Latent Mean Differences Research Articles

Secondary Students’ Perceptions of Teaching Quality and Teacher-Family Interaction: A Structural Equation Model

AbstractStudents’ perceptions of teaching quality and teachers’ interaction with their families, along with student self-efficacy and interest in a subject matter to performance. However, little is known about how student perceptions of teaching quality and teacher-family interactions in the subject of English relate to performance in reading comprehension. The current study explores how the teacher-related perceptions of New Zealand senior secondary students relate to their reading attitudes and tested reading comprehension. Partially mediated structural equation modelling revealed that students’ perceptions of teaching quality directly and positively influenced student self-efficacy, interest, and performance and indirectly through self-efficacy. However, perceptions of teacher-family interactions negatively influenced student self-efficacy and performance and were not related to interest. Although the model was statistically invariant across demographic variables, there were large latent mean differences across school decile bands for reading performance. These findings reinforce the importance of student perceptions of good teaching and raise challenges about how teachers interact with families.

Assessing mental health professionals’ attitudes toward routine outcome monitoring across Eastern and Western cultures: an examination of invariance and latent mean differences†

Objective To investigate the attitudes of mental health professionals towards Routine Outcome Monitoring (ROM) in Eastern and Western cultures. Method Two samples of American (N = 455) and Chinese (N = 505) mental health professionals completed the prevalent Monitoring and Feedback Attitudes Scale (MFA). We tested the measure’s psychometric characteristics, measurement invariance, and latent mean difference across cultures. Results Confirmatory factor analysis (CFA) found a two-factor structure of the MFA within both cohorts. The MFA subscales showed excellent internal consistency in both the Chinese and American samples. The MFA demonstrated partial scalar invariance between the two cultural groups, supporting the comparison of latent means among Chinese and American professionals. Chinese professionals perceived greater harm from ROM than their American counterparts. Conclusions The findings show that the MFA is a valid tool to evaluate and compare the US and Chinese mental health professionals’ attitudes toward ROM, suggesting that the measure may be useful in both Eastern and Western cultures.

연령대별 디지털 리터러시 척도의 구인동등성 검증 및 잠재평균 분석

This study tested the equivalence of a digital literacy measurement tool across different age groups and examined differences in latent means. It measured digital literacy using the same items among 994 elementary students, 348 middle school students, and 220 university students representing childhood, adolescent, and early-adulthood groups, respectively. The digital literacy scale originally developed for the adolescent group was modified to accommodate the vocabulary and contexts appropriate for childhood and early-adulthood groups. The results confirmed complete configural and metric invariance across the groups, though scalar invariance was partially achieved. After partial scalar invariance was established, partial strict invariance was secured for comparing latent means. The comparison revealed that the childhood group scored higher on data and content literacy than the adolescents, but lower on community literacy. The early-adulthood group scored significantly higher than the adolescent group on data and media literacy. This paper interprets of these results, discussing the implications and limitations of the study.

연령대별 디지털 리터러시 척도의 구인동등성 검증 및 잠재평균 분석

This study tested the equivalence of a digital literacy measurement tool across different age groups and examined differences in latent means. It measured digital literacy using the same items among 994 elementary students, 348 middle school students, and 220 university students representing childhood, adolescent, and early-adulthood groups, respectively. The digital literacy scale originally developed for the adolescent group was modified to accommodate the vocabulary and contexts appropriate for childhood and early-adulthood groups. The results confirmed complete configural and metric invariance across the groups, though scalar invariance was partially achieved. After partial scalar invariance was established, partial strict invariance was secured for comparing latent means. The comparison revealed that the childhood group scored higher on data and content literacy than the adolescents, but lower on community literacy. The early-adulthood group scored significantly higher than the adolescent group on data and media literacy. This paper interprets of these results, discussing the implications and limitations of the study.

Measurement invariance of the Suicide Cognitions Scale-Revised (SCS-R).

The Suicide Cognitions Scale-Revised (SCS-R) is a unidimensional measure of suicidal cognitions theorized to assess the suicide belief system. Several solutions have been proposed for the Suicide Cognitions Scale and SCS-R (e.g., bifactor model with two specific factors, bifactor model with two specific factors, three correlated factors model). Research indicates the endorsement of thoughts of suicide and suicide-related cognitions varies across demographics. Thus, the current investigation tested the measurement invariance (MI) of the SCS-R across gender, race, and sexual orientation within these proposed solutions and a unidimensional model. A national sample of N = 10,625 adults completed an online survey that included the SCS-R and self-report measures of demographics. Results indicated that the bifactor model with three specific factors, the bifactor model with two specific factors, and the three correlated factors models achieved scalar invariance across gender, race, and sexual orientation; a unidimensional model was not scalar invariant by gender. Tests of latent mean differences revealed significant differences in the general factor (i.e., suicidal belief system) and the specific unlovability, unbearability, and unsolvability factors between a few demographic groups. Implications for theory, measurement, and modeling are discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Examining measurement invariance of a managerial coaching scale across traditional and virtual employee groups: implications for expanding managerial coaching research

ABSTRACT Managers and human resource development (HRD) professionals are increasingly interacting with employees who work remotely and on-site at the office. These workers also span across multiple generations, each of which is unique in their needs for growth, learning, development, performance improvement, and fulfiling their career aspirations. Managerial coaching is a well-established approach that can develop, support, and improve work-related outcomes for employees. Yet, studies focusing on managerial coaching in virtual contexts are limited. To advance research on managerial coaching in virtual contexts, it is important to use rigorously tested and psychometrically verified scales. Therefore, the purpose of this study was to test measurement invariance and assess latent mean differences between virtual and traditional groups with data collected with one of the most widely used managerial coaching instruments. Data were obtained via MTurk in this cross-sectional, multi-survey design. Methodologically this study is unique in HRD because groups were equated with Propensity Score Matching before data analysis. Results indicate full scalar invariance between groups. Furthermore, traditional employees perceived that their managers exhibited more coaching behaviours than those perceived by virtual employees. Findings from this study provide support for the psychometric properties of this scale for these groups. Implications for research and practice and future research directions are discussed.

Testing the Measurement Invariance and Latent Mean Difference of the Creative Trait Motivation Scale for Dance Students

Testing the Measurement Invariance and Latent Mean Difference of the Creative Trait Motivation Scale for Dance Students

Measurement of the self-determination continuum of motivation for engaging in alcohol-related harm reduction behaviors: Improved content coverage of the Treatment Self-Regulation Questionnaire.

Self-determination theory (SDT) proposes that people are more likely to engage in behaviors that reduce the harms associated with alcohol use if they do so for more self-determined reasons. There is growing support for this proposal, but the Treatment Self-Regulation Questionnaire (TSRQ), which assesses the self-determination continuum of motivation for engaging in alcohol-related harm reduction behaviors, lacks content coverage. We generated additional items to improve the content coverage of the TSRQ and evaluated its psychometric properties. We also compared two randomly assigned instruction sets that referenced "responsible drinking" or "protective behavioral strategies" (PBS). Participants (n = 2,236) were college students from psychology departments at 10 universities in eight U.S. states who reported past-month alcohol use and completed the revised TSRQ; the online survey was completed for partial course credit. Exploratory factor analysis supported a three-factor structure representing autonomous motivation, controlled motivation, and amotivation for the PBS version. This factor structure was confirmed using exploratory structural equation modeling for both the PBS and responsible drinking versions. Scalar invariance was achieved across instruction sets. Latent mean differences showed that those who received the PBS version had lower autonomous and controlled motivation scores, but higher amotivation scores. Associations of the three TSRQ factors with alcohol-related outcomes were consistent with SDT, and the magnitude of these associations did not differ across instruction sets. More accurate assessment of the self-determination continuum of motivation for alcohol-related harm reduction behaviors will improve research on this topic which has promising implications for alcohol interventions. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

BECK ANXIETY INVENTORY: MEASUREMENT INVARIANCE, LATENT MEAN COMPARISON, AND RELIABILITY IN ADULTS FROM BUENOS AIRES

This research aimed to provide new psychometric evidence of the Argentine Beck Anxiety Inventory (BAI) version in a sample of adults from Argentina. More specifically, the objectives were: 1) test the factor structure of the BAI; 2) assess gender, education, age, and region invariance; 3) test the latent factor mean difference across gender, education, age, and region; 4) analyze the reliability of the BAI; 5) provide population-based norms. The sample was composed of 1,410 adults ranging between 18 and 65 years (M = 32.41; SD = 10.45). Results showed that the BAI should be interpreted as unidimensional. Full configural, metric, and scalar invariance across gender, education, age, and region were obtained for the unidimensional model. The BAI presented adequate reliability values. Latent mean comparisons showed that men experienced more anxiety than women, that college-educated experienced more anxiety than non-college, that older people experienced more anxiety than younger people, and that people living in Greater Buenos Aires experienced more anxiety than those from Buenos Aires City. The BAI is an invariant measure of anxiety symptoms with good psychometric properties.

Measurement Invariance of the Children's Social Desirability Scale-Short Version (CSD-S) Across Gender, Grade Level, and Race/Ethnicity.

Social desirability bias is a methodological challenge with participant self-reports. Social desirability measures are often used to control the potential effects of social desirability bias on participant self-reports, but whether these measures are interpreted similarly across different demographic groups remains unclear. This study examines the factor structure and measurement invariance of the Children's Social Desirability Scale-Short version (CSD-S) across gender, grade level, and race/ethnicity with a large sample of children and adolescents in the United States. Results indicate a close fit to a one-factor model. Tests of measurement invariance show partial strong invariance across gender (male vs. female) and grade level (elementary vs. middle vs. high schoolers) as well as strong invariance across race/ethnicity (Asian vs. Black/African American vs. Hispanic/Latine vs. White vs. Bi/Multiracial). Latent mean differences were found across grade level and race/ethnicity but not gender, with lower grades reporting higher social desirability than upper grades and Bi/Multiracials reporting lower social desirability than Asians and Hispanics/Latines. Findings provide preliminary evidence for the use of CSD-S in detecting social desirability bias and assessing meaningful social desirability differences in diverse children and adolescents.

Measurement invariance of the inventory of teacher-student relationships among middle and high school students.

Teacher-student relationship quality (TSRQ) predicts academic motivation (Wentzel, 1997), school engagement, and academic achievement (Hughes, 2011). However, TSRQ appears to differ across demographics. For example, boys and racially/ethnically minoritized students consistently have poorer relationships with their teachers than girls and White students (Koomen & Jellesma, 2015; Murray et al., 2008). Ensuring that TSRQ is consistently conceptualized across individuals will allow demographic differences on TSRQ to be compared. The present study aims to further validate a survey instrument used to measure TSRQ, called the Inventory of Teacher-Student Relationships (IT-SR; Murray & Zvoch, 2011). Participants included 3,541 middle and high school students in a large school district in the Southeastern United States. The results of the study confirmed the hypothesized three-factor structure of the instrument. The instrument demonstrated configural, metric, and scalar invariance across race/ethnicity (Black/African American, White, Hispanic/Latinx, and multiracial) and partial metric and scalar invariance across gender (boys and girls) and school level (middle school and high school). Significant latent mean differences were found, where boys, Black/African American students, Hispanic/Latinx students, and high school students reported lower scores on various factors on the IT-SR compared to girls, White students, multiracial students, and middle school students, respectively. Results support future research and applied use of the IT-SR with middle and high school students. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Reliability, Factorial Invariance, Latent Mean Differences of the Goal Orientations in Exercise Scale (GOES) and the Relation of Ego-Task Profiles with Resilience and Life Satisfaction in Ecuadorian Sample

The Achievement Goal Theory has been commonly used in the literature and the relation between the goal orientation construct and aspects of athletes’ personality and well-being is an important element to be considered in the sports field. The objectives of this study were: (i) to confirm the reliability, factorial invariance and existence of latent mean differences of the Goal Orientations in Exercise Scale (GOES) in an Ecuadorian sample across sex; (ii) to identify the existence of subgroups of Ecuadorian students with different profiles of goal orientations to exercise; and iii) to determine the differences in resilience and life satisfaction among these profiles. A total of 597 Ecuadorian students from the Sports Sciences degree participated in this study. Confirmatory factorial analysis supported the two-dimensional (ego-task) GOES structure. The GOES has an adequate reliability and structural invariance across sex. Significant latent mean differences on the ego orientation subscale were found across sex but not in the task one. Cluster analysis identified three task-ego orientation profiles: (1) moderately high on task and low on ego; (2) low on task and slightly high on ego; (3) high in both task and ego orientation. Post hoc comparisons showed statistical differences in resilience between the three profiles and in life satisfaction between the profile 3 and profiles 1 and 2. The current study allows the possibility of using the GOES as an invariant and reliable scale in Ecuadorian sample and initiates the use of goal orientation profiles and their relationship with resilience and life satisfaction in the sports field.

Willingness to Use Digital Health Screening and Tracking Tools for Public Health in Sexual Minority Populations in a National Probability Sample: Quantitative Intersectional Analysis.

Little is known about sexual minority adults' willingness to use digital health tools, such as pandemic-related tools for screening and tracking, outside of HIV prevention and intervention efforts for sexual minority men, specifically. Additionally, given the current cultural climate in the United States, heterosexual and sexual minority adults may differ in their willingness to use digital health tools, and there may be within-group differences among sexual minority adults. This study compared sexual minority and heterosexual adults' willingness to use COVID-19-related digital health tools for public health screening and tracking and tested whether sexual minority adults differed from each other by age group, gender, and race or ethnicity. We analyzed data from a cross-sectional, national probability survey (n=2047) implemented from May 30 to June 8, 2020, in the United States during the height of the public health response to the COVID-19 pandemic. Using latent-variable modeling, heterosexual and sexual minority adults were tested for differences in their willingness to use digital health tools for public health screening and tracking. Among sexual minority adults, specifically, associations with age, gender, and race or ethnicity were assessed. On average, sexual minority adults showed greater willingness to use digital health tools for screening and tracking than heterosexual adults (latent factor mean difference 0.46, 95% CI 0.15-0.77). Among sexual minority adults, there were no differences by age group, gender, or race or ethnicity. However, African American (b=0.41, 95% CI 0.19-0.62), Hispanic or Latino (b=0.36, 95% CI 0.18-0.55), and other racial or ethnic minority (b=0.54, 95% CI 0.31-0.77) heterosexual adults showed greater willingness to use digital health tools for screening and tracking than White heterosexual adults. In the United States, sexual minority adults were more willing to use digital health tools for screening and tracking than heterosexual adults. Sexual minority adults did not differ from each other by age, gender, or race or ethnicity in terms of their willingness to use these digital health tools, so no sexual orientation-based or intersectional disparities were identified. Furthermore, White heterosexual adults were less willing to use these tools than racial or ethnic minority heterosexual adults. Findings support the use of digital health tools with sexual minority adults, which could be important for other public health-related concerns (eg, the recent example of mpox). Additional studies are needed regarding the decision-making process of White heterosexual adults regarding the use of digital health tools to address public health crises, including pandemics or outbreaks that disproportionately affect minoritized populations.

Validation of factor structure of the neurodevelopmental parent report for outcome monitoring in down syndrome: confirmatory factor analysis.

The Neurodevelopmental Parent Report for Outcome Monitoring (ND-PROM), initially developed to monitor developmental and behavioral functions in children with autism spectrum disorder (ASD), assesses symptoms across a wide range of domains relevant in Down syndrome (DS). Psychometric properties of ND-PROM were assessed in 385 individuals with DS and 52 with a combined diagnosis of DS and ASD (DS+ASD), whose caregivers completed the ND-PROM questionnaire for a clinical visit in a specialized Down syndrome program at a tertiary pediatric hospital. Confirmatory factor analysis was conducted to evaluate the internal structure validity of the ND-PROM. Measurement invariance was assessed, with a comparison group of 246 individuals with ASD, and latent mean differences between the DS and ASD-only groups, as well as the combined DS+ASD groups, were assessed. Findings support the existence of the 12 clinically-derived factors in the DS population: Expressive Language, Receptive Language, Adaptive skills/Toileting, Social Emotional Understanding, Social Interaction, Independent Play, Sensory Processes, Challenging Behaviors, Impulse/ADHD, and Mental Health. Differences in response patterns of development and behaviors were observed between those with DS and those with ASD, including those with DS having higher abilities in nonverbal communication, social emotional understanding, and social interaction, and fewer restricted and repetitive behaviors and interests, impulsivity or ADHD symptoms, and mental health concerns compared to those with ASD. Individuals in the DS+ASD group had more difficulties with expressive and receptive language, nonverbal and social communication, social interaction, independent play, and adaptive skills than either the DS-only group or the ASD-only groups. The ND-PROM has a desirable factor structure and is a valid and clinically useful tool that captures a range of distinct and independent areas of developmental and behavioral functioning in DS, for individuals with and without an ASD diagnosis.

Are anxiety, depression, and stress distinguishable in Italian adolescents? an examination through the Depression Anxiety Stress Scales-21.

The adolescents' ability to discriminate between different negative emotional states is still under debate. The Depression Anxiety Stress Scales-21 (DASS-21) serves as a useful tool to unravel this issue, yet the literature on its structural validity in young people is ambiguous. Therefore, this study aimed to expand knowledge on the emotional experience of youth by investigating the factor structure and psychometric properties of the DASS-21 in Italian adolescents. Six hundred fifty-five students (60.6% girls) aged 14-18 (M = 16.3 ± 1.29) completed an online survey containing the DASS-21 and Positive and Negative Affect Schedule (PANAS). To evaluate the factor structure of the DASS-21, several alternative models were tested, also adopting an Exploratory Structural Equation Modeling (ESEM) procedure. Measurement invariance, reliability, validity, and latent means differences were addressed. The ESEM model with three correlated factors of Depression, Anxiety, and Stress yielded the best fit to the data, supporting a hierarchical structure of the DASS-21. In addition, this model was invariant across sex and age groups. The Anxiety scale predicted both positive and negative affect, while Depression predicted positive affect only. Finally, girls scored higher than boys on Anxiety and Depression, but no age differences emerged. Overall, our results indicate that anxiety, depression, and stress are distinguishable in Italian adolescents but, simultaneously, share an underlying condition of general distress, which may explain the comorbidity between internalizing problems. Such findings are discussed in terms of clinical and preventive implications for the adolescent population.

Provoked and unprovoked aggression in toddlerhood: Evaluating measurement invariance and latent means across gender, age, and time.

This paper re-examined the factor structure of a recently developed parent report of aggression, the Provoked and Unprovoked Aggression Questionnaire, and evaluated measurement invariance and latent mean differences across gender, age, and time. Participants were 333 mothers of toddlers (younger age group: n = 167, 53.9% boys, Mage Time 1 = 18.30 months, SD = 0.45; older age group: n = 166, 48.8% boys, Mage Time 1 = 24.29 months, SD = 0.38) who filled out provoked (eightitems) and unprovoked (fiveitems) aggression scales twice over a 3-month period. We found evidence for the configural, metric, and scalar measurement invariance-a requirement needed to make a meaningful comparison between aggression means across gender, age groups, and time. When comparing means for boys and girls, gender differences emerged earlier for unprovoked than provoked aggression. Also, the frequency of provoked aggression increased among younger, but not older, toddlers. By developing a brief parental report of aggression in toddlerhood, we hope to fill a void of early aggression measures that tap the contextual variability in aggressive behaviors, and thereby stimulate more research to further our understanding of different types of aggression and their correlates in toddlerhood.

School emphasis on academic success: the role of principal qualifications.

The purpose of the present study was to relate a principal's qualifications with a school's emphasis on academic success. Participants were n = 206 principals of respective schools in Saudi Arabia that took part in the study as a function of the TIMSS-2019 assessment. Principals were administered the eleven-item "School Emphasis on Academic Success" scale. A binary covariate defining low and high principal qualifications was computed. The Multiple Indicators Multiple Causes (MIMIC) model was involved with the latent means of both a general and three specific factors being regressed on the covariate to evaluate latent mean differences across differentially qualified principals. Furthermore, each one of the instrument's indicators was regressed on the principal covariate to evaluate the presence of Differential Item Functioning (DIF) or in other words additional effects due to item content. Results indicated a significant omnibus effect for the general factor only, with highly qualified principals holding significantly more positive beliefs about how parents, teachers, and students feel about their school's emphasis on academic success. Further analyses at the item level indicated that "teacher expectations" were the single item presenting a DIF effect with highly qualified principals having stronger beliefs about their teacher's expectations of student success over and above the latent factor mean. Results are discussed on how they inform educational policy and practice.

Investigating latent mean differences in achievement emotions among Chinese secondary EFL learners: A gender and grade perspective.

The control-value theory (CVT) of achievement emotions posits that achievement emotions are significantly associated with the key indicators of academic outcomes, including academic motivation, engagement, and performance. Existing studies have tested the theoretical hypothesis of the CVT in a variety of cultures, disciplines, and samples. However, evidence is limited for whether there are gender and grade differences in achievement emotions, especially in the context of English as a Foreign Language (EFL). 1,460 Chinese secondary school students (male N = 671; female N = 789; seventh-graders N = 731; eighth-graders N = 729) took part in the study. Confirmatory factor analyses and multi-group analyses were conducted to explore the possible gender and grade differences in EFL-related achievement emotions. Results indicated that there are gender or grade differences in EFL-related enjoyment, anxiety, and boredom, while hope and pride did not. Both limitations and implications are discussed.

Validation of an ultra-short global quality of life scale in a large population-based health survey.

Quality of life (QoL) assessment is essential in health-related research and clinical settings, offering insights into individuals' well-being and functioning. This study validated the Essential QoL-3 (EQoL-3), an ultra-short scale assessing essential dimensions of QoL, for use in epidemiological research and clinical settings. Data from a 2021 national survey in Norway (N = 17,487) were used. Three items on the EQoL-3 assess life satisfaction, happiness, and meaningfulness on a 0-10 scale. Discriminant validity was assessed by comparing the EQoL-3 with the Satisfaction with Life Scale (SWLS) and by examining latent mean differences between individuals with adverse life experiences (ALE+) and those without such experiences (ALE-). Convergent validity was evaluated through latent regression analyses comparing the EQoL-3 with a perceived mastery scale and a mental distress scale. The discriminant validity of the EQoL-3 was less than optimal when compared with the SWLS. Nonetheless, a multigroup confirmatory factor analysis revealed that the EQoL-3 score was 1.42 (95% CI = 1.33-1.50, p < 0.001) lower in the ALE+ group compared to the ALE- group, providing support for discriminant validity. Convergent validity was established with a positive association between EQoL-3 and mastery (β = 1.75, 95% CI = 1.70-1.80, p < 0.001) and a negative association between the EQoL-3 and mental distress (β = -2.64, 95% CI = -2.71/-2.59, p < 0.001). The EQoL-3 is a reliable measure of QoL. Its streamlined nature facilitates quick administration, making it a valuable tool for clinicians and researchers in diverse settings. Its inclusion of the eudaimonic dimension, as well as its exclusion of health items in the measure itself, distinguishes it from traditional HQoL measures, making it suitable for mental health and substance use disorder research.

Greek adaptation of the Teruel Orthorexia Scale (TOS) in adults from the Republic of Cyprus: A bidimensional model may not be universal

The Teruel Orthorexia Scale (TOS) is a 17-item instrument designed to assess distinct facets of Orthorexia Nervosa (ON) and Healthy Orthorexia (HO). While a bidimensional model of TOS scores has been supported in diverse national and linguistic groups, the psychometric properties of the TOS have not been previously assessed in Greek-speaking populations. To rectify this, we assessed the psychometric properties of a novel Greek translation of the TOS in a sample of adults from Cyprus. A total of 1248 respondents (710 women, 538 men) completed the Greek TOS, as well as previously validated measures of perfectionism, obsessive-compulsive symptomatology, eating restriction, negative affect, and appearance evaluation. Our results showed that a bidimensional model of the TOS had less-than-adequate fit when modelled using both confirmatory factor analysis and exploratory structural equation modeling (ESEM). Conversely, both exploratory factor analysis and ESEM supported extraction of a 3-factor model consisting of a HO facet and separate components of emotional orthorexia and cognitive-social orthorexia. This 3-factor model showed a lack of measurement bias (measurement invariance across gender identity and lack of differential item functioning as function of age and body mass index), but there were differences in latent factor means as function of respondent age and body mass index. The 3-factor model showed adequate evidence of construct validity, with the latent emotional orthorexia and cognitive-social orthorexia facets showing significant and moderate associations with the additional constructs measured in the survey. Broadly speaking, these findings support the psychometric properties of a 3-factor model of the Greek TOS, but also suggest that the bidimensional model of the TOS may not have universal applicability. We encourage further assessments of factorial validity in other national and linguistic contexts.