Meaning in Life Questionnaire: Validation for Chinese Adolescents

Abstract

Background:

Previous psychometric analyses of the Meaning in Life Questionnaire (MLQ; Steger et al. 2006) in the Chinese adolescent population are limited methodologically.

Methods:

This study validates the MLQ in a sample of 1794 adolescents (62.4% girls; mean age = 16.6 years) in 24 high schools in Yunnan, China, a representative sample in terms of provincial socioeconomic status.

Results:

Confirmatory factor analysis validates the two-factor structure of the MLQ, and several indices of scale reliability confirm its reliability. However, homogeneity and response patterns at the item level highlight issues with the last two items of the scale, suggesting that potential latent dimensions such as subsamples should be considered for those items.

Conclusions:

It is concluded that the MLQ has adequate psychometric performance as a measurement instrument for meaning in life in Chinese adolescents, paving the way for its future application in research and clinical practice.

Introduction

Meaning in life plays an important role in mental well-being. Its importance came to the fore during the coronavirus disease 2019 (COVID-19) pandemic when people faced uncertainty and change that challenged their understandings of meaningfulness, and meaning in life continues to be relevant for navigating the post-pandemic era and its discontents.^1–3 Interventions that enhance perceived meaning in life in individuals have been implemented in the wake of negative events such as earthquakes ⁴ and following the pandemic. ⁵

The Meaning in Life Questionnaire (MLQ ⁶ ) has been widely used as an instrument for measuring the subjective sense or experience that life is meaningful, although its utility as a measurement instrument for Chinese adolescents is not well established. Our study validates the MLQ in a representative sample of high school students including ethnic minority students in China’s Yunnan Province, using a stratified approach which accounts for social-economic factors within different regions. Previous psychometric analyses of the MLQ in Chinese adolescent populations have thus far been limited methodologically, and there are discrepant findings in previous studies. For example, an early study ⁷ based on college students in Beijing suggested that a 9-item scale was better than alternative versions, and yet subsequent studies of high school students^4,8 have used the original 10-item scale to retain integrity. Previous work is also limited by its inadequate reporting of samples, sampling, ⁹ and sole focus on Han Chinese. Although it is possible that meaning of life could vary in relation to ethnicity, the scale’s utility for surveying ethnic minority adolescents in China remains unclear. ⁴ We can find only one study ¹⁰ that included ethnic minorities in their sample, but the percentage of the total sample was small (29; 1.5%), even though ethnicity plays an important role in positive psychology. ¹¹ Previous validation studies are also limited to classical test theory (CTT), ⁹ and statistical tests such as factor analysis and conventional reliability tests,^7,8 even though these approaches have been considered insufficient for scale validation, ¹² leaving the item-level properties of the MLQ open to question. It is clear, then, that previous studies of the MLQ in China have only cursorily examined its psychometric properties, and further work is needed to validate the scale for adolescent respondents which is the aim of the study reported in this article.

Method

The participants were residents in six regions in Yunnan Province. The study used a two-level stratified sampling approach to ensure a representative sample in terms of provincial socioeconomic status. Gross domestic product (GDP) per capita was first used to cluster regions following the ranking of gross regional product by region in the census; the GDP per capita values of 16 regions were analyzed using k-mean cluster analysis. As a result, three clusters of regions based on GDP per capita were identified (see Supplementary Materials for the regions). To best match the ratios of cluster sizes, six regions were selected for each cluster, namely, Kunming (17%), Honghe (17%), Dali, Zhaotong, Pu’er, and Xishuangbanna (67%). Kunming was selected because it is the capital of Yunnan Province whereas Yuxi was excluded, even though it was a cluster 1 region, as its economy is unusually more reliant on the tobacco industry. Candidate regions in the other two clusters were randomly chosen.

At the school level, we used convenience sampling whereby administrators in schools in the selected regions were approached for permission to conduct the survey, although three declined due to scheduling issues. The administrators were issued with a QR code linked to the online survey used in the study. The form teachers in each school then sent the QR code to students in their class group on WeChat (the inclusion criterion was being a registered tenth grader student in the school). Students first read the instructions of the online survey and then answered its questions using their mobile phones after clicking on a consent agreement button. Participation was voluntary and data collection was anonymous and mostly happened during the Chinese New Year holiday time when the students were not in school, that is, not monitored by a teacher who might have influenced their self-reporting. The data collection took place in 2024.

As a result, 1794 adolescents in 24 secondary schools (tenth graders; 62.4% girls; M_age = 16.6 years; R_age = 15–19 years) completed the MLQ. Nearly half (42.6%) of those surveyed are Han Chinese, the numerically largest ethnic group in China and the world. All questions in the survey with quantitative responses were compulsory; thus, there are no missing values in the dataset. The psychometric analysis involved four stages following comprehensive guidelines and criteria, ¹² as detailed below. Following comprehensive validation statistics, a descriptive analysis between subscales was also conducted. Data were managed in IBM SPSS. All analyses were conducted in R.

Results

Factor Analysis

Results from confirmatory factor analysis confirm the two-factor structure of the MLQ previously reported ⁶ since its two subscales, the presence of meaning in life and the search for meaning in life respectively, each represent one factor or dimension. The model had an adequate-good fit ¹³ , RMSEA = 0.10, CFI/TLI = 0.96/0.94, and SRMR = 0.07, which outperformed the one-factor structure model, RMSEA = 0.22, CFI/TLI = 0.79/0.72, SRMR = 0.10.

Psychometric Analysis

Using several indices of scale reliability, the MLQ and its subscales were tested using the CTT including Cronbach’s alpha (α), Guttman’s lambda 6 (G6), and McDonald’s omega (ω). Both the presence (α = 0.80, G6 = 0.84, ω = 0.85) and search (α = 0.90, G6 = 0.91, ω = 0.92) subscales are reliable. Given the 7-point ordinal scale nature of all items, polychoric correlations were used to calculate an α index for both subscales, resulting in robust outcomes with the search subscale being more reliable than the presence subscale.

Homogeneity of Scale

Non-parametric item response theory techniques were used to evaluate homogeneity and response patterns at the item level. While the entire 10-item questionnaire can be regarded as homogeneous (H = 0.48), the ninth item (i.e., “My life has no clear purpose”) shows a poor fit in the scale. When the ninth item is removed, the unidimensional nature of the remaining items can be assumed following the automatic item selection algorithm. ¹² When item response functions were examined using the MIIO method, which checks the manifest invariant item ordering, the tenth item (i.e., “I am searching for meaning in my life”) was identified as a deviant item (crit. = 192 > 80 ¹² ). A test of the final eight items found no idiosyncratic response patterns across the items.

Subscale Descriptives

Finally, descriptive analyses on the total scores were conducted. Two total scores were created by summing up all item scores. As shown in Figure 1, the highly correlated (Pearson’s r = 0.57) composite scores showed a wide range and acceptable distribution: present sum score = 21.94/5.81 (M/SD), range = 5–35 and search sum score = 23.97/6.40 (M/SD), range = 5–35. Using 15% as the cutoff point, ¹² these sum scores showed no ceiling or floor effects.

Figure 1.

Note: The 1 and 2 sigma correlation ellipses are also displayed on the scatterplot. Density lines are plotted over the histograms.

Ethics Approval

The study was given ethics approval by the Academic Committee of the School of Foreign Languages and Literature at Yunnan Normal University dated September 14, 2023. Written informed consent was taken from all the participants. School administrators gave permission to conduct research in their schools. The study was carried out in accordance with the principles enunciated in the Declaration of Helsinki.

Discussion

Building on previous validation studies, the thorough psychometric evaluation of the MLQ in this study demonstrates adequate psychometric performance for assessing meaning in life in Chinese adolescents including ethnic minority adolescents. Our focus on the item-level properties of the MLQ in contrast to previous studies of conventional statistical tests^7,8 offers a more sophisticated scale validation. This finding paves the way for future applications of the measurement instrument in psychological and social science research among Chinese-speaking people and in a wide range of topics to explore their associations with meaning in life. The validated MLQ is also useful for mental health professionals working in child and adolescent mental health ¹⁴ including interventions to enhance perceived meaning in life in individuals.^4,5 This is particularly important in China where students experience excessive stress in their daily lives, especially from high expectations for academic achievement, ¹⁵ and where there is a demonstrated need for support in handling stressful situations and in personal growth and development.

The thorough sampling strategy and validation of the MLQ in this study overcome incomplete information and reporting in previous studies about samples, sampling, and statistics^4,7–9 which have made it difficult to ascertain whether the MLQ is appropriate for Chinese adolescents. However, although the strong scalability and reliability of the MLQ’s subscales in the present study suggest that its application can lead to reliable and meaningful research findings and better-informed decisions, the bifactorial structure indicates nuances in the questionnaire items proposed previously. ⁶ The sum scores of both subscales demonstrate diversity, yet our item response investigation highlights issues with the last two items which challenge the homogeneity or invariant ordering patterns of the scale. This finding is extremely salient for the sample in this study that consists only of tenth-grade students in the same province. As indicated, ¹² potential latent dimensions such as subsamples may need to be considered for those items in future studies. Although our study has strengths including its representative sample and validation process, the restricted age and grade of the sample are its main limitation as the developmental stage of the adolescents might have limited their comprehension of some items in the questionnaire. Future research should include adolescents of different ages and different ethnic minorities.

Supplemental Material

Supplemental material for this article is available online.