Validating Measures of Scanned Information Exposure in the Context of Cancer Prevention and Screening Behaviors

Abstract

Individuals may obtain health information, particularly from the mass media, without engaging in purposeful information searches (called scanning). This study used the Seeking and Scanning Behavior Survey of the General Population (SSBG), a nationally representative survey of adults aged 40–70 years (n = 2,489), to validate measures of scanned information exposure about cancer prevention and screening behaviors. Scanned exposure measures concerning specific behaviors (exercise; fruit and vegetable consumption; dieting; and mammogram, prostate-specific antigen (PSA) text, and colonoscopy screening) have good face validity and are convergent across behaviors (mean correlation across six preventive behaviors = 0.50, sd = 0.09). These measures can be discriminated from measures of general media exposure (mean r = 0.23, sd = 0.02) and seeking exposure for the same behaviors (mean r = 0.25, sd = 0.06). Scanned information exposure was associated with weekly volume of newspaper coverage for two of six behaviors, providing additional evidence of nomological validity. Scanned information exposure at the first round of measurement was associated with identical exposure 1 year later (mean r = .41, sd = .04). Scanned exposure measures also were significantly associated with five of the six preventive behaviors. These results provide evidence that scanned information exposure measures are valid indicators of the construct. Researchers might consider their use to capture scanned media influence on cognitions and behaviors.

This publication was made possible by grant number 5P50CA095856 from the National Cancer Institute. (Center of Excellence in Cancer Communication Research [CECCR], University of Pennsylvania). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. This publication was also made possible by support from the Robert Wood Johnson Health and Society Scholars Program. The authors also thank Kyung Lee, Anca Romantan, J. Sanford Schwartz, A. Susana Ramírez, Rebekah Nagler, Katrina Armstrong, Angela DeMichele, Martin Fishbein, Dominick Frosch, Stacy Gray, Shawnika Hull, Lela Jacobsohn, Annice Kim, Giang Nguyen, Minsun Shim, Aaron Smith-McLallen, Robin Stevens, and Norman Wong for contributions to the research.

Notes

¹For colonoscopy and PSA testing, respondents were first asked whether they had heard of the test. Respondents who had never heard of the test skipped to the next prevention or screening behavior. Pretest interviews revealed that nearly all females within the sampled age range had heard of the mammogram. As a result, female respondents were not asked whether they had heard of the mammogram, but were provided with a brief description of the procedure.

²One cannot measure the ability of scanned information exposure to predict other variables in its true sense with cross-sectional data, as predictive validity is defined as “the ability of a measure to predict scores on a criterion measured in the future” (Schutt, 2004, p. 114). Since we are not attempting to make causal claims, but merely to show that the scanning measures are associated with measures we would expect them to be, cross-sectional associations are sufficient in this case.

³Analyses for colonoscopy were restricted to those aged 50 and over, for whom the behavior is recommended (USPSTF, 2002; n = 1,505 or 60% of the sample). The USPSTF task force does not currently make any recommendation that men of a specific age get PSA tests. As there was no specific recommendation at the time, we decided to include all men (40 and over) in the analyses (n = 1,216 or 49%). All sampled females were included in analyses regarding mammography (n = 1,273 or 51%). Analyses for diet behavior were restricted to those with a body mass index (BMI) of 25 or higher, in accordance with the National Institutes of Health (NIH); and National Heart, Lung, and Blood Institute's (NHLBI's) definition of overweight (NIH's NHLBI, 1998; n = 1674 or 67% of the sample).

Note. Each behavior-specific scanned information exposure measure has a range of 0–4. The summary measure has a range of 0–20.

Note. Sample ns range from 1,159 for PSA-related scanning to 2,422 for variables involving both males and females.

^a Intra-scan correlations describe pair-wise correlations between scanned exposure measures (e.g., scanned exposure about mammography with scanned exposure about colonoscopy; scanned exposure about colonoscopy with scanned exposure about dieting to lose weight).

^b Scanned exposure with general media use correlations describe those between each scanned exposure measure and the general media use index.

^c Scanned exposure with sought exposure across all behaviors refers to correlations between each scanned exposure measure and each sought exposure measure e.g., (scanned exposure about mammography with sought exposure about colonoscopy, etc.).

^d Intra-behavior correlations describe pair-wise correlations between scanned exposure and sought exposure about the same behavior (e.g., scanned exposure about mammography with sought exposure about mammography, etc.).

Note. The colonoscopy analysis is restricted to individuals over the age of 49 and the diet analysis is restricted to those with a BMI of 25 or greater. Mammogram and PSA analyses are each restricted to the relevant gender.