Assessing Old Order Amish Outpatients With the MCMI–III

Abstract

In this study, we examined Millon Clinical Multiaxial Inventory–III (MCMI–III; Millon, 2009) characteristics in an Old Order Amish outpatient sample (n = 166), with a comparison group of Old Order Amish who were not receiving mental health treatment at the time of testing (n = 80). We also graphically compared the 2 Amish groups to a non-Amish psychiatric sample in the literature. Consistent with our hypotheses, the Old Order Amish outpatients scored significantly higher than the Old Order Amish comparison group on the majority of MCMI–III scales, with mostly medium effect sizes, suggesting that the MCMI–III is a useful personality instrument in discriminating between Old Order Amish clinical and nonclinical groups. In addition, the Amish outpatients scored similar to a non-Amish psychiatric sample in the literature on most personality scales. Future MCMI–III studies with the Amish are needed to replicate and generalize our findings.

Acknowledgments

This research was funded by a private donation to Philhaven Hospital. The authors would like to thank Ken Sensenig at Mennonite Central Committee, Paul Yoder, PhD, at Oaklawn, and the Old Order Amish administrator at Green Pasture (who prefers to remain unnamed) for their help with data collection.

Notes

For the remainder of the article, we refer to “Old Order Amish” as simply “Amish” to be concise. However, to be sure, there are many different Anabaptist groups who label themselves as “Amish” (Weber, Cates, & Carey, 2010).

For a more detailed literature review of research on Amish personality and affective functioning, see Knabb, Vogt, and Newgren (2010).

The original version of this article was unavailable to the authors in that it is a master's thesis published in 1952.

Due to the challenge in collecting data from this private religious group, we were unable to utilize random sampling for our study. Instead, we collected MCMI–III data on all Amish who were willing to participate. This study design, as a result, prevented us from obtaining Amish clinical and nonclinical samples that are representative of the larger Amish community, and, thus, limits the generalizability of our findings. Overall, the MCMI–III profiles within our convenience sample might or might not be different from potential profiles of those who declined to take the test. Additional MCMI–III research is necessary that utilizes random sampling methods with the Amish.

Although propensity score analysis (see Rosenbaum & Rubin, 1983) has been used in recent years to match nonequivalent groups, it is currently unclear how large of a sample size is needed to benefit from this method of reducing sampling bias (see Luellen, Shadish, & Clark, 2005). Thus, we decided not to use propensity scoring in our statistical matching strategy.

The chi-square test results should be interpreted with caution in that the two groups were not matched on the demographic variables of age, gender, and state of residence. Rather, the percentages are provided for descriptive purposes only to examine the prevalence rates of MCMI–III personality traits in each sample.

Like the prevalence rate comparisons, the graphical comparison between the Amish clinical, Amish nonclinical, and non-Amish psychiatric groups should be interpreted with caution in that the three groups were not matched on key demographic variables, which might have altered the results. Rather, the graph is offered for descriptive purposes only to offer a preliminary comparison of Amish and non-Amish MCMI–III profiles.