Selective Exposure in the Age of Social Media

Main text

1. 1pp. 1043, 1044, 1045 and 1048, in-text citations

   Published

   (as advocated by Sears & Freeman, 1967) [...] habitual de facto selective exposure (Sears & Freeman, 1967) [...] (e.g., Sears & Freeman, 1967) [...] (Sears & Freeman, 1967, pp. 209-212)

   Correction

   (as advocated by Sears & Freedman, 1967) [...] habitual de facto selective exposure (Sears & Freedman, 1967) [...] (e.g., Sears & Freedman, 1967) [...] (Sears & Freedman, 1967, pp. 209-212)

   Basis The reference list entry is Sears, D. O., & Freedman, J. L. (1967), Public Opinion Quarterly 31(2).

2. 2p. 1051, Study 1 results

   Published

   although the MSNBC selection rate for Democrats does not show a significant difference between the partisan label–only condition (M = 0.30, SD = 0.27) and the condition that also contains social endorsement cues (M = 0.27, SD = 0.21), conditions t(234) = .92, p = .18, one-tailed, Cohen’s d = .57

   Correction

   although the MSNBC selection rate for Democrats does not show a significant difference between the partisan label–only condition (M = 0.30, SD = 0.27) and the condition that also contains social endorsement cues (M = 0.27, SD = 0.21), conditions t(234) = .92, p = .18, one-tailed, Cohen’s d = .12

   Basis The printed means and standard deviations give d = (0.30 − 0.27)/0.24 = .12, and the t statistic gives the same value (2t/√df = .12). A d of .57 is incompatible with t(234) = .92.

3. 3p. 1052, Study 1 results

   Published

   while they were 18% less likely to select such an article with weaker endorsements χ2(1, N = 187) = 15.09, p < .001, φ = .14. [...] while selecting such articles with weaker endorsements at a rate about 1% lower than chance χ2(1, N = 187) = 1.73, p = .19, φ = .14.

   Correction

   while they were 18% less likely to select such an article with weaker endorsements χ2(1, N = 187) = 15.09, p < .001, φ = .28. [...] while selecting such articles with weaker endorsements at a rate about 1% lower than chance χ2(1, N = 187) = 1.73, p = .19, φ = .10.

   Basis For a test with one degree of freedom, φ = √(χ²/N). The printed .14 equals √(3.89/187), which uses the χ² of the endorsement-only test (3.89, N = 111) in place of each test's own statistic. The endorsement-only φ = .19 on the same page is correct.

4. 4p. 1053, Study 1 results

   Published

   χ2(1, N = 299) = 23.12, p < .001, φ = .11.

   Correction

   χ2(1, N = 299) = 23.12, p < .001, φ = .28.

   Basis √(23.12/299) = .28. The printed .11 equals √(3.89/299); see the correction to p. 1052.

5. 5p. 1055, Study 2 sample

   Published

   Twelve persons were removed because they did not complete the experiment, selected nothing, or had heard about the study beforehand, leaving 141 respondents for the analysis.

   Correction

   Twelve persons were removed because they did not complete the experiment, selected nothing, or had heard about the study beforehand, leaving 141 respondents. The analyses below use 140 of them.

   Basis Table 2 reports 140 groups (participants) in every model, and the paired t test on the same page has 139 degrees of freedom, which also implies 140 participants.

6. 6p. 1057, Table 2, random effects

   Published

   σα 0.224 (0.473), 0.220 (0.469), 0.221 (0.470)

   Correction

   Corrected random-effects rows of Table 2.

   Random effects	Social cue	Partisan cue	Both
   σα (SD of participant intercepts)	0.473	0.469	0.470
   σα² (variance of participant intercepts)	0.224	0.220	0.221

   Basis Each printed σα is the square of the number printed beneath it in parentheses (0.473² = 0.224, 0.469² = 0.220, 0.470² = 0.221). The row printed the variance of the participant intercepts as σα and their standard deviation in the position of a standard error.

7. 7p. 1061, References

   Published

   Conlisk, J. (1996). Why bounded rationality? Journal of Economic Literature, 34(2), 669-700. doi:10.1126/science.151.3712.867-a

   Correction

   Conlisk, J. (1996). Why bounded rationality? Journal of Economic Literature, 34(2), 669-700.

   Basis The printed DOI belongs to a 1966 item in Science. Crossref lists no DOI for the Journal of Economic Literature article.