know or who are uncertain about their risk? In a research program spanning nearly a decade, we have discovered that treating uncertainty about disease risk as a nuisance instead of a meaningful response threatens the validity and generalizability of health behavior theories and risk perception research for five main reasons.
 

  1. Uncertainty about disease risk, as indicated by don’t know (DK) responses, is prevalent in the U.S. population. DK responding ranged from 1.6% to 69.3%3-7 depending on the study population, whether a DK response option was or was not provided, and the health risk in question.
  2. We found no evidence that DK responding is due to participants’ reluctance to think carefully about the survey items (i.e., “satisficing”).5, 6
  3. DK responding occurs more often among people with limited formal education, health knowledge, health literacy, and numeracy, or who are members of marginalized racial or ethnic groups3-5, 7-9. There is also strong evidence that DK responders tend to avoid threatening health information.5, 10 Thus, excluding DK responses from a dataset reduces the relevance of the study for the very populations who are most at risk of having poor health and who may have the least health knowledge. It also violates the assumption that “missing data” from DK responses is missing completely at random, which means that it is inappropriate to impute the missing perceived risk values or to exclude participants with missing values from analyses.11
  4. DK responding is associated with less engagement in health behaviors such as physical activity, health information seeking, and adherence to breast and colorectal cancer screening guidelines.4, 12 It is also associated with lower enrollment in research studies that provide personalized genetic risk information.13
  5. Not including a DK option when asking participants their perceived risk for a disease, results in spuriously low levels of perceived risk in a sample, and may obscure the perceived risk-behavior relationship.6
     

Our research, and that of others14, suggests that DK responding is more of a meaningful response option than a nuisance and, in fact, may highlight a unique population in need of specialized intervention. That it affects members of populations that are marginalized suggests that ignoring it may exacerbate disparities. We recommend that researchers add DK response options to survey items assessing perceived risk and that they examine whether their results differ based on whether participants did or did not mark DK.

As researchers expand investigation into the nature, meaning, and consequences of DK responding, this will (1) clarify our conceptual understanding of perceived risk as a psychological phenomenon and its relation to health behavior, (2) improve the ability of our theories to describe, explain, and predict health behavior, (3) increase the effectiveness of health behavior interventions, and (4) develop risk perception interventions specifically targeted at individuals who are uncertain about their risk.

 

References

  1. Conner M, Norman P, editors. Predicting Health Behaviour. Buckingham/Philadelphia: Open University Press; 1995.
  2. Sheeran P, Harris PR, Epton T. Does heightening risk appraisals change people's intentions and behavior? A meta-analysis of experimental studies. Psychol Bull. 2014;140(2):511-43.
  3. Waters EA, Hay JL, Orom H, Kiviniemi MT, Drake BF. “Don’t Know” Responses to Risk Perception Measures: Implications for Underserved Populations. Med Decis Making. 2013;33(2):271-81.
  4. Hay JL, Orom H, Kiviniemi MT, Waters EA. "I don't know" my cancer risk: exploring deficits in cancer knowledge and information-seeking skills to explain an often-overlooked participant response. Med Decis Making. 2015;35(4):436-45.
  5. Orom H, Schofield E, Kiviniemi MT, Waters EA, Biddle C, Chen X, et al. Low Health Literacy and Health Information Avoidance but Not Satisficing Help Explain "Don't Know" Responses to Questions Assessing Perceived Risk. Med Decis Making. 2018;38(8):1006-17.
  6. Kiviniemi MT, Ellis EM, Orom H, Waters EA, Hay JL. 'Don't know' responding and estimates of perceived risk: failing to provide a 'don't know' response systematically biases laypeople's perceived risk estimates. Health Risk Soc. 2020;22(1):69-85.
  7. Waters EA, Klein WMP, Moser RP, Yu M, Waldron WR, McNeel TS, et al. Correlates of unrealistic risk beliefs in a nationally representative sample. J Behav Med. 2011;34(3):225-35.
  8. Orom H, Kiviniemi MT, Shavers VL, Ross L, Underwood W, 3rd. Perceived risk for breast cancer and its relationship to mammography in Blacks, Hispanics, and Whites. J Behav Med. 2013;36(5):466-76.
  9. Kiviniemi MT, Orom H, Waters EA, McKillip M, Hay JL. Education-based disparities in knowledge of novel health risks: The case of knowledge gaps in HIV risk perceptions. Br J Health Psychol. 2018;23(2):420-35.
  10. Hay JL, Kiviniemi MT, Orom H, Waters EA. Using NCI-Designated Cancer Center Catchment-Area Data to Understand an Ignored but High-Need Constituent: People Uncertain or Avoidant about Their Cancer Risk. Cancer Epidemiol Biomarkers Prev. 2019;28(12):1955-7.
  11. Graham JW. Missing Data Analysis: Making It Work in the Real World. Annu Rev Psychol. 2009;60:549-76.
  12. Waters EA, Kiviniemi MT, Orom H, Hay JL. "I don't know" My Cancer Risk: Implications for Health Behavior Engagement. Ann Behav Med. 2016.
  13. Hay JL, Meyer White K, Sussman A, Kaphingst K, Guest D, Schofield E, et al. Psychosocial and Cultural Determinants of Interest and Uptake of Skin Cancer Genetic Testing in Diverse Primary Care. Public health genomics. 2019;22(1-2):58-68.
  14. Bruine de Bruin W, Carman KG. Measuring Risk Perceptions: What Does the Excessive Use of 50% Mean? Med Decis Making. 2011.