Constrained k‐means on cluster proportion and distances among clusters for longitudinal data analysis

Clustering individuals by measures of similarity or dissimilarity at trajectories of changes in longitudinal data enables determination of typical patterns of development and growth. The present research proposes a new constrained k‐means method with lower bound constraints on cluster proportions and distances among clusters at focused variables and time points to fulfill various needs in clustering longitudinal data. The method assumes a large number of clusters at the onset and iteratively deletes and combines clusters according to these constraints. An additional property of the proposed constrained k‐means includes direct estimation of the unknown number of clusters. Simulation results clearly show the usefulness of the method for extracting clusters in plausible, real‐life analysis including non‐normality within clusters, and the proposed algorithm works well and convergence of the estimates is satisfactory. An actual example using Japanese longitudinal data regarding sleep habits and mental health is presented to verify the utility of the proposed constrained k‐means.     

Direction of dependence in measurement error models

Methods to determine the direction of a regression line, that is, to determine the direction of dependence in reversible linear regression models (e.g., x→y vs. y→x), have experienced rapid development within the last decade. However, previous research largely rested on the assumption that the true predictor is measured without measurement error. The present paper extends the direction dependence principle to measurement error models. First, we discuss asymmetric representations of the reliability coefficient in terms of higher moments of variables and the attenuation of skewness and excess kurtosis due to measurement error. Second, we identify conditions where direction dependence decisions are biased due to measurement error and suggest method of moments (MOM) estimation as a remedy. Third, we address data situations in which the true outcome exhibits both regression and measurement error, and propose a sensitivity analysis approach to determining the robustness of direction dependence decisions against unreliably measured outcomes. Monte Carlo simulations were performed to assess the performance of MOM-based direction dependence measures and their robustness to violated measurement error assumptions (i.e., non-independence and non-normality). An empirical example from subjective well-being research is presented. The plausibility of model assumptions and links to modern causal inference methods for observational data are discussed.     

Football hooliganism: comparing self‐awareness and social identity theory explanations

The emergence of research interest in group violence in general and football hooliganism in particular has been explained by various models generally relying on situational or individual differences accounts. Yet, these two research traditions have largely evolved independently, showing little or no interaction. In the present study (N = 109), we integrate measures of these two approaches and the results reveal that social identity was more predictive of self‐reported physical aggression than of loss of private and public self‐awareness. Moreover, attitudes towards violence were the most marked predictor variables of both physical and verbal aggression. In the discussion, the moderator effect of social identity and attitudes about violence on physical aggression is elaborated upon. Copyright © 2007 John Wiley & Sons, Ltd.     

Analysis of Interactions and Nonlinear Effects with Missing Data: A Factored Regression Modeling Approach Using Maximum Likelihood Estimation

Abstract

When estimating multiple regression models with incomplete predictor variables, it is necessary to specify a joint distribution for the predictor variables. A convenient assumption is that this distribution is a multivariate normal distribution, which is also the default in many statistical software packages. This distribution will in general be misspecified if predictors with missing data have nonlinear effects (e.g., x²) or are included in interaction terms (e.g., x·z). In the present article, we introduce a factored regression modeling approach for estimating regression models with missing data that is based on maximum likelihood estimation. In this approach, the model likelihood is factorized into a part that is due to the model of interest and a part that is due to the model for the incomplete predictors. In three simulation studies, we showed that the factored regression modeling approach produced valid estimates of interaction and nonlinear effects in regression models with missing values on categorical or continuous predictor variables under a broad range of conditions. We developed the R package mdmb, which facilitates a user-friendly application of the factored regression modeling approach, and present a real-data example that illustrates the flexibility of the software.     

Randomization tests as alternative analysis methods for behavior‐analytic data

Randomization statistics offer alternatives to many of the statistical methods commonly used in behavior analysis and the psychological sciences, more generally. These methods are more flexible than conventional parametric and nonparametric statistical techniques in that they make no assumptions about the underlying distribution of outcome variables, are relatively robust when applied to small‐n data sets, and are generally applicable to between‐groups, within‐subjects, mixed, and single‐case research designs. In the present article, we first will provide a historical overview of randomization methods. Next, we will discuss the properties of randomization statistics that may make them particularly well suited for analysis of behavior‐analytic data. We will introduce readers to the major assumptions that undergird randomization methods, as well as some practical and computational considerations for their application. Finally, we will demonstrate how randomization statistics may be calculated for mixed and single‐case research designs. Throughout, we will direct readers toward resources that they may find useful in developing randomization tests for their own data.     

Validity and Adverse Impact Potential of Predictor Composite Formation

Previous research on the validity and adverse impact (AI) of predictor composite formation focused on the merits of regression‐based or ad hoc composites. We argue for a broader focus. Ad hoc chosen composites are usually not Pareto‐optimal, whereas the regression‐based composite represents only one element from the total set of Pareto‐optimal composites and can, therefore, provide only limited information on the potential for validity and AI reduction of forming predictor composites when both validity and AI are of concern. In that case, other Pareto‐optimal composites may provide a better benchmark to decide on the merits of the predictor composite formation. We summarize a method to determine the set of Pareto‐optimal composites and apply the method to a representative collection of selection predictors. The application shows that the assessment of the AI and validity of predictor composite formation can differ substantially from the one arrived at when considering only regression‐based composites.     

Neurocognitive predictors of performance‐based functional capacity in bipolar disorder

Neurocognitive impairment can predict functional capacity in individuals with bipolar disorder, though little research has examined whether different neurocognitive domains impact specific types of tasks. This study examined the relationship between several neurocognitive variables and the UCSD Performance‐Based Skills Assessment (UPSA; Patterson et al., 2011) to identify the domains and tests that best predict the performance across the subscales. Forty‐seven euthymic participants who were diagnosed with either Bipolar I or Bipolar II were recruited and assessed on a battery of neuropsychological measures and the UPSA. Correlational and regression analyses were run to identify neurocognitive predictors of UPSA subscales. Per the literature, verbal learning and memory and executive function composites were first examined. Verbal learning and memory predicted the Communication subscale and Total score variables above and beyond the estimated FSIQ and symptom rating scales. In a secondary exploratory analysis, the Benton Judgment of Line Orientation subtest predicted the Finance subscale while the California Verbal Learning Test predicted the UPSA total score. Verbal learning and memory emerged as the strongest predictor of functional capacity, suggesting that this domain should be investigated in future mediational and longitudinal studies with the UPSA.     

A simulated annealing methodology for clusterwise linear regression

In many regression applications, users are often faced with difficulties due to nonlinear relationships, heterogeneous subjects, or time series which are best represented by splines. In such applications, two or more regression functions are often necessary to best summarize the underlying structure of the data. Unfortunately, in most cases, it is not known a priori which subset of observations should be approximated with which specific regression function. This paper presents a methodology which simultaneously clusters observations into a preset number of groups and estimates the corresponding regression functions' coefficients, all to optimize a common objective function. We describe the problem and discuss related procedures. A new simulated annealing-based methodology is described as well as program options to accommodate overlapping or nonoverlapping clustering, replications per subject, univariate or multivariate dependent variables, and constraints imposed on cluster membership. Extensive Monte Carlo analyses are reported which investigate the overall performance of the methodology. A consumer psychology application is provided concerning a conjoint analysis investigation of consumer satisfaction determinants. Finally, other applications and extensions of the methodology are discussed.     

A criterion for selecting variables in a regression analysis

Methods are given for deciding whether to use some or no predictor variables in a regression analysis. Previously obtained results on the more general problem, whether to usek ork — r predictor variables are reviewed with emphasis on applications.     

Predictors of re‐employment: A question of attitude,behavior, or gender?

This longitudinal study examined the predictive value of attitudes, personal‐related variables, job search behaviour, and demographic variables on re‐employment among 142 assembly workers who had been made redundant. Participants completed a questionnaire within a week after leaving their jobs, and another 15 months later. Results of hierarchical logistic regression revealed that gender (being male), was the strongest predictor of re‐employment. Willingness to relocate and desire to change occupation also increased the odds of re‐employment 15 months after dismissal. On the other hand ‐ having children at home and anonymous‐passive job‐search behaviour, which is more prevalent among women, decreased the odds for re‐employment. The study is contributing to research by revealing gender differences in job search behaviour and the importance of focusing qualitative differences instead of merely quantitative measures in job‐search behaviour. And even more important, despite attitude and job‐search behaviour, there is still differences that seems to be related to gender and family responsibility.     

A mixed‐method approach to sense of coherence,health behaviors,self‐efficacy and optimism: Towards the operationalization of positive health attitudes

Posadzki, P., Stockl, A., Musonda, P. & Tsouroufli, M. (2010). A mixed‐method approach to sense of coherence, health behaviors, self‐efficacy and optimism: Towards the operationalization of positive health attitudes. Scandinavian Journal of Psychology, 51, 246–252. This study discusses the results of a cross‐sectional survey of healthy college students in Poland. More specifically, it describes, explores, and explains the relationships between psychological variables/models such as health behaviors (HB), sense of coherence (SOC), level of optimism (LOO), and self‐efficacy (SE) among college students. These separate constructs have also been used to operationalize a positive health attitude (PHA) as a novel construct. The social survey was carried out at three higher education institutions in Poland in January 2006. The random sample of 455 undergraduate students was taken from five different faculties: Physiotherapy, Physical Education, Tourism and Recreation, English Philology and Polish Philology. Four reliable and validated research tools were used to collect the data: Juczynsky’s Health Behaviour Inventory (HBI); Antonovsky’s Sense of Coherence Questionnaire (SOC‐29); Schwarzer & Jerusalem’s Generalized Self‐Efficacy Scale (GSES); and Seligman’s Scale (SS). The results indicate statistically significant differences (p < 0.001) between these four variables: for example, the healthier health behaviors the stronger the sense of coherence, level of optimism and self‐efficacy. It was also demonstrated that LOO, SOC, SE, and HB correlate with one another. Finally, these variables create an explicit empirical‐theoretical pattern. All the research results from REGWQ tests, Pearson’s correlation coefficient and cluster analysis suggest the existence of conceptual similarities between these four variables and/or the existence of some broader scientific construct such as PHA. However, this needs to be examined further. These results could be a good indicator for future research among different faculties or age groups.     

Uncovering the diverse cultural bases of social identity: Ingroup ties predict self‐stereotyping among individualists but not among collectivists

On what basis do people form their social identities? To investigate this issue, the present research investigates cross‐cultural differences in self‐stereotyping, a key outcome of social identification. In particular, the research tests the hypothesis that ingroup ties are a stronger predictor of self‐stereotyping among people from individualist cultures than among people from collectivist cultures. In Study 1, university students (N = 117) completed measures of ingroup ties and self‐stereotyping with respect to an intimacy group (family and friends). Consistent with predictions, ingroup ties significantly predicted self‐stereotyping among individualists but not among collectivists. Study 2 (N = 104) found a similar pattern of results among members of the global internet community who considered either an intimacy group (their friends), a task group (their work group) or a social category (their gender). These results indicate that people in individualist cultures are more likely than those in collectivist cultures to base their social identities on ingroup ties. The implications of these results are discussed in relation to self‐categorization theory's depersonalization account of social identification.     

Posttraumatic Growth and HIV: A Study on Associations of Stigma and Social Support

The present study investigated the influence of HIV‐related stigma and social support on posttraumatic growth (PTG) in adults with HIV (N = 126). The study examined if social support moderated the relationship between stigma and PTG. Results from the study revealed that the predictor variables contributed significantly to PTG following an HIV diagnosis; however, no significant interaction effect between the 2 variables was found. Implications for counselors and directions for future research are provided.     

Associations of self‐esteem and temperament traits to self‐ and teacher‐reported social status among classmates

Hintsanen, M., Alatupa, S., Pullmann, H., Hirstiö‐Snellman, P. & Keltikangas‐Järvinen, L. (2010). Associations of self‐esteem and temperament traits to self‐ and teacher‐reported social status among classmates. Scandinavian Journal of Psychology 51, 488–494. The present study examined the validity of self‐ and teacher‐reported social status and its relations to self‐esteem and temperament in a geographically representative cohort of Finnish adolescents (N = 3941, mean age 15.1 years). High agreement was found between self‐ and teacher‐rated social status (r = 0.42). Different aspects of self‐esteem were differently related to social status. When other temperament and self‐esteem variables were included in the analyses, strongest predictor for self‐rated social status in both genders was social self‐esteem (p < 0.01) and for teacher‐rated social status general self‐esteem (p < 0.01). The strongest temperamental predictors of social status were lower inhibition in girls (p < 0.01, self‐ and teacher‐rated) and higher impulsivity (p < 0.01, self‐rated) and activity (p < 0.01, teacher‐rated) in boys. The present findings are consistent with the view that social functioning and peer relations are associated with individual differences in self‐concept and temperament dimensions.     

Automatic thoughts and meta‐cognition as predictors of depressive or anxious symptoms: A prospective study of two trajectories

The purpose of this prospective study was to explore the Automatic Thought Questionnaire Negative (ATQ‐30‐N) and the Meta‐cognitions Questionnaire (MCQ‐30) as predictors in the development of depressive or anxious symptoms. A sample (N = 201) completed the ATQ‐30‐N, MCQ‐30, and the Hopkins Symptom Checklist‐25 (HSCL‐25) twice with a three month interval. The HSCL‐25 measures both depressive and anxiety symptoms. Separate multiple hierarchical regression analyses indicated that the ATQ‐30‐N was a positive predictor for levels of depressive symptoms, while the MCQ‐30 was a predictor of both levels of anxiety and depressive symptoms, at follow‐up, when controlling for gender, age and pre‐test levels of symptoms. However, the MCQ‐30 did not predict future levels of depressive symptoms, when levels of automatic negative thoughts measured by the ATQ‐30‐N were statistically controlled for. The findings suggested that the ATQ‐30‐N predicts future levels of depressive symptoms, while the MCQ‐30 primarily predicts future levels of anxiety.     

The relationships between contact,status and prejudice: An integrated threat theory analysis of Hindu–Muslim relations in India

In the context of Hindu–Muslim relations in India, the present study (N = 87) utilized Integrated Threat Theory (Stephan & Stephan, 2000 ) to examine the mediating roles of intergroup anxiety, realistic and symbolic threats and the moderating role of group membership (Hindu vs. Muslim) in the relationships between cross‐community contact, relative in‐group status and prejudice. Overall, intergroup anxiety and realistic, but not symbolic, threat emerged as proximal predictors of prejudice and partial mediators between the predictor and criterion variables. But these findings were qualified by majority (Hindu) versus minority (Muslim) group membership. As predicted, while symbolic threat was a predictor of prejudice for Hindus, realistic threat was a paramount predictor for Muslims. In‐group status was as a significant predictor for low‐status minority group only. The results are discussed with reference to their potential implications for future research and interventions aimed at improving intergroup relations. Copyright © 2008 John Wiley & Sons, Ltd.     

Self‐efficacy or perceived behavioural control: Which influences consumers' physical activity and healthful eating behaviour maintenance?

This research investigated the role of a commercial program, Michelle Bridges 12 Week Body Transformation (MB12WBT), that uses a variety of tools including social support to influence weight management behaviours such as exercising and eating healthfully. The need for research into weight management remains current given obesity rates have continued to increase in recent decades. Weight management is a complex behaviour that for many consumers involves barriers that influence their commitment to continue exercising and eating healthily. The Theory of Planned Behaviour is frequently used to explain, and to a lesser extent predict, behaviours such as physical activity and healthy eating; however, there is much debate as to whether a perceived behavioural control variable or self‐efficacy is the best predictor of behaviour. Structural equation modelling was used to analyse data from a sample of 724 respondents who answered a baseline and three month follow‐up survey. The analysis revealed self‐efficacy and perceived behavioural control were two distinct constructs. Self‐efficacy was a better predictor of behaviour than perceived behavioural control variables as suggested in the original Theory of Planned Behaviour. Furthermore, self‐efficacy had a significant impact on exercising and eating healthfully behaviour. The findings and implications for both behaviour change theory and practice are discussed.     

Test anxiety in UK schoolchildren: Prevalence and demographic patterns

Background. Despite a large body of international literature concerning the antecedents, correlates of and treatments for test anxiety, there has been little research until recently using samples of students drawn from the UK. There is a need to establish some basic normative data for test anxiety scores in this population of students, in order to establish whether international research findings may generalize to UK schoolchildren. Aim. To collect some exploratory data regarding test anxiety scores in a sample of UK schoolchildren, along with socio‐demographic variables identified in the existing literature as theoretically significant sources of individual and group differences in test anxiety scores. Sample. Key Stage 4 students (1348): 690 students in the Year 10 cohort and 658 students in the Year 11 cohort, drawn from seven secondary schools in the North of the UK. Method. Data on test anxiety were collected using a self‐report questionnaire, the Test Anxiety Inventory ( Spielberger, 1980 ) and additional demographic variables through the Student Profile Questionnaire. The factor structure of the Test Anxiety Inventory was explored using principal components analysis and multiple regression analysis used to predict variance in self‐reported test anxiety scores from individual and group variables. Results. The principal components analysis extracted two factors, worry and emotionality, in line with theoretical predictions. Gender, ethnic and socio‐economic background were identified as significant predictors of variance in test anxiety scores in this dataset. Whether English was an additional, or native, language of students did not predict variance in test anxiety scores and year group was identified as a predictor of emotionality scores only. Conclusion. Variance in the test anxiety scores of Key Stage 4 students can be predicted from a number of socio‐demographic variables. Further research is now required to assess the implications for assessment performance, examination arrangements and appropriateness of using a North American measure of test anxiety in a UK context.     

Religion,Prejudice, and Authoritarianism: Is RWA a Boon or Bane to the Psychology of Religion?

In research on religiosity and prejudice, right‐wing authoritarianism (RWA) has been studied alongside variables such as fundamentalism and orthodoxy. Four concerns regarding research on the relationship between RWA and religiosity are identified: (1) the overlap of religiosity and prejudice within the RWA scale; (2) the inflation of relationships by correlating part‐whole measures; (3) covariation in the extremes of the construct hiding the possible independence of components within RWA; and (4) statistical artifacts arising in multiple regression from the combination of these factors. We elaborate these four issues and then demonstrate how they can lead to different interpretations of some previously published data. The article concludes with suggestions for the management and resolution of these issues that may allow RWA to continue to be used in religiosity and prejudice research and how it might evolve to become the boon to researchers that they seek.     

Is the Third‐Person Effect Real? A Critical Examination of Rationales,Testing Methods,and Previous Findings of the Third‐Person Effect on Censorship Attitudes

Regarding the effect of the third‐person perception on censorship attitudes, the present study first highlighted logical weaknesses of previously proposed rationales and limitations of previously used statistical models to test the effect. This study reanalyzed data from past research on the effect of the third‐person perception. In Study 1, the average effects of the other‐self perceptual gap in the media influence on censorship attitudes were estimated based on reports from 13 previous studies (total N = 6,414). Study 2 reanalyzed Schmierbach, Boyle, Xu, and McLeod's (2011) correlational data (N = 692). The results of these two studies showed that the presumed effect on others is a stronger predictor of censorship attitudes than the other‐self differential in perceived media effect.