Associations Between Reward and Future-Related Orientations and General and Specific Mental Health Issues in Adolescence

Adolescence is characterised by a peak in sensation seeking accompanied by gradually developing self-control skills. Adolescents typically show steeper delay discounting performance than other age groups; a feature that is transdiagnostically related to a variety of mental health disorders. However, delay discounting performance is not a singular mental process but involves both risk/reward and future orientation elements, usually operationalised as probability/risk and time discounting tasks, respectively. To clarify the specific relations between the risk/reward and future orientation elements of delay discounting and different types of mental health problems, two bi-factor models and a series of structural equation models (SEMs) were fitted to multi-informant (parent and adolescent self-reported) mental health data from a large UK study. A transdiagnostic promotive role of future orientation was found using bi-factor modelling to separate general and dimension-specific mental health variation; however, this was limited to parent reports. In addition, future orientation was negatively associated with conduct problems and ADHD symptoms, but positively associated with emotional problems. Risk aversion was negatively associated with conduct problems, but positively associated with emotional and peer problems. The findings highlight that risk/reward and future orientation elements of delay discounting play partly distinct roles in different mental health problems and can serve both promotive and risk roles during adolescence. Findings also illuminate which elements of delay discounting should be intervention targets for different mental health concerns.

Are they equivalent? An examination of task variants of approximate number comparison

Nonverbal numerical ability supports individuals' numerical information processing in everyday life and is also correlated with their learning of mathematics. This ability is typically measured with an approximate number comparison paradigm, in which participants are presented with two sets of objects and instructed to choose the numerically larger set. This paradigm has multiple task variants, where the two sets are presented in different ways (e.g., two sets are presented either simultaneously or sequentially, or two sets are presented either intermixed or separately). Despite the fact that different task variants have often been used interchangeably, it remains unclear whether these variants measure the same aspects of nonverbal numerical ability. Using a latent variable modeling approach with 270 participants (Mage = 20.75 years, SDage = 2.03, 94 males), this study examined the degree to which three commonly used task variants tapped into the same construct. The results showed that a bi-factor model corresponding to the hypothesis that task variants had both commonalities and uniqueness was a better fit for the data than a single-factor model, corresponding to the hypothesis that task variants were construct equivalent. These findings suggested that task variants of approximate number comparison did not measure the same construct and cannot be used interchangeably. This study also quantified the extent to which general cognitive abilities were involved in both common and unique parts of these task variants.

Bi-factor and Second-Order Copula Models for Item Response Data

Bi-factor and second-order models based on copulas are proposed for item response data, where the items are sampled from identified subdomains of some larger domain such that there is a homogeneous dependence within each domain. Our general models include the Gaussian bi-factor and second-order models as special cases and can lead to more probability in the joint upper or lower tail compared with the Gaussian bi-factor and second-order models. Details on maximum likelihood estimation of parameters for the bi-factor and second-order copula models are given, as well as model selection and goodness-of-fit techniques. Our general methodology is demonstrated with an extensive simulation study and illustrated for the Toronto Alexithymia Scale. Our studies suggest that there can be a substantial improvement over the Gaussian bi-factor and second-order models both conceptually, as the items can have interpretations of discretized maxima/minima or mixtures of discretized means in comparison with discretized means, and in fit to data.

Neuroticism vulnerability factors of anxiety symptoms in adolescents and early adults: an analysis using the bi-factor model and multi-wave longitudinal model

Background

Neuroticism and stress are important vulnerability factors in the development and outcome of anxiety symptoms. However, as neuroticism is a heterogeneity trait, it is still unclear how different neuroticism factors contribute to anxiety symptoms independently or in conjunction with stress. Thus, different factors of neuroticism were extracted in the present longitudinal study using the bi-factor model. The prediction effect of these different factors on anxiety symptoms and their combined effects with stress in both adolescent and adult samples were examined.

Method

Participants (592 adolescents and 638 young adults) in Hunan China were included. In the initial assessment in our longitudinal study, participants were asked to complete measurements that assessed neuroticism, stress, and anxiety symptoms. Next, a monthly assessment of stress and anxiety symptoms was completed for the subsequent 6 months. The bi-factor model was used to extract different factors of neuroticism. The hierarchical linear model was used to analyze longitudinal multi-wave data.

Result

Several model fit indices were used to evaluate the bi-factor model fit for neuroticism (adolescent: Tucker-Lewis index (TLI) = 0.957, comparative fit index (CFI) = 0.973, RMSEA = 0.040, Chi-Square = 80.471; early adults: TLI = 0.957, CFI = 0.973, RMSEA = 0.042, Chi-Square = 88.465). The results of hierarchical linear modeling analyses indicated that the general factor of neuroticism possessed a predictive effect on anxiety symptoms (adolescents: F = 36.77, p < 0.0001, early adults: F = 30.44, p < 0.0001); The negative effect factor only had the prediction effect on anxiety symptoms in early adults (adolescents: F = 0.65, p > 0.05; early adults: F = 4.84, p < 0.05); No prediction of self-reproach factor was found on anxiety symptoms (adolescents: F = 3.79, p > 0.05; early adults: F = 0.02, p > 0.05); the interactive effects of the general factor and stress on anxiety symptoms were only found in early adulthood (adolescents: F = 0.13, p > 0.05; early adults: F = 11.55, p < 0.01).

Conclusion

Our results suggested that the bi-factor model achieved a satisfactory fit for neuroticism measurement and supported that the anxiety symptoms were induced by the main effects of the general factor in both age samples and the negative factor only in adults. The general factor of neuroticism, but not the negative factor could make an additive effect for anxiety symptoms in face of stress, which meant that the homogeneity of neuroticism played a more significant role in further anxiety symptoms than heterogeneity when coping with stress.

Confirmatory and bi-factor analysis of the Short Form Health Survey 8 (SF-8) scale structure in a German general population sample

BACKGROUND

The SF-8 is a short form of the SF-36 Health Survey, which is used for generic assessment of physical and mental aspects of health-related quality of life (HRQoL). Each of the 8 dimensions of the SF-36 is covered by a single item in the SF-8. The aim of the study was to examine the latent model structure of the SF-8.

METHOD

One-, two- and three dimensional as well as bi-factor structural models were defined and estimated adopting the ML- as well as the WLSMV-algorithm for ordinal data. The data were collected in a German general population sample (N = 2545 persons).

RESULTS

A two- (physical and mental health) and a three-dimensional CFA structure (in addition overall health) represent the empirical data information adequately [CFI = .987/.995; SRMR = .024/.014]. If a general factor is added, the resulting bi-factor models provide a further improvement in data fit [CFI = .999/.998; SRMR = .001]. The individual items are much more highly associated with the general HRQoL factor (loadings: .698 to .908) than with the factors physical, mental, and overall health (loadings: -.206 to .566).

CONCLUSIONS

In the SF-8, each item reflects mainly general HRQoL (general factor) as well as one of the three components physical, mental, and overall health. The findings suggest in particular that the evaluation of the information of the SF-8 items can be validly supplemented by a general value HRQoL.

Executive Functions and Externalizing Symptoms: Common and Unique Associations

In discussing the four papers in this special issue, we provide our perspective on the authors' contributions and suggest directions for future research. First, we highlight the usefulness of the bi-factor model for investigating relations among specific aspects of executive functions (EFs) and externalizing symptoms. Next, we examine the role of EFs as a protective factor that can moderate the relation between risk factors - specifically, callous-unemotional behaviors - on externalizing symptoms. And finally, we address the contributions of innovative measurement approaches to understanding the relations between EFs and externalizing symptoms, using the state-space grid methodology as an example.

Exploring the item sets of the Recovering Quality of Life (ReQoL) measures using factor analysis

Purpose

This paper presents two studies exploring the latent structure of item sets used in the development of the Recovering Quality of Life mental health outcome measures: ReQoL-10 and ReQoL-20.

Method

In study 1, 2262 participants completed an initial set of 61 items. In study 2, 4266 participants completed a reduced set of 40 items. Study 2 evaluated two formats of the questionnaires: one version where the items were intermingled and one where the positively worded and negatively worded items were presented as two separate blocks. Exploratory and confirmatory factor analyses were conducted on both datasets where models were specified using ordinal treatment of the item responses. Dimensionality based on the conceptual framework and methods effects reflecting the mixture of positively worded and negatively worded items were explored. Factor invariance was tested across the intermingled and block formats.

Results

In both studies, a bi-factor model (study 1: RMSEA = 0.061; CFI = 0.954; study 2: RMSEA = 0.066; CFI = 0.971) with one general factor and two local factors (positively worded questions and negatively worded questions) was preferred. The loadings on the general factor were higher than on the two local factors suggesting that the ReQoL scale scores can be understood in terms of a general factor. Insignificant differences were found between the intermingled and block formats.

Conclusions

The analyses confirmed that the ReQoL item sets are sufficiently unidimensional to proceed to item response theory analysis. The model was robust across different ordering of positive and negative items.

Electronic supplementary material

The online version of this article (10.1007/s11136-018-2091-1) contains supplementary material, which is available to authorized users.

Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): results from a German representative population sample

BACKGROUND

The Beck Hopelessness Scale (BHS) has been the most frequently used instrument for the measurement of hopelessness in the past 40 years. Only recently has it officially been translated into German. The psychometric properties and factor structure of the BHS have been cause for intensive debate in the past.

METHODS

Based on a representative sample of the German population (N = 2450) item analysis including item sensitivity, item-total correlation and item difficulty was performed. Confirmatory factor analyses (CFA) for several factor solutions from the literature were performed. Multiple group factor analysis was performed to assess measurement invariance. Construct validity was assessed via the replication of well-established correlations with concurrently assessed measures.

RESULTS

Most items exhibited adequate properties. Items #4, #8 and #13 exhibited poor item characteristics- each of these items had previously received negative evaluations in international studies. A one-dimensional factor solution, favorable for the calculation and interpretation of a sum score, was regarded as adequate. A bi-factor model with one content factor and two method factors (defined by positive/negative item coding) resulted in an excellent model fit. Cronbach's alpha in the current sample was .87. Hopelessness, as measured by the BHS, significantly correlated in the expected direction with suicidal ideation (r = .36), depression (r = .53) and life satisfaction (r = -.53). Strict measurement invariance could be established regarding gender and depression status. Due to limited research regarding the interpretation of fit indices with dichotomous data, interpretation of CFA results needs to remain tentative.

CONCLUSION

The BHS is a valid measure of hopelessness in various subgroups of the general population. Future research could aim at replicating these findings using item response theory and cross-cultural samples. A one-dimensional bi-factor model seems appropriate even in a non-clinical population.

SCL-90-R emotional distress ratings in substance use and impulse control disorders: One-factor, oblique first-order, higher-order, and bi-factor models compared

To fully understand the dimensionality of an instrument in a certain population, rival bi-factor models should be routinely examined and tested against oblique first-order and higher-order structures. The present study is among the very few studies that have carried out such a comparison in relation to the Symptom Checklist-90-R. In doing so, it utilized a sample comprising 2593 patients with substance use and impulse control disorders. The study also included a test of a one-dimensional model of general psychological distress. Oblique first-order factors were based on the original a priori 9-dimensional model advanced by Derogatis (1977); and on an 8-dimensional model proposed by Arrindell and Ettema (2003)-Agoraphobia, Anxiety, Depression, Somatization, Cognitive-performance deficits, Interpersonal sensitivity and mistrust, Acting-out hostility, and Sleep difficulties. Taking individual symptoms as input, three higher-order models were tested with at the second-order levels either (1) General psychological distress; (2) 'Panic with agoraphobia', 'Depression' and 'Extra-punitive behavior'; or (3) 'Irritable-hostile depression' and 'Panic with agoraphobia'. In line with previous studies, no support was found for the one-factor model. Bi-factor models were found to fit the dataset best relative to the oblique first-order and higher-order models. However, oblique first-order and higher-order factor models also fit the data fairly well in absolute terms. Higher-order solution (2) provided support for R.F. Krueger's empirical model of psychopathology which distinguishes between fear, distress, and externalizing factors (Krueger, 1999). The higher-order model (3), which combines externalizing and distress factors (Irritable-hostile depression), fit the data numerically equally well. Overall, findings were interpreted as supporting the hypothesis that the prevalent forms of symptomatology addressed have both important common and unique features. Proposals were made to improve the Depression subscale as its scores represent more of a very common construct as is measured with the severity (total) scale than of a specific measure that purports to measure what it should assess-symptoms of depression.

Cross-cultural confirmation of bi-factor models of a symptom distress measure: Symptom Checklist-90-Revised in clinical samples

Four decades have elapsed since the introduction for clinical and research purposes of the Symptom Checklist-90(-R). Yet, its underlying dimensional structure has not been clearly delineated. A shift has been observed in the methods utilized-from predominantly exploratory factor analytic in nature in the first two decades or so to different confirmatory methods in recent years. A need remains to search for a structure that remains invariant across samples and nations. In that context, the present study attempted to replicate and extend recent findings yielded in a Hungarian general population sample (N=2,874) with two psychiatric patient samples from Hungary (N=972) and The Netherlands (N=1,902). In doing so, four models were contrasted: the one-factor model, Derogatis' nine factor model, a second-ordered factor model, and a bi-factor model. The bi-factor model was shown to yield the closest fit to the data in both countries. Further studies are needed to determine the stable number and kind of subscale scores that reflect the specific (primary) symptoms best, that is, those subscales with minimal shared variance with the overall general psychological distress dimension.

Somatic symptom reports in the general population: Application of a bi-factor model to the analysis of change

OBJECTIVE

To investigate the latent structure of somatic symptom reports in the general population with a bi-factor model and apply the structure to the analysis of change in reported symptoms after the emergence of an uncertain environmental health risk.

METHODS

Somatic symptoms were assessed in two general population environmental health cohorts (AMIGO, n=14,829 & POWER, n=951) using the somatization scale of the four-dimensional symptom questionnaire (4DSQ-S). Exploratory bi-factor analysis was used to determine the factor structure in the AMIGO cohort. Multi-group and longitudinal models were applied to assess measurement invariance. For a subsample of residents living close to a newly introduced power line (n=224), we compared a uni- and multidimensional method for the analysis of change in reported symptoms after the power line was put into operation.

RESULTS

We found a good fit (RMSEA=0.03, CFI=0.98) for a bi-factor model with one general and three symptom specific factors (musculoskeletal, gastrointestinal, cardiopulmonary). The latent structure was found to be invariant between cohorts and over time. A significant increase (p<.05) was found only for musculoskeletal and gastrointestinal symptoms after the power line was put into operation.

CONCLUSIONS

In our study we found that a bi-factor structure of somatic symptoms reports was equivalent between cohorts and over time. Our findings suggest that taking this structure into account can lead to a more informative interpretation of a change in symptom reports compared to a unidimensional approach.

Dimensionality of the Center for Epidemiologic Studies Depression Scale: an exploratory bi-factor analytic study

Objective

The Center for Epidemiologic Studies Depression Scale (CES-D) is a widely used instrument for measuring depressive symptoms. Though conventional factor analytic evaluations supported the use of four sub-scales for the CES-D, existing studies have yet to adopt the bi-factor analytic approach in psychometric assessment of the 20-item inventory. The present study aimed to apply both confirmatory factor analysis and exploratory bi-factor analysis to evaluate the dimensionality of the CES-D.

Methods

Current scoring practice of the CES-D (single-factor, four-factor, and second-order models) was tested using confirmatory factor analyses in a sample of 706 Chinese persons with insomnia and depressive symptoms. As an alternative, exploratory bi-factor analysis was conducted to examine the utility of the general depression factor and specific factors.

Results

Existing measurement models on the CES-D did not provide an adequate model fit to the data in terms of model fit indices and discriminant validity. The bi-factor model revealed a general depression factor that accounted for the majority of the item variance. The three specific factors (somatic symptoms, positive affect, and interpersonal problems) provided little unique information over and above the general factor and plausibly represent a methodological artifact rather than a substantive factor.

Conclusion

The present study demonstrated empirical support for the bi-factor model as a realistic representation of the underlying structure of the CES-D. Researchers and clinicians are better served by simply using a single measure of depression.