Erratum Re: “Coping With Racism: Moderators of the Discrimination—Distress Link Among Mexican-Origin Adolescents”

Erratum:

After the publication of Park, Wang, Williams, & Alegría (2018), we discovered that one of the families in the sample violated the inclusion criteria. Specifically, one inclusion criterion specified that adolescents resided with biological mothers of Mexican descent; however, there was one stepmother (not a biological mother) of Mexican descent. Thus, this one family with the stepmother was dropped from the sample, and we re-ran the analyses. Results are shown in Table 1.

Table 1.

Results from Multilevel Moderation Analyses by Mental Health Outcome

Moderator	${\mathit{\beta }}_{40}\mathit{ }$Within-person	${\mathit{\beta }}_{21}\mathit{ }$Cross-level	${\mathit{\beta }}_{03}\mathit{ }$Between-person
Externalizing problems
Anger-Out	3.70 (.017)*	.16 (.007)	.03 (.619)
Anger Suppression	1.55 (.162)	−.11 (.140)	−.25 (.001)
Ethnic Identity Commitment	−14.19 (.020)	−.50 (.045)	−.54 (.051)
Ethnic Identity Exploration	−7.62 (.032)	−.61 (.019)	−.23 (.473)
Social Support
Family	−8.86 (.039)	−.38 (.005)	−.05 (.754)
Friends	−1.95 (.559)	−.34 (.022)	.18 (.223)
Sig Other	−.97 (.705)	−.54 (.001)	−.02 (.911)
Internalizing problems
Anger-Out	2.97 (.051)	.16 (.007)	.02 (.728)
Anger Suppression	1.08 (.477)	.11 (.262)	.13 (.180)
Ethnic Identity Commitment	3.41 (.671)	−.35 (.300)	−.26 (.447)
Ethnic Identity Exploration	5.34 (.243)	−.37 (.298)	.60 (.145)
Social Support
Family	1.86 (.590)	−.41 (.018)	−.20 (.248)
Friends	2.16 (.483)	−.18 (.331)	−.11 (.569)
Sig Other	2.65 (.434)	−.69 (.001)	.02 (.929)

$Note: {\mathrm{\beta }}_{40}$ represents the average within-person level moderation effect, quantifying how the time-varying moderator moderates the within-person relation between perceived discrimination and the outcome; ${\mathrm{\beta }}_{21}$ is a cross-level moderation effect, quantifying how person levels of the moderator moderate the within-person relation between perceived discrimination and the outcome; and ${\mathrm{\beta }}_{03}$ measures the between-person level moderation effect, quantifying how person levels of the moderator moderate the between-person relation between perceived discrimination and the outcome. Anger-Out = Outward Anger Expression; Sig Other = Significant Other. The values outside the parentheses are the point estimates, and the values inside the parentheses are p values; p values <.05 have been bolded. The missing data handling method was full information likelihood estimation.

^*This result was still statistically significant (p =.029) when a truncated version of the Externalizing Problems measure was used.

The overall pattern of results remained the same. That is, all significant results remained significant, and all non-significant results remained non-significant in Table 1.

However, one minor exception was the result associated with the footnote under Table 1. Originally, the footnote had pointed out that using a truncated version of the Externalizing Problems measure led to a trend-level result (p =.06) for outward anger expression as a moderator. According to the re-analysis, the truncated version of the Externalizing Problems measure still led to a statistically significant result (p =.029) for outward anger expression as a moderator. This corroborates the pattern of findings in Table 1 and underscores the significance of the results.

These results would not change our Discussion section, apart from making the cautionary note unnecessary regarding the within-level exacerbating effect of outward anger expression on the discrimination—externalizing problems association. In other words, this re-analysis only strengthens the original results.