Comparing groups on latent variables: a structural equation modeling approach

Structural equation modeling (SEM) provides a dependable framework for testing differences among groups on latent variables (constructs, factors). The purpose of this article is to illustrate SEM-based testing for group mean differences on latent variables. Related procedures of confirmatory factor analysis and testing for measurement invariance across compared groups are also presented in the context of rehabilitation research.     

The (non)comparability of the correlation effect size across different measurement procedures: a challenge to meta-analysis as a tool for identifying "evidence based practices"

Meta-analysis is becoming a principal tool for research synthesis and for the identification and justification of evidence based practices. A fundamental assumption in meta-analysis is that effect sizes based upon different measures are comparable. Recent work has challenged this assumption in the case of the standardized mean difference. In this article it is shown that population universe (true) score level correlation effect sizes, for the relationship between two constructs A and B, based upon different measures will be comparable only if construct validity invariance holds across the measures used to make inferences to A and the measures used to make inferences to B. The results of a simulation study are also reported which show that the results of a meta-analysis may be significantly and adversely affected by violations of construct validity invariance. Finally, it is concluded that the theoretical results obtained in this article, and the results of the simulation study, combine to suggest that the role of meta-analysis in the synthesis of social work research, and in the identification of evidence based practices, be de-emphasized until important questions about the sensitivity of meta-analysis to violations of construct validity invariance are answered.     

Differences in Relationship Quality Measurement and Covariate Influence Between Individuals in First and Second Marriages: A Propensity Score Analysis

Few studies have examined relationship quality construct measurement and covariate influence differences between first married and remarried individuals. We used multiple-sample confirmatory factor analysis and structural regression invariance testing to assess relationship quality measurement and covariate influence differences between these two groups. We account for selection bias on 11 observed selection variables by comparing outcomes before and after propensity score greedy matching procedures. Using a sample of 2,030 first married and 488 remarried individuals from the Relationship Evaluation Survey, prematching results indicate some significant differences between groups. Postmatching results, however, indicate that all measurement and structural parameters can be constrained to be equal between groups without worsening model fit. Implications, limitations, and future directions for research are discussed.     

Measurement invariance matters: A case made for the ORTOFIN

In this paper, it is argued that the measurements of an underlying, latent variable cannot straightforwardly be used in group comparisons without testing whether the measurements relate to the latent variable in the same way for all groups. The procedure for testing this is discussed and is illustrated by assessing measurement invariance across groups with a different socio-economic status (SES) for the ORTOFIN, a scale measuring an individuals’ orientation toward finances [Loix, E., Pepermans, R., Mentens, C., Goedee, M., & Jegers, M. (2005). Orientation toward finances: Development of a measurement scale. The Journal of Behavioral Finance, 6, 192–201]. The results show that one factor of the ORTOFIN, i.e. ‘‘financial information”, is invariant across SES groups while ‘‘personal financial planning” lacks measurement invariance. Implications of these results are discussed.     

Cross‐Cultural Validity of Bowen Theory in Southern European Countries

We present a comparison between Italy, Spain, Greece, and Portugal to test the cross‐cultural validity of the Differentiation of Self Inventory Revised (DSI‐R) Scales of emotional cut‐off (EC) and emotional reactivity (ER). Our study focuses on the importance of in‐depth analyses of the differentiation of self‐constructs from a cross‐cultural perspective and reflects on the possibility of conceptualising the universality of Bowen theory in southern European countries. The study involved 1,807 healthy individuals. All participants completed the DSI‐R (Skowron & Schmitt, 2003). It was found that the measurement models had a suitable fit for each country group. Nevertheless, in our study, the two DSI‐R subscales, ER and EC, highlighted a weak invariance, specifically demonstrating only the invariance of the pattern factor loadings and the magnitude of factor loadings in the four samples. The empirical and clinical implications of these data are discussed.     

No two ageisms are the same: testing measurement invariance in ageism experience across Europe

The article focuses on the topic of measurement invariance (equivalence) testing in comparative research. Measurement invariance is the basic requirement for comparative data and should be addressed by researchers before any international comparison begins. In the article, multiple-group confirmatory factor analysis is first performed to describe three methods of measurement invariance testing (configural, metric, and scalar). Subsequently, this procedure is applied to the case of measuring experience with ageism against seniors in the quantitative European Social Survey, Round 4. Measurement invariance in the concept is tested across 29 states participating in the survey. Hypotheses about the sources of the lack of measurement invariance between some countries are then tested by comparing the parameters of the models identified for different groups of European states. Finally, evidence of international comparability of items measuring experience with ageism against seniors is summarized and implications for other areas of comparative social research are discussed.     

Measuring Constructs in Family Science: How Can Item Response Theory Improve Precision and Validity?

This article provides family scientists with an understanding of contemporary measurement perspectives and the ways in which item response theory (IRT) can be used to develop measures with desired evidence of precision and validity for research uses. The article offers a nontechnical introduction to some key features of IRT, including its orientation toward locating items along an underlying dimension and toward estimating precision of measurement for persons with different levels of that same construct. It also offers a didactic example of how the approach can be used to refine conceptualization and operationalization of constructs in the family sciences, using data from the National Longitudinal Survey of Youth 1979 (n = 2,732). Three basic models are considered: (a) the Rasch and (b) two‐parameter logistic models for dichotomous items and (c) the Rating Scale Model for multicategory items. Throughout, the author highlights the potential for researchers to elevate measurement to a level on par with theorizing and testing about relationships among constructs.     

Teen Dating Violence,Sexual Harassment,and Bullying Among Middle School Youth: Examining Measurement Invariance by Gender

This study investigated measurement invariance by gender among commonly used teen dating violence (TDV), sexual harassment, and bullying measures. Data were collected from one cohort of seventh‐grade middle school students (N = 754) from four schools. Using structural equation modeling, exploratory and confirmatory factor analyses assessed measurement models and tested measurement invariance by gender for aggression measures. Analyses invoked baseline data only. Physical and psychological TDV perpetration measures achieved strict measurement invariance, while bullying perpetration demonstrated partial strict invariance. Electronic TDV and sexual harassment perpetration achieved metric/scalar invariance. Study findings lend validation to prior and future studies using these measures with similar populations. Future research should increase attention to measurement development, refinement, and testing among study measures.     

Scale invariance and similar invariance conditions for bankruptcy problems

A frequent motivation for the use of scale invariance in the bankruptcy literature is that it imposes that the outcome of a bankruptcy problem does not depend on the units of measurement. We show that this interpretation is not correct. Scale invariance is an invariance condition that applies when all amounts are multiplied by a constant (without change of units). With this interpretation in mind, it is natural to consider other invariance conditions, for example one that applies when all amounts are increased by the same constant. In this paper, we analyze the consequences of several invariance conditions. An erratum to this article can be found at     

Commitment and Sacrifice in Emerging Adult Romantic Relationships

The development of commitment and satisfaction with sacrifice has been associated with relationship health and stability. Given the importance of early romantic relationships, the purpose of the present study was to explore how these constructs relate and develop in an emerging adult population (n = 246). Through an initial confirmatory factor analysis, we found the two constructs can be distinct. Then, using interdependence theory and a cross-lagged model over three time points, we found support for a bidirectional relationship between commitment and satisfaction with sacrifice. In addition, we found a main effect for relationship length and instability (i.e., cycling) on Time 1 and Time 3 dedication respectively.     

The structure of harassment and abuse in the workplace: a factorial comparison of two measures

The structures of two measures examining negative experiences in the workplace, one focusing primarily on sexual harassment (SEQ) and one focusing on workplace abuse (GWA), were examined in detail. This article investigated whether the five subscales for the relatively unexplored measure (GWA) are reliably measured by a single underlying construct. It also investigated whether the two workplace-based measures are distinct but related constructs and the consistency of their factor structure across genders. Using a large and diverse organizational survey derived from a Midwestern university, analyses supported the distinctiveness of the two measures and showed that the factor structures for the two constructs were remarkably similar across genders. Analyses also suggested that indices of extreme behavior within each of the constructs were not reliably measured. The findings have important implications for data collection strategies in research focused on negative workplace experiences. This study provides considerable support for the continued use of both measures in research investigating the impact of adverse workplace environment on health.     

Development and validation of the Early Executive Functions Questionnaire: A carer-administered measure of Executive Functions suitable for 9- to 30-month-olds

Executive functions (EFs) enable us to control our attention and behavior in order to set and work toward goals. Strong EF skills are linked to better academic performance, and greater health, wealth, and happiness in later life. Research into EF development has been hampered by a lack of scalable measures suitable for infancy through to toddlerhood. The 31-item Early Executive Functions Questionnaire (EEFQ) complements temperament measures by targeting cognitive and regulatory capabilities. Exploratory Factor Analysis (n = 486 8- to 30-month-olds) and Confirmatory Factor Analysis (n = 317 9- to 30-month-olds) indicate Inhibitory Control, Flexibility, and Working Memory items load onto a common “Cognitive Executive Function (CEF)” factor, while Regulation items do not. The CEF factor shows strong factorial measurement invariance for sex, and partial strong factorial measurement invariance for age. CEF and Regulation scores show limited floor and ceiling effects, good internal consistency, short-term stability, and convergent validity with carer-report measures of attentional control. The EEFQ is sensitive to developmental change. Results indicate that the widely overlooked period between late infancy and early toddlerhood may be a sensitive period for EF development. The low-resource demands of the EEFQ afford the possibility to study emergent EFs at scale; opening up new opportunities in basic developmental and intervention research.     

At the intersection of criminal justice involvement and sexual orientation: Dynamic networks and health among a population-based sample of young Black men who have sex with men

Mass incarceration of young Black men has a significant impact on their network composition and stability that, in turn, may have major implications for health and well-being. A sub-group of young Black men with criminal justice involvement (CJI) identify as gay, bisexual or are non-identified men who have sex with men (hereafter MSM). This paper focuses on the potential effects of CJI on the composition of Black MSM social and sexual networks, their stability over time, and concomitant health and social outcomes. We use data from the UConnect study, a population-based cohort of young Black MSM 16–29 years of age (n = 618) selected using respondent-driven sampling in Chicago from 2013 to 2016. Both confidant and sexual network name generators and interpreters were administered at 9-month intervals over three waves of data collection. Ego and dyadic-level data were collected on behaviors prevalent among MSM and including factors associated with network CJI, network stability, and health outcomes. Generalized Structural Equation Models (GSEM) were utilized to determine the relationship between CJI network composition, network stability and behaviors prevalent among young Black MSM and their networks. In the UConnect cohort, 46% had at least once been detained, arrested or spent time in jail or prison. In addition, 20% of participants reported incident CJI over the study period. Respondents with a history of CJI were significantly more likely to have CJI homophily in their confidant and sexual networks. Multivariate analyses demonstrate that the association between one’s history of CJI, housing instability and drug use is partially explained by one’s network CJI. In addition, a higher prevalence of network CJI is associated with increased turnover in the confidant network, and this network instability is also related to important health and social outcomes. This analysis describes the networks of criminal justice involved men among a representative sample of young Black MSM and demonstrates the relationship between CJI network homophily, network stability and their impact on several key health and social outcomes relevant to this population.     

The Role of Beliefs in Sexual Behavior of Adolescents: Development and Validation of an Adolescent Sexual Expectancies Scale

This article reports on the development and psychometric properties of the Adolescent Sexual Expectancies Scale (ASEXS). Data were obtained from three annual longitudinal surveys of youth aged 10–17 at the first administration (N=932 at Wave 3). Confirmatory factor analyses indicated that 4 correlated factors corresponding to Social Risk, Social Benefit, Health Risk, and Pleasure adequately represented the expectancy items. High alpha coefficients and stability across survey waves indicated good internal and test–retest reliability for these factors. Latent variable autoregressive analyses indicated that Social Risk expectancies predicted changes in sexual behavior over the course of the study. Multigroup analyses established that the measurement and predictive models were equivalent for males and females and across age groups. The ASEXS provides researchers with a useful measure for examining adolescents' expectancy beliefs about the consequences of sexual behavior.     

Are mating strategies and mating tactics independent constructs?

This study explored the constructs of mating tactics and mating strategy. These constructs are conceptually related but distinct. In current research, the measurement of one of these constructs is often viewed as being indicative of the other. Therefore, an exploration of these constructs will enhance understanding of study outcomes in this research area. Self-report measures of mating tactics and strategies were administered to 183 female participants, aged 18-45 years. The Escalating Sexual Encounters Questionnaire (ESEQ, Greer & Buss, 1994), the Derogatis' Sexual Experience Scale (Derogatis & Melisaratos, 1979), the Sexual Strategies Measure (SSM, Schmitt, 1996), the Sociosexual Orientation Inventory (Simpson & Gangestad, 1991), and two questions assessing age at menarche and total number of sexual partners were administered. Exploratory factor analysis with oblique rotation produced two distinct factors reflecting a "tactic"-based factor and a "strategy"-based factor. This finding is consistent with viewing mating tactics and mating strategies as distinct and varying independently. An important implication of this study is that measurement of mating tactics is not indicative of underlying mating strategies in women. Further, four patterns of female mating style emerged upon review of participant factor scores and are discussed within an evolutionary context.     

Dyadic Supervision Evaluation: An Actor–Partner Relational Model

This article contributes to research practices in marital and family therapy, specifically the dyadic and development over time in clinical supervision, and describes and applies methodological strategies to develop measurements congruent with the systemic and developmental principles of the field. This project evaluates the psychometric properties of the dyadic supervision evaluation (DSE) in terms of measurement equivalence and causality. A structural equation analysis is conducted utilizing the actor–partner interdependent model resulting in a goodness of fit. This study presents a longitudinal model for evaluating the supervisory relationship and proposes a model of clinical supervision evaluation. The relationship among latent constructs in the DSE and its limitations related to inference and application are discussed.     

Measuring organizational effectiveness to develop strategies to promote retention in public child welfare

Public child welfare agencies are under pressure to improve organizational, practice and client outcomes. Related to all of these outcomes is the retention of staff. Employee intent to remain employed may be used as a proxy for actual retention. In this study public child welfare staff in one Midwestern state were surveyed using the Survey of Organizational Excellence (Lauderdale, 1999) and the Intent to Remain Employed (Ellett, Ellett, & Rugutt, 2003) scales to assess the extent to which constructs such as perceptions of organizational culture, communication and other areas of organizational effectiveness were associated with intent to remain employed. A number of statistically significant relationships were identified which were presented to the public agency for use in the development of strategies for organizational improvement. Data were also analyzed regionally and based on urban/suburban/rural status to enable development of targeted approaches. This case study presents an example of how ongoing measurement of organizational effectiveness can be used as a strategy for organizational improvement over time in the child welfare system.     

Measuring social networks cross-culturally

This paper explores the reliability and validity of a cross-cultural survey approach to network measurement using data from two ethnic groups in rural West Africa. Over a 2-week test-retest period, no significant differences in network size were detected. Network composition was less stable over the same period, particularly among peripheral versus core network members. Comparisons with support networks identified through lifecycle interviews and 24 h recall support the construct validity of the survey approach in both ethnic groups.     

Assessing the Validity and Reliability of the Perceptions of the Consent to Sex Scale

Although sexual assault prevention education tends to focus on consent promotion as a means to reduce rates of sexual assault, little is known about how college students consent to sexual activity. The current study aimed to better understand college students' consent via the systematic development of the Consent to Sex Scale (CSS), utilizing mixed methods via three phases and two waves of data collection. In Phase 1, qualitative data were collected from college students (n = 185) to provide a foundation for item writing. In Phase 2, closed-ended items were written for a quantitative instrument and reviewed by a team of experts. In Phase 3, a quantitative survey, including items written in Phase 2, was administered to college students (n = 685); the measure was assessed for its psychometric properties. Exploratory factor analysis was utilized, resulting in a five-factor solution. The CSS and corresponding factors demonstrated high internal consistency reliability and expected gender differences, supporting the construct validity of the measure. The CSS assesses college students' cues for indicating consent to sex, a construct not addressed by previous measures. The validated scale may be useful in future research to better understand how consent relates to other behaviors or constructs.