Connecting Internet Use with Gaps in Cancer Knowledge

Abstract

This study applies the knowledge gap hypothesis to the specific domain of Internet use for cancer information. In particular, this study examines disparities in online information seeking by education and ethnicity, and subsequent gaps in cancer knowledge. Perceived risk of cancer and ease of connection to the Internet are concerned as contingent factors influencing knowledge gaps. A subsample of the 2003 Health Information National Trends Survey—those who have ever been online—was analyzed. Data supported the hypothesis that high education groups and White Americans were more likely to use the Internet for cancer information than were their counterparts, and online information seeking enlarged to some degree the cancer knowledge gaps between education groups. Perceived cancer risk had a weak but significant three-way interaction effect with ethnicity and online seeking on cancer knowledge, which suggests the importance of motivation in attenuating the knowledge gaps. The moderating role of ease of connection to the Internet was not supported. Discussion about the findings and further suggestions are offered.

View correction statement:

Erratum

Notes

¹This rate was calculated using the “Standard Definitions” of the American Association for Public Opinion Research (2001, for detailed information about the HINTS administration, see Nelson et al., 2004).

²The criteria for the eligibility of knowledge items were developed to cope with limitations of secondary data analysis. The topics of knowledge were first restricted into lifestyle behaviors and screening tests. Items that were thought to measure beliefs, not knowledge, were not eligible for inclusion (e.g., how much do you agree or disagree with the statement that there is not much people can do to lower their chances of getting cancer?). In addition, to maximize statistical power, items that were either randomly assigned to a sub sample or asked only of relevant sub groups (e.g., only females got a question about mammograms and only males got a questions about PSA) were not included.

³Measures of information holding tap what respondents believe they know about cancer prevention and screening instead of what they must know (Yows et al., 1991). Including both open-ended items and close-ended factual items is expected to reduce potential middle-class bias of knowledge items (Bonfadelli, 2002; Viswanath & Finnegan, 1996).

⁴In order to reduce potential problems with multicollinearity, all variables were centered before their interaction terms were computed (Cronbach, 1987; Eveland, 1997). That is, the mean value of each variable was subtracted from the original value prior to the computation of the product terms (e.g., education multiplied by Internet use).

⁵There was one exception to this pattern, despite a weak association, such that Asian Americans and other non-Hispanic, non-African minorities showed particularly greater increase in online seeking when an easy-to-use Internet connection was available at their home (ß = .03, p< .05).

⁶In addition to the regression models reported in Table 2, four different models were tested after each type of knowledge was further divided into open-ended knowledge (i.e., knowledge held by respondents regardless of its accuracy) and factual knowledge. Patterns observed were consistent with those presented in Table 2, but respondents' education level had stronger associations with knowledge holding than factual knowledge for both cancer prevention and detection. Interestingly, the two-way interaction between education and Internet use for cancer information on screening knowledge (see Table 2) remained statistically significant when predicting knowledge holding whereas it was not significant when predicting factual knowledge about cancer screening; the three-way interaction term among ethnicity, Internet use for cancer information, and perceived cancer risk (which was statistically significant in Hispanic Americans as shown in Table 2) had a significant association with knowledge holding but not with factual knowledge about lifestyle behaviors. Further research is necessary for the relationship between online information seeking and knowledge holding.

⁷As the relationship between education and cancer knowledge was positive and the relationship between Internet use and knowledge was positive, a positive sign for the interaction term indicates a stronger relationship between education and knowledge for respondents with higher levels of Internet use.

⁸The predicted values of screening knowledge, which were acquired from multiple regression after controlling for other relevant variables, were used for the examination of these correlations.

⁹Those who have sought information about cancer on the Internet in the past 12 months (N = 1,357) were asked “Overall, how useful was the cancer-related information you got from the Internet?” (1 = not at all to 4 = very useful; M = 3.34, SD = . 64).