Gallery Presentation International Positive Psychology Association 7th IPPA World Congress 2021

(Re)examining the dimensionality of the Multidimensional Existential Meaning Scale: Using the bifactor model and ancillary bifactor measures (#258)

Jonah Li 1
  1. Indiana University-Bloomington, Bloomington, INDIANA, United States

Background

Growing research interests in meaning in life (MIL) have catalyzed the development of Multidimensional Existential Meaning Scale (MEMS)1, reflecting the recent tripartite MIL conceptualization2-4 (comprehension, purpose, mattering). Factor analytic results using student samples supported MEMS’s correlated three-factor structure1. Despite this advancement, the correlated three-factor model cannot distinguish variance shared across all items from variance specific to factors, which should be assessed in multidimensional constructs5. Bifactor modeling could conduct this assessment5. Also, ancillary bifactor measures can assist interpreting dimensionality6.

Hypotheses

This study sought to (re)examine MEMS’s dimensionality via comparing three competing models (unidimensional, correlated factors, bifactor) and examining ancillary bifactor measures. Given previous factor analytic results1, the correlated factors model was hypothesized to show adequate fit and be supported by ancillary bifactor measures.

Sample

Sample 1 included 643 community adults (Mage = 32.04; SD = 9.36) recruited from Amazon Turk(MTurk). Sample 2 comprised 346 patients with chronic illnesses (e.g., chronic pain, mood disorders) (Mage = 36.02, SD = 12.68) recruited from MTurk. Sample 3 included 956 college students (Mage = 19.94, SD = 2.78) recruited from a university. 

Design

CFA tested MEMS’s dimensionality with each sample. Competing measurement models were examined using maximum likelihood estimation with robust standard errors (MLR). Model fit was evaluated based on existing fit indices’ guidelines7. The unidimensional model was nested within the three-factor correlated model, which was nested within the bifactor model. Scaled chi square difference test, AIC, BIC were used for model comparison8-9.

Results

Across three samples, the correlated factors and bifactor solution demonstrated good fit, while the unidimensional model demonstrated poor fit. Model fit comparisons suggested that the bifactor solution fit the best. Ancillary bifactor measures collectively suggested conceptualizing MEMS as unidimensional.

Contribution

By clarifying MEMS’s dimensionality, the findings challenge the recent tripartite MIL model and caution against using its subscales’ scores. 

  • Keywords: Meaning and Purpose