Article

Missing Not at Random Models for Latent Growth Curve Analyses

Department of Psychology, Arizona State University,Tempe, AZ 85287–1104, USA.
Psychological Methods (Impact Factor: 4.45). 03/2011; 16(1):1-16. DOI: 10.1037/a0022640
Source: PubMed

ABSTRACT

The past decade has seen a noticeable shift in missing data handling techniques that assume a missing at random (MAR) mechanism, where the propensity for missing data on an outcome is related to other analysis variables. Although MAR is often reasonable, there are situations where this assumption is unlikely to hold, leading to biased parameter estimates. One such example is a longitudinal study of substance use where participants with the highest frequency of use also have the highest likelihood of attrition, even after controlling for other correlates of missingness. There is a large body of literature on missing not at random (MNAR) analysis models for longitudinal data, particularly in the field of biostatistics. Because these methods allow for a relationship between the outcome variable and the propensity for missing data, they require a weaker assumption about the missing data mechanism. This article describes 2 classic MNAR modeling approaches for longitudinal data: the selection model and the pattern mixture model. To date, these models have been slow to migrate to the social sciences, in part because they required complicated custom computer programs. These models are now quite easy to estimate in popular structural equation modeling programs, particularly Mplus. The purpose of this article is to describe these MNAR modeling frameworks and to illustrate their application on a real data set. Despite their potential advantages, MNAR-based analyses are not without problems and also rely on untestable assumptions. This article offers practical advice for implementing and choosing among different longitudinal models.

  • Source
    • "The required assumption is that missing data are Bmissing at random,^ which means that there is no discernible pattern to the missingness once measured variables are controlled. If this assumption cannot be met (as is often the case), sensitivity tests should be conducted to probe the likely impact that missing data might have on the estimates of the intervention effect (Enders 2011; Imai 2009; Muthen et al. 2011). "
    [Show abstract] [Hide abstract]
    ABSTRACT: A decade ago, the Society of Prevention Research (SPR) endorsed a set of standards for evidence related to research on prevention interventions. These standards (Flay et al., Prevention Science 6:151-175, 2005) were intended in part to increase consistency in reviews of prevention research that often generated disparate lists of effective interventions due to the application of different standards for what was considered to be necessary to demonstrate effectiveness. In 2013, SPR's Board of Directors decided that the field has progressed sufficiently to warrant a review and, if necessary, publication of "the next generation" of standards of evidence. The Board convened a committee to review and update the standards. This article reports on the results of this committee's deliberations, summarizing changes made to the earlier standards and explaining the rationale for each change. The SPR Board of Directors endorses "The Standards of Evidence for Efficacy, Effectiveness, and Scale-up Research in Prevention Science: Next Generation."
    Full-text · Article · Apr 2015 · Prevention Science
  • Source
    • "All these models suggested a 3-class solution; characteristics of subgroups were comparable to those of the MAR-solution; at most, the deviation with regard to the assignment of individuals to a specific class was N = 2. We did not apply selection modeling as our data did not meet the central assumptions of these models due to the discrete measurement of cultural identification in our study (see Enders, 2011). Table 1 "
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent research suggests that cultural groups comprise subgroups differing in the combination of their self-identifications with their heritage and host cultures and following distinct trajectories of acculturation. This study aimed at identifying such subgroups, predicting group membership by pre-migration factors, and testing for acculturation-related experiences with the host culture over time. The sample comprised 366 adolescent diaspora migrants (59% female, 16 years old) from the former Soviet Union to Germany. A person-oriented and longitudinal approach using growth mixture modeling revealed three subtypes of cultural identification change. The first subgroup (Idealists) comprised adolescents with high and stable identification with their host culture and low but increasing identification with their heritage culture. The second group (Skeptics) showed low and stable identification with their host culture and high but decreasing identification with their heritage culture. The third group (Realists) reported medium-level and stable identification with both host and heritage cultures. Group comparisons showed premigration differences: Idealists and Realists most likely spoke the host-culture language (i.e., German) as a native language and Idealists reported the highest level of maternal education. Membership in subgroups also related to adolescents’ perception of their current situation in the host culture: Idealists reported less acculturation-related hassles with respect to language and socio-cultural adaptation. Skeptics perceived lower expectations by natives to get in contact with natives and higher expectations to spend time with members of their immigrant group than the average immigrant. Results highlight the heterogeneity in immigrant populations and their specific needs before and after the actual migration.
    Full-text · Article · May 2014 · Journal of Cross-Cultural Psychology
  • Source
    • "Data analysis was performed in four steps. First, the integrity of the data was inspected through the evaluation of data distributions and scatterplots and by conducting sensitivity analyses using pattern mixture models (Enders, 2011) to determine the degree to which missing data impacted observed results. Second, two exploratory factor analyses (EFA) were performed. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The therapeutic model underlying Acceptance and Commitment Therapy (ACT) is reasonably well-established as it applies to chronic pain. Several studies have examined measures of single ACT processes, or subsets of processes, and have almost uniformly indicated reliable relations with patient functioning. To date, however, no study has performed a comprehensive examination of the entire ACT model, including all of its component processes, as it relates to functioning. The present study performed this examination in 274 individuals with chronic pain presenting for an assessment appointment. Participants completed a battery of self-report questionnaires, assessing multiple aspects of the ACT model, as well as pain intensity, disability, and emotional distress. Initial exploratory factor analyses examined measures of the ACT model and measures of patient functioning separately with each analysis identifying three factors. Next, the fit of a model including ACT processes on the one hand and patient functioning on the other was examined using Structural Equation Modeling. Overall model fit was acceptable and indicated moderate correlations among the ACT processes themselves, as well as significant relations with pain intensity, emotional distress, and disability. These analyses build on the existing literature by providing, to our knowledge, the most comprehensive evaluation of the ACT theoretical model in chronic pain to date.
    Full-text · Article · May 2014 · Behavior therapy
Show more