PREFACE: The purpose of this Master’s thesis was to determine which attributes of an early intervention (EI) mental health service would encourage treatment initiation and sustained engagement. This research was motivated by the bourgeoning interest in patient-centered care, particularly the incorporation of patient preferences into service design and implementation. Additionally, the research was inspired by the use of marketing research methodology in healthcare. Two conjoint surveys were formulated with the purpose of asking two questions: what EI service attributes will increase the likelihood of someone (1) initiating contact with an EI service and attending their first appointment, and (2) remaining engaged in treatment. A literature search determined which attributes would be the most relevant and important for conceptualizing an EI service. These attributes were narrowed down with the help of focus groups, key informant interviews, the expertise of the authors, and in the case of the second survey, was also informed by the first survey’s results. Each attribute was assigned four levels and these multi-level attributes were formulated into the two aforementioned surveys that were completed by mental health patients, their families, and mental health professionals. Chapter 1 of this thesis contains a short overview of the research literature investigating the benefits of EI services, some of the reasons why many patients may not receive such services, and some potential strategies to enhance patients’ initial contact and ongoing engagement with such services. In particular, the central tenet of this thesis is that patient engagement with EI services will be enhanced if service design considers and incorporates the preferences of patients and their families with regard to the attributes that characterize the service and its delivery. This hypothesis is explored using discrete choice conjoint experimental (DCE) methods to identify important service attributes regarding patient initiation and engagement. Given that DCEs are the central methodology of this thesis, Chapter 1 also includes an introduction to these methods and their unique benefits. Each of these service attribute questions posed above is addressed in a separate survey and experiment. Therefore, the rationale, methods, results and conclusions of each experiment are described in separate chapters (Chapters 2 & 3). It should be noted that these two chapters are written in the form of stand-alone scientific reports, each of which is about to be submitted for publication to peer-reviewed journals. Finally, the thesis concludes with a General Discussion (Chapter 4), which attempts to frame the two studies, and this line of inquiry more generally, in the broader research literature and highlight their clinical and policy implications. It should also be noted that, given that there exists substantial conceptual overlap between the two experiments and the main issues described in both the General Introduction and General Discussion, the reader may encounter some repetition throughout the thesis. / Mental illness places a large burden on individuals and society-at-large, a problem that becomes much worse the longer it is left untreated. Early intervention (EI) can mitigate this burden; however, those experiencing emerging mental illnesses often do not seek help promptly. Patient-centered care, such as shared-decision making models of mental healthcare, may reduce barriers to treatment. A central tenet of patient-centered care is that patient engagement and service utilization increases when patient preferences are incorporated into clinical services. In the current thesis, discrete choice conjoint experiments (DCE) were used to elicit the preferences of patients and their families, as well as the hypothesized preferences of patients according to mental health professionals, in two surveys. The first survey aimed to identify the attributes of an EI service that would encourage people experiencing psychiatric symptoms to initiate contact with a service and attend their first appointment (Chapter 2). The second survey sought to determine which service attributes would encourage someone to remain engaged with mental health treatment (Chapter 3). Both surveys used Latent Class Analysis to segment the study populations into identifiable subgroups based on shared preferences, and Randomized First Choice simulations to predict which service delivery model each of these identified groups would most likely use. The results of these studies have several implications for current and future mental health services. Effective EI mental health services should include rapid access to services, a range of treatment options, and effective crisis response. Moreover, future DCE studies should focus on replicating these results using more heterogeneous samples and improving DCE methods. / Thesis / Master of Science (MSc)

This thesis examines family formation in Scotland, with a focus on having a first birth. I argue that fertility behaviour must be researched as part of a process of family formation across the life course and informed by the conceptual framework of the Theory of Planned Behaviour. Starting from geographical patterns in fertility rates, housing and partnership, I establish relationships between shared social attitudes, partnership, housing, and family formation behaviour. I use cross-sectional and panel data from the Scottish Social Attitude Survey, the Scottish Census, the Scottish Longitudinal Study and the British Household Panel Study. The methodological contribution lies in the use of a Latent Class Analysis to identify shared social attitudes groups, which are then incorporated in more generic models predicting fertility intentions and first birth outcomes. The findings indicate that social attitudes and norms are important for family formation. While the Second Demographic Transition Theory argued that social norms are being replaced by more independent choices, more recent fertility literature has called for more attention to social norms. The current thesis thus supports these recent calls by showing how social norms might be (indirectly) measured in quantitative research. A key focus is on relationships between housing and family formation, especially normative ideas of 'proper' family housing. I demonstrate that family housing is a significant predictor of first births, at least for women. The gendered nature of family formation is also demonstrated by the different indicators found to predict men's and women's fertility intentions and realisations. I conclude that improved understanding of fertility behaviour requires longitudinal research that goes beyond the usual fertility indicators such as partnership and recognises the importance of gender differences, housing, shared social attitudes, and, above all, the continuing relevance of social norms in the family formation process.

The closing decades of the 20th century have witnessed a large increase in the numbers of young people remaining in education post-16 rather than entering the labour market. Concurrently, overall educational attainment in General Certificate of Secondary Education (GCSE) qualifications in England and Wales has steadily increased since their introduction in 1988. The 1990s represent a key period of change in these trends. Some sociologists argue that processes of detraditionalisation have occurred whereby previous indicators of social inequality, such as social class, are less relevant to the transitions of young people from school to work. Sociologists from other traditions argue that inequalities persist in the stratification of educational attainment by the family backgrounds of young people but that these factors have changed during this period. This thesis is an investigation of the influence of family background factors upon GCSE attainment during the 1990s. This includes extensive statistical analysis of measures of parental occupation, parental education and family structure with gender, ethnicity, school type and housing tenure type within the Youth Cohort Study of England and Wales. These analyses include over 100,000 respondents in 6 cohorts of school leavers with the harmonisation of data from cohort 6 (1992) to the Youth Cohort Time Series for England, Wales and Scotland 1984-2002 (Croxford, Ianelli and Shapira 2007). By adding the 1992 data to existing 1990s cohorts, the statistical models fitted apply to the complete set of 1990s cohorts and are therefore able to provide insight for the whole of this period. Strong differentials by parental occupation persist throughout the 1990s and do not diminish despite the overall context of rising attainment. This relationship remains net of the other factors listed, irrespective of the measure of parental occupation or the GCSE attainment outcome measure used. This builds upon and supports previous work conducted using the Youth Cohort Study and suggests that stratification in educational attainment remains a significant factor. Gender and ethnicity remain further sources of persistent stratification in GCSE attainment. Following a discussion of the weighting system and features of the Youth Cohort Study as a dataset, a thorough investigation of missing data is included, with the results of multiply imputed datasets used to examine the potential for missing data to bias estimates. This includes a critique of these approaches in the context of survey data analysis. The findings from this investigation suggest the importance of survey data collection methods, the limitations of post-survey bias correction methods and provide a thorough investigation of the data. The analysis then develops and expands previous work by investigating variation in GCSE attainment by subjects studied, through Latent Class Analysis of YCS cohort 6 (1992). Of the four groups identified in the model, a clear division is noted between those middle-attaining groups with respect to attainment in Science and Mathematics. GCSE attainment in combinations of subjects studied is stratified particularly with respect to gender and ethnicity. This research offers new insight into the role of family background factors in GCSE attainment by subject combination.

This study was part of a larger multi-informant longitudinal study with a sample of 319 children (52% male, 48% female) ages 4-12 (mean= 7 years 9 months) whose parents had recently filed for divorce. Three annual waves of data from four informants were used for analysis: child self-report, mother, teacher, and observer report. The purpose of the study was to add to the understanding of multi-informant research and children’s psycho-social well-being. The first goal was to determine the consensus of children’s psycho-social well-being scores within informant across time, within child across informant, and between children over time. The second goal was to determine factors that contribute to the levels of consensus, such as, child gender, child age, child ethnicity, and length of parents’ separation, maternal baseline depressive symptoms score, and timing of the teacher questionnaire. The third goal was to determine if children could be classified into meaningful psycho-social well-being groups. Lastly, a visual diagnostic tool, the “patchwork”, was created using a random sample of eight prototypical cases of group membership based on predicted probabilities. This tool displayed the four informants scores, and child and informant characteristics. A single measure of child psycho-social well-being was created for each informant to compare rater consensus in hierarchical linear modeling. Latent class analysis was used to determine groupings. The HLM results indicate that 53% of the variance is within informants across time, 31% is within child across raters, and only 16% is between child over time. As expected, results showed more consensus of informants’ scores among girls than boys, the greatest consensus for children in middle childhood over other age groups, among Non-Hispanic White children compared to other ethnicities, and among spring reports than fall reports from teachers. Maternal baseline depressive symptoms score was significantly related to level of consensus of reporters, with greatest consensus when mother’s baseline depressive symptoms scores are at the mean (15.47). Mother’s scores of children’s psycho-social well-being decrease from highest scores of when baseline depressive symptoms score is 0, decreasing -.02 with each point increase in baseline depressive symptoms score. The results of the latent class analysis show two latent classes with maternal baseline depressive symptoms as a covariate best fit the data, one class with psycho-social well-being scores above the mean (N=258), and one with scores below the mean (N=61). Baseline data alone sufficiently models these groups and is chosen for parsimony over latent transition analysis. In sum, this study demonstrated benefits of multi-method multi-informant research, while acknowledging the strengths and biases that influence informant consensus of children’s psycho-social well being / text

Introduction: Mood disorders are a serious burden on Canadians. Physical and sedentary activity are easily modifiable risk factors for many diseases. An association with depression could have important implications Objective: To investigate any cross-sectional or longitudinal association between physical activity, sedentary activity, and depression in the National Longitudinal Survey of Children and Youth (NLSCY). Methods: These studies used both a stacked cross-sectional and a trajectory/latent class analysis design Univariate and multivariate multinomial logistic regressions were used to assess the relationship between physical and sedentary activity and depression using the emotional disorder-anxiety scale for children and youth available in the NLSCY. Results: When accounting for covariates, physically inactive respondents had increased odds of higher depressive symptom scores. Sedentary activity was only statistically significantly associated with depressive symptoms in the cross-sectional design. Conclusions: Physical inactivity is significantly associated to depressive symptomatology. The relationship between sedentary activity and depression is inconclusive.

Traditional latent class analysis (LCA) considers entropy R2 as the only measure of effect size. However, entropy may not always be reliable, a low boundary is not agreed upon, and good separation is limited to values of greater than .80. As applications of LCA grow in popularity, it is imperative to use additional sources to quantify LCA classification accuracy. Greater classification accuracy helps to ensure that the profile of the latent classes reflect the profile of the true underlying subgroups. This Monte Carlo study compared the quantification of classification accuracy and confidence intervals of three effect sizes, entropy R2, I-index, and Cohen’s d. Study conditions included total sample size, number of dichotomous indicators, latent class membership probabilities (γ), conditional item-response probabilities (ρ), variance ratio, sample size ratio, and distribution types for a 2-class model. Overall, entropy R2 and I-index showed the best accuracy and standard error, along with the smallest confidence interval widths. Results showed that I-index only performed well for a few cases.

The thesis "The Potential of Latent Class Analysis: the Czech Television Audience Case Study" deals with latent class analysis and it's potential as a segmentation method. Three different approaches are examined - latent class analysis for joint data from six research waves, simultaneous latent class analysis for each wave separately and latent class analysis for joint data transferred into dichotomous form. The results of all three methods are compared through secondary analysis of data from case study focused on audience's perception of the Czech Television; their strong and weak points are described and the most suitable solution is chosen. Based on these findings, four identified segments of the audience are then interpreted from the point of their characteristic features and sociodemographic parameters. This paper is supported by works of classic authors in the area of latent class analysis; the thesis refers to their findings that also served as basis for composition of the research hypotheses.

Ce mémoire a été fait sous la tutelle de Jean Proulx. Il a été évalué par un jury composé de Denis Lafortune, Jean Proulx et Catherine Arseneault. Suite à l'évaluation du mémoire, ce dernier a été recommandé à la liste d'honneur du doyen et du recteur, et a été soumis pour le prix du meilleur mémoire de la FESP 2022. / Encore à ce jour, le portrait de l’agresseur sexuel sadique rapporté dans la littérature ne fait pas consensus, et ce notamment dû aux limites des études desquelles il découle (ex. : définition du sadisme sexuel qui ne fait pas consensus, échantillon hétérogène, temporalité fixée à l’âge adulte). Dans l’objectif de combler ces limites, nous avons examiné, à partir d’un échantillon composé de 206 agresseurs sexuels de femmes adultes, incarcérés au Québec (Canada), et sur la base d’une définition du sadisme sexuel validée empiriquement, la Severe Sexual Sadism Scale (SESAS), et d’une centaine de variables s’étalant de l’enfance à la phase post-délictuelle, ce qui distingue les agresseurs sexuels sadiques (n = 69) des non-sadiques (n = 137). Pour ce faire, des analyses bivariées (chi-deux) ont été réalisées. Ces analyses ont révélé que les agresseurs sexuels sadiques et non-sadiques de femmes adultes se distinguent sur plusieurs aspects, notamment de nature développementale (ex. : victimisation avant l’âge de 18 ans), psychologique (ex. : personnalité, psychopathologie), sexologique (ex. : fantaisies sexuelles déviantes) et criminologique (ex. : modus operandi). En outre, nous avons examiné, auprès du même échantillon, si le sadisme sexuel constitue ou non un phénomène hétérogène. À cette fin, des analyses de classes latentes ont été réalisées sur les items de la SESAS, lesquelles nous ont révélé que le sadisme sexuel ne constitue pas un phénomène hétérogène, se manifestant différemment selon les caractéristiques de l’agresseur sexuel. Les résultats susmentionnés et leurs implications théoriques et cliniques sont discutés. / To date, there has been no consensus in the literature on the characteristics of the sadistic sexual aggressor. This has been largely the result of the limitations of studies of this phenomenon (e.g., idiosyncratic definitions of sexual sadism, heterogeneous samples, temporality fixed at adulthood). In an effort to address these limitations, we identified the distinguishing features of sadistic (n = 69) and nonsadistic (n = 137) sexual aggressors against adult women, all of whom were incarcerated in Quebec (Canada), using the Severe Sexual Sadism Scale (SESAS), an empirically validated measurement instrument for sexual sadism, as well as other variables. Bivariate (chi-square) analyses revealed that sadistic and nonsadistic sexual aggressors of adult women differed in several ways, including developmental (e.g., victimization before age 18), psychological (e.g., personality, psychopathology), sexological (e.g., deviant sexual fantasies), and criminological (e.g., modus operandi). To determine whether sexual sadism is a heterogeneous phenomenon, latent class analyses were conducted on the SESAS items. These analyses revealed that sexual sadism is not heterogeneous, manifesting itself differently depending on the characteristics of the sexual aggressor. The above results and their theoretical and clinical implications are discussed.

This study was designed to empirically derive latent classes of mothers who were sexually abused during childhood and to assess the association between depression, alcohol/drug use, supportive intimate partner, and specific classes. One hundred six women between the ages of 20 and 44 years (M = 27) who reported having been sexually abused during childhood (CSA) and 158 non-CSA mothers between the ages of 20 and 43 years (M = 23) were interviewed and assessed along six parenting dimensions. Logistic regression models evaluated the association between psychoemotional variables and specific classes. The final model consisted of three classes--53.2%, 31.7%, and 15.2%. Alcohol/drug use was not statistically significantly associated with either class. Maternal depression and intimate partner support were differentially associated with the three parenting classes. Empirical support is provided for distinct classes of mothers sexually abused in childhood. The data-driven categorization of CSA mothers provides research and clinical directions for future parenting of survivors of childhood sexual abuse.

L’objectif de la présente étude est de soutenir l’utilisation d’une approche écologique pour l’analyse des données d’une évaluation standardisée. Nous implémentons cette approche en utilisant une Analyse de Classes Latentes (ACL) avec covariables sur les données du test de lecture du Programme d’Analyse des Systèmes Éducatifs de la CONFEMEN (PASEC) des élèves francophones de 6e année du primaire au Cameroun (N=617) et montrons comment une telle approche peut susciter un regard nouveau sur les résultats de cette évaluation. En effet, les programmes d’Évaluations standardisées à Larges Échelles (ELE) visent à évaluer les apprentissages et les compétences des individus et fournissent des données d’aide à la décision dans beaucoup de pays (Hogan, 2017 ; Loye, 2011 ; Wagemaker, 2014). Dans le cadre de ces ELE, les performances sont estimées uniquement sur la base des réponses que fournissent les candidats et donc de leurs capacités cognitives (Zumbo et al, 2015). Pourtant, de plus en plus de recherches suggèrent de considérer la performance à un test comme un phénomène qui se réalise dans un réseau interconnecté de connaissances, de caractéristiques individuelles et de contextes particuliers (McNamara, 2007 ; McNamara et Roever, 2006 ; Mislevy, 2018 ; Zumbo et al, 2015). L’approche écologique, qui s’inscrit dans cette perspective, s’intéresse aux hypothèses contextuelles, sociales et culturelles lors de l’estimation des performances à un test (McNamara, 2007 ; McNamara et Roever, 2006 ; Zumbo et al, 2015). Nos résultats mettent en évidence des écologies de performances en lecture qui varient en fonction des régions. Dans chacune de ces écologies, les profils de performances en lecture qui s’y dégagent dépendent des caractéristiques des élèves et des caractéristiques des milieux scolaires et extrascolaires dans lesquels ils évoluent. Nous mettons ainsi l’emphase sur le caractère situé des performances à un test permettant une lecture plus équitable des performances des différents candidats. Ce faisant, nous formulons des recommandations qui tiennent compte des réalités contextuelles des candidats contrairement aux recommandations uniformes souvent formulées à la suite des résultats aux ELE. / The purpose of this study is to support the use of an ecological approach for the analysis of standardized data assessment. This study is implimented by Latent Classes Analysis (ACL) with covariables on the test reading data of the « Programme d’Analyse des Systèmes Éducatifs de la CONFEMEN » (PASEC) french-speaking students of 6th grade primary school in Cameroon (N = 617) and shows how this approach can change the way we interpret the results of this study. Indeed, Standardized Large Scale Assessments (ELE) programs aim to appraise the learning and skills of individuals and provide decision support data in many countries (Hogan, 2017 ; Loye, 2011; Wagemaker, 2014). In the context of these ELEs, performance is estimated solely based on the responses provided by the candidates and therefore of their cognitive abilities (Zumbo et al, 2015). Yet a growing body of research suggests viewing test performance as a phenomenon that occurs in an interconnected network of knowledge, individual characteristics, and particular contexts (McNamara, 2007 ; McNamara and Roever, 2006 ; Mislevy, 2018 ; Zumbo et al. , 2015). The ecological approach, which is part of this perspective, is interested in the contextual, social and cultural hypotheses when estimating the performance of a test (McNamara and Roever, 2006 ; McNamara, 2007 ; Zumbo et al, 2015). Our results highlight reading performance ecologies that vary by region. In each of these ecologies, the reading performance profiles that emerge depend on the characteristics of the students and the characteristics of the school and out-of-school environments in which they operate. We therefore put the emphasis on the situated character of the performances on a test allowing a fairer reading of the performances of the different candidates. In doing so, we formulate recommendations that take into account the contextual realities of applicants, unlike the standard recommendations often made following ELE results.