Rumors About Cancer: Content, Sources, Coping, Transmission, and Belief

Abstract

Using a sense-making and threat management framework in rumor psychology, the authors used an exploratory web survey (n = 169) to query members of online cancer discussion groups about informal cancer statements heard from nonmedical sources (i.e., cancer rumors). Respondents perceived that rumors helped them cope. Dread rumors exceeded wish rumors; secondary control (control through emotional coping) rumors outnumbered primary control (direct action) rumors. Rumor content focused on cancer lethality, causes, and suffering. Rumors came primarily from family or friends in face-to-face conversations. Respondents discussed rumors with medical personnel primarily for fact-finding purposes, but with nonmedical people for altruistic, emotional coping, or relationship enhancement motives. Transmitters (vs. nontransmitters) considered rumors to be more important, were more anxious, and felt rumors helped them cope better, but did not believe them more strongly or feel that they were less knowledgeable about cancer. Most respondents believed the rumors; confidence was based on trust in family or friends (disregarding source nonexpertise) and concordance with beliefs, attitudes, and experience. Results point toward the fruitfulness of using rumor theory to guide research on cancer rumors and suggest that rumors help people achieve a sense of emotional control for dreaded cancer outcomes, inform the social construction of cancer, and highlight the continuing importance of nonelectronic word of mouth.

Acknowledgments

This material is based upon work supported by the National Science Foundation under Grant No. BCS-0527371. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. A portion of the findings in this article were presented in the second author's master's thesis.

Notes

¹All research reported here was approved by a university institute review board.

Note. A total of 190 elements were unitized from 169 rumor statements (17 statements contained two elements; 5 contained three elements).

²The parenthetical clarification “This information … ” was added after the first 18 respondents to increase the proportion of responses that included cancer rumors.

³To check reliability, a blind judge subsequently categorized the 190 units using a list and description of all subcategories; agreement was very high (κ = .84, n = 187).

⁴Interpretation of Fleiss's κ followed Landis and Koch (1977, p. 165) and Viera and Garrett (2005, p. 362).

⁵The first three categories were generated a priori on the basis of rumor literature (discussed earlier); the last two arose from common keywords and phrases in the responses.

⁶Mann-Whitney U ps for predicted directional effects are one-tailed; otherwise, they are two-tailed. Effect size rs throughout text were calculated following Field (2009) or Rosenthal and Rosnow (2008).

⁷No differences across minority status were evident.

⁸There was a significantly higher proportion of dread rumors for minorities (86% of 51) than nonminorities (66% of 90), χ²(1) = 7.10, p < .008, but no difference for control type.

⁹Percentages sum exceeds 100 because responses were not mutually exclusive.

Note. Results are based on rumor transmission to nonmedicals.

*p < .05. ***p < .001. All p values are one-tailed.

¹⁰Transmission antecedents for nonminority transmitters (n = 86) and nontransmitters (n = 15) were comparable, but transmitters (median = 4.00) were also more confident than nontransmitters (median = 3.00; U = 408.50, z = −2.34, p = .02); ns were too small for comparisons between minority transmitters and nontransmitters. Across minority status, minorities (median =3.00) had more cancer uncertainty (vs. nonminorities; median = 3.00; U = 2201.00, z = −2.34, p = .02), but no differences in importance, anxiety, or confidence.

¹¹No transmission or motivation differences across minority status reached significance.

Note. N = 139 codeable responses. No differences were obtained across minority status.