Can A High Sense of Control and John Henryism be Bad for Mental Health?

Abstract

Social stratification profoundly affects mental health. Specifically, substantial empirical evidence finds that higher status promotes mental health via a higher sense of control and a propensity to cope actively with problems. An unresolved issue, though, is whether the effects of sense of control and active coping on mental health are uniformly beneficial across levels of socioeconomic status. Perceived control and John Henryism, an active coping style, may undermine mental health, especially for lower-status persons, who lack resources. Using data from the National Comorbidity Survey for African Americans, Hispanics, and whites, we find that both sense of control and John Henryism tend to be monotonically related to positive mental health regardless of socioeconomic status and race/ethnicity.

ACKNOWLEDGMENTS

We thank Anastasia S. Vogt Yuan and Blair Wheaton for helpful comments and suggestions.

NOTES

Notes

¹ Coping styles differ from coping strategies, which are behavioral and/or cognitive attempts to manage a stressor that is specific to a role or an event. Problem-focused coping strategies are aimed at the stressors themselves, whereas emotion-focused strategies attempt to deal with the emotions that stressors evoke (Menaghan 1983; Lazarus and Folkman 1984; Thoits 1995). Coping styles, too, can be more or less problem focused or emotion focused.

² We excluded two other items, “blame yourself for things” and “feel hopeless about the future,” because they could be confounded with the sense of control.

³ Mirowsky and Ross (1991:142–43) suggested creating a sense of control scale by subtracting summed scores on the powerful others and chance items from summed scores on the internal control items. That is, sense of control = (I − P − F)/k_I + k_P + k_F), where k is the number of items measuring each dimension. Substantively, the higher internal control is, relative to perceived control by chance and powerful others, the higher is sense of control. This measure is identical to the one used here when it is scaled to range from 0 to 3.

Mirowsky and Ross also suggested creating two related measures. First, an indicator of agreement bias summed scores on all the items in the internal control, change, and powerful others scales: Agreement = (I + P + F)/k_I + k_P + k_F). We did not include this variable in the final models, as its high correlation with sense of control (r = −.55) created some collinearity problems in the John Henryism regressions (Table 3). It had significant effects only on John Henryism and substance use disorder, to which it was positively related. Its inclusion did not change the effects of the other predictors on mental health outcomes. Second, a measure of the tendency to attribute outcomes to powerful others rather than chance or fate subtracted scores on the chance scale from scores on the powerful others scale. Although this measure was tangential to our research questions, we included it in initial regression analyses of mental health outcomes (not shown). Its effects never approached significance, and the results were nearly identical to those reported.

⁴ Three other items in the National Comorbidity Survey (NCS) that were designed to measure John Henryism were excluded because they factored on a different dimension. These items were “Sometimes I feel that if anything is to be done right, I have to do it myself,” “It is important for me to be able to do things the way I want to do them rather than the way other people want me to do them,” and “I like doing things that other people thought could not be done.” Ford et al. (2000) also excluded the first two items from their scale, because the items had very low item-to-scale correlations for both African Americans and whites.

⁵ The items loaded on four factors corresponding to the measures of John Henryism, internal control, chance, and powerful others. Because of the theoretical linkage between John Henryism and internal control and the presence of an item in each scale that mentions hard work, it was not surprising that John Henryism and sense of control were correlated (r = .34). Collinearity was not a problem, however, and the two variables had strong, independent effects.

⁶ In analyses not shown, we added the square of John Henryism as a predictor of the three mental health outcomes. It was never significant, indicating that the protective effects of John Henryism increased linearly.

⁷ Mirowsky and Ross (1990) used age, education, income, and race/ethnicity as indicators of status. We included gender as a status as well, although it did not significantly predict sense of control.

⁸ Predicted distress would be at a minimum at those points. It would for 100.283 for hypothetical low-status respondents and 91.326 for hypothetical high-status respondents.