Examination of the latent structure of the psychological sense of school membership scale

Despite its widespread use, there has been limited examination of the underlying factor structure of the Psychological Sense of School Membership (PSSM) scale. The current study examined the psychometric properties of the PSSM to refine its utility for researchers and practitioners using a sample of 504 Australian high school students. Results from exploratory and confirmatory factor analyses indicated that the PSSM is a multidimensional instrument. Factor analysis procedures identified three factors representing related aspects of students’ perceptions of their school membership: caring relationships, acceptance, and rejection

A unified view on the effect of fibers and loading on sfrc creep through linear projection to latent structures

Creep of Steel Fiber Reinforced Concrete (SFRC) under flexural loads in the cracked state and to what extent different factors determine creep behaviour are quite understudied topics within the general field of SFRC mechanical properties. A series of prismatic specimens have been produced and subjected to sustained flexural loads. The effect of a number of variables (fiber length and slenderness, fiber content, and concrete compressive strength) has been studied in a comprehensive fashion. Twelve response variables (creep parameters measured at different times) have been retained as descriptive of flexural creep behaviour. Multivariate techniques have been used: the experimental results have been projected to their latent structure by means of Principal Components Analysis (PCA), so that all the information has been reduced to a set of three latent variables. They have been related to the variables considered and statistical significance of their effects on creep behaviour has been assessed. The result is a unified view on the effects of the different variables considered upon creep behaviour: fiber content and fiber slenderness have been detected to clearly modify the effect that load ratio has on flexural creep behaviour.

Differences between younger and older adults in the structure of suicidal intent and its correlates

OBJECTIVE: To analyze differences in the variables associated with severity of suicidal intent and in the main factors associated with intent when comparing younger and older adults. DESIGN: Observational, descriptive cross-sectional study. SETTING: Four general hospitals in Madrid, Spain. PARTICIPANTS: Eight hundred seventy suicide attempts by 793 subjects split into two groups: 18-54 year olds and subjects older than 55 years. MEASUREMENTS: The authors tested the factorial latent structure of suicidal intent through multigroup confirmatory factor analysis for categorical outcomes and performed statistical tests of invariance across age groups using the DIFFTEST procedure. Then, they tested a multiple indicators-multiple causes (MIMIC) model including different covariates regressed on the latent factor "intent" and performed two separate MIMIC models for younger and older adults to test for differential patterns. RESULTS: Older adults had higher suicidal intent than younger adults (z = 2.63, p = 0.009). The final model for the whole sample showed a relationship of intent with previous attempts, support, mood disorder, personality disorder, substance-related disorder, and schizophrenia and other psychotic disorders. The model showed an adequate fit (chi²[12] = 22.23, p = 0.035; comparative fit index = 0.986; Tucker-Lewis index = 0.980; root mean square error of approximation = 0.031; weighted root mean square residual = 0.727). All covariates had significant weights in the younger group, but in the older group, only previous attempts and mood disorders were significantly related to intent severity. CONCLUSIONS: The pattern of variables associated with suicidal intent varies with age. Recognition, and treatment of geriatric depression may be the most effective measure to prevent suicidal behavior in older adults.

New Advancements of Scalable Statistical Methods for Learning Latent Structures in Big Data

Constant technology advances have caused data explosion in recent years. Accord- ingly modern statistical and machine learning methods must be adapted to deal with complex and heterogeneous data types. This phenomenon is particularly true for an- alyzing biological data. For example DNA sequence data can be viewed as categorical variables with each nucleotide taking four different categories. The gene expression data, depending on the quantitative technology, could be continuous numbers or counts. With the advancement of high-throughput technology, the abundance of such data becomes unprecedentedly rich. Therefore efficient statistical approaches are crucial in this big data era.

Previous statistical methods for big data often aim to find low dimensional struc- tures in the observed data. For example in a factor analysis model a latent Gaussian distributed multivariate vector is assumed. With this assumption a factor model produces a low rank estimation of the covariance of the observed variables. Another example is the latent Dirichlet allocation model for documents. The mixture pro- portions of topics, represented by a Dirichlet distributed variable, is assumed. This dissertation proposes several novel extensions to the previous statistical methods that are developed to address challenges in big data. Those novel methods are applied in multiple real world applications including construction of condition specific gene co-expression networks, estimating shared topics among newsgroups, analysis of pro- moter sequences, analysis of political-economics risk data and estimating population structure from genotype data.

Un poison même à petite dose : étude de la structure latente et des facteurs développementaux du sadisme sexuel

Les délinquants sexuels sadiques sont généralement décrits comme une entité clinique particulière commettant des délits graves. Or, la notion même de sadisme sexuel pose un nombre important de problèmes. Parmi ceux-ci, on retrouve des problèmes de validité et de fidélité. Perçu comme une maladie dont on est atteint ou pas, le sadisme a été étudié comme si les sadiques étaient fondamentalement différents. À l’heure actuelle, plusieurs travaux laissent croire que la majorité des troubles psychologiques se présentent comme une différence d'intensité (dimension) plutôt qu’une différence de nature (taxon). Même si la conception médicale prévaut encore en ce qui concerne le sadisme sexuel, plusieurs évoquent l’idée qu’il pourrait être mieux conceptualisé à l’aide d’une approche dimensionnelle. En parallèle, nos connaissances sur les facteurs contributifs au développement du sadisme sexuel sont limitées et reposent sur de faibles appuis empiriques. Jusqu'à présent, très peu d'études se sont intéressées aux facteurs menant au développement du sadisme sexuel et encore moins ont tenté de valider leurs théories. En outre, nos connaissances proviennent majoritairement d'études de cas portant sur les meurtriers sexuels, un sous-groupe très particulier de délinquants fréquemment motivé par des intérêts sexuels sadiques. À notre connaissance, aucune étude n'a proposé jusqu'à présent de modèle développemental portant spécifiquement sur le sadisme sexuel. Pourtant, l'identification de facteurs liés au développement du sadisme sexuel est essentielle dans notre compréhension ainsi que dans l'élaboration de stratégie d'intervention efficace. La présente thèse s'inscrit dans un contexte visant à clarifier le concept de sadisme sexuel. Plus spécialement, nous nous intéressons à sa structure latente, à sa mesure et à ses origines développementales. À partir d'un échantillon de 514 délinquants sexuels évalué au Massachusetts Treatment Center, la viabilité d’une conception dimensionnelle du sadisme sexuel sera mise à l’épreuve à l'aide d'analyses taxométriques permettant d'étudier la structure latente d'un construit. Dans une seconde étape, à l'aide d'analyses de Rasch et d'analyses appartenant aux théories de la réponse à l'item à deux paramètres, nous développerons la MTC Sadism Scale (MTCSS), une mesure dimensionnelle du sadisme sexuel. Dans une troisième et dernière étape, un modèle développemental sera élaboré à l'aide d'équations structurales. La présente thèse permettra de contribuer à la clarification du concept de sadisme sexuel. Une clarification de la structure latente et des facteurs développementaux permettra de saisir les devis de recherche les plus à même de capturer les aspects essentiels. En outre, ceci permettra d'identifier les facteurs pour lesquels une intervention est la plus appropriée pour réduire la récidive, ou la gravité de celle-ci.

Un poison même à petite dose : étude de la structure latente et des facteurs développementaux du sadisme sexuel

Les délinquants sexuels sadiques sont généralement décrits comme une entité clinique particulière commettant des délits graves. Or, la notion même de sadisme sexuel pose un nombre important de problèmes. Parmi ceux-ci, on retrouve des problèmes de validité et de fidélité. Perçu comme une maladie dont on est atteint ou pas, le sadisme a été étudié comme si les sadiques étaient fondamentalement différents. À l’heure actuelle, plusieurs travaux laissent croire que la majorité des troubles psychologiques se présentent comme une différence d'intensité (dimension) plutôt qu’une différence de nature (taxon). Même si la conception médicale prévaut encore en ce qui concerne le sadisme sexuel, plusieurs évoquent l’idée qu’il pourrait être mieux conceptualisé à l’aide d’une approche dimensionnelle. En parallèle, nos connaissances sur les facteurs contributifs au développement du sadisme sexuel sont limitées et reposent sur de faibles appuis empiriques. Jusqu'à présent, très peu d'études se sont intéressées aux facteurs menant au développement du sadisme sexuel et encore moins ont tenté de valider leurs théories. En outre, nos connaissances proviennent majoritairement d'études de cas portant sur les meurtriers sexuels, un sous-groupe très particulier de délinquants fréquemment motivé par des intérêts sexuels sadiques. À notre connaissance, aucune étude n'a proposé jusqu'à présent de modèle développemental portant spécifiquement sur le sadisme sexuel. Pourtant, l'identification de facteurs liés au développement du sadisme sexuel est essentielle dans notre compréhension ainsi que dans l'élaboration de stratégie d'intervention efficace. La présente thèse s'inscrit dans un contexte visant à clarifier le concept de sadisme sexuel. Plus spécialement, nous nous intéressons à sa structure latente, à sa mesure et à ses origines développementales. À partir d'un échantillon de 514 délinquants sexuels évalué au Massachusetts Treatment Center, la viabilité d’une conception dimensionnelle du sadisme sexuel sera mise à l’épreuve à l'aide d'analyses taxométriques permettant d'étudier la structure latente d'un construit. Dans une seconde étape, à l'aide d'analyses de Rasch et d'analyses appartenant aux théories de la réponse à l'item à deux paramètres, nous développerons la MTC Sadism Scale (MTCSS), une mesure dimensionnelle du sadisme sexuel. Dans une troisième et dernière étape, un modèle développemental sera élaboré à l'aide d'équations structurales. La présente thèse permettra de contribuer à la clarification du concept de sadisme sexuel. Une clarification de la structure latente et des facteurs développementaux permettra de saisir les devis de recherche les plus à même de capturer les aspects essentiels. En outre, ceci permettra d'identifier les facteurs pour lesquels une intervention est la plus appropriée pour réduire la récidive, ou la gravité de celle-ci.

Measuring risk and blameworthiness in the sentencing of criminal defendants : an exploratory study in Western Australia's Higher Courts

The focal concern perspective dominates quantitative explorations of judicial sentencing. A critical argument underlying this perspective is the role of judicial assessments of risk and blameworthiness. Prior research has not generally explored how these two concepts fit together. This study provides an empirical test of the focal concerns perspective by examining the latent structure among the measures traditionally used in sentencing research, and investigates the extent to which focal concerns can be applied in a non-US jurisdiction. Using factor analysis (as suggested by prior research), we find evidence of distinct factors of risk and blameworthiness, with separate and independent effects on sentencing outcomes. We also identify the need for further development of the focal concerns perspective, especially around the role of perceptual shorthand.

Client's assessment of architects' performance in building delivery process: Evidence from Nigeria

The overall intent of this research is to provide architects with information that can be used to improve their performance so as to optimally satisfy the client's requirements and achieve high-quality overall project performance in Nigerian construction industry. Architect performance criteria were identified based on literature within the domain of architect responsibilities. The assessment of architects’ performance was carried out through a questionnaire survey of clients of recently completed building projects in Nigeria. Analysis of data includes comparison of criteria using importance–performance index analysis. Factor analysis was carried out on criteria where architects are falling below average, to group and explore the latent structure of the criteria in the data. The results showed that the architect needs to focus on management skills and ability, buildability, design quality, project communication, project integration and client focus. These results would encourage architects to perform better within their full responsibilities in the building delivery process and deliver high-quality projects within Nigerian construction industry.

Identification of Problem-based learning components and their effects on graduates' competencies

This study investigated the relationship between higher education and the requirement of the world of work with an emphasis on the effect of problem-based learning (PBL) on graduates' competencies. The implementation of full PBL method is costly (Albanese & Mitchell, 1993; Berkson, 1993; Finucane, Shannon, & McGrath, 2009). However, the implementation of PBL in a less than curriculum-wide mode is more achievable in a broader context (Albanese, 2000). This means higher education institutions implement only a few PBL components in the curriculum. Or a teacher implements a few PBL components at the courses level. For this kind of implementation there is a need to identify PBL components and their effects on particular educational outputs (Hmelo-Silver, 2004; Newman, 2003). So far, however there has been little research about this topic. The main aims of this study were: (1) to identify each of PBL components which were manifested in the development of a valid and reliable PBL implementation questionnaire and (2) to determine the effect of each identified PBL component to specific graduates' competencies. The analysis was based on quantitative data collected in the survey of medicine graduates of Gadjah Mada University, Indonesia. A total of 225 graduates responded to the survey. The result of confirmatory factor analysis (CFA) showed that all individual constructs of PBL and graduates' competencies had acceptable GOFs (Goodness-of-fit). Additionally, the values of the factor loadings (standardize loading estimates), the AVEs (average variance extracted), CRs (construct reliability), and ASVs (average shared squared variance) showed the proof of convergent and discriminant validity. All values indicated valid and reliable measurements. The investigation of the effects of PBL showed that each PBL component had specific effects on graduates' competencies. Interpersonal competencies were affected by Student-centred learning (β = .137; p < .05) and Small group components (β = .078; p < .05). Problem as stimulus affected Leadership (β = .182; p < .01). Real-world problems affected Personal and organisational competencies (β = .140; p < .01) and Interpersonal competencies (β = .114; p < .05). Teacher as facilitator affected Leadership (β = 142; p < .05). Self-directed learning affected Field-related competencies (β = .080; p < .05). These results can help higher education institution and educator to have informed choice about the implementation of PBL components. With this information higher education institutions and educators could fulfil their educational goals and in the same time meet their limited resources. This study seeks to improve prior studies' research method in four major ways: (1) by indentifying PBL components based on theory and empirical data; (2) by using latent variables in the structural equation modelling instead of using a variable as a proxy of a construct; (3) by using CFA to validate the latent structure of the measurement, thus providing better evidence of validity; and (4) by using graduate survey data which is suitable for analysing PBL effects in the frame work of the relationship between higher education and the world of work.

The Prospective and Retrospective Memory Questionnaire: A population-based random sampling study

The Prospective and Retrospective Memory Questionnaire (PRMQ) has been shown to have acceptable reliability and factorial, predictive, and concurrent validity. However, the PRMQ has never been administered to a probability sample survey representative of all ages in adulthood, nor have previous studies controlled for factors that are known to influence metamemory, such as affective status. Here, the PRMQ was applied in a survey adopting a probabilistic three-stage cluster sample representative of the population of Sao Paulo, Brazil, according to gender, age (20-80 years), and economic status (n=1042). After excluding participants who had conditions that impair memory (depression, anxiety, used psychotropics, and/or had neurological/psychiatric disorders), in the remaining 664 individuals we (a) used confirmatory factor analyses to test competing models of the latent structure of the PRMQ, and (b) studied effects of gender, age, schooling, and economic status on prospective and retrospective memory complaints. The model with the best fit confirmed the same tripartite structure (general memory factor and two orthogonal prospective and retrospective memory factors) previously reported. Women complained more of general memory slips, especially those in the first 5 years after menopause, and there were more complaints of prospective than retrospective memory, except in participants with lower family income.

Assessment of depression in medical patients: a systematic review of the utility of the beck depression inventory-II

To perform a systematic review of the utility of the Beck Depression Inventory for detecting depression in medical settings, this article focuses on the revised version of the scale (Beck Depression Inventory-II), which was reformulated according to the DSM-IV criteria for major depression. We examined relevant investigations with the Beck Depression Inventory-II for measuring depression in medical settings to provide guidelines for practicing clinicians. Considering the inclusion and exclusion criteria seventy articles were retained. Validation studies of the Beck Depression Inventory-II, in both primary care and hospital settings, were found for clinics of cardiology, neurology, obstetrics, brain injury, nephrology, chronic pain, chronic fatigue, oncology, and infectious disease. The Beck Depression Inventory-II showed high reliability and good correlation with measures of depression and anxiety. Its threshold for detecting depression varied according to the type of patients, suggesting the need for adjusted cut-off points. The somatic and cognitive-affective dimension described the latent structure of the instrument. The Beck Depression Inventory-II can be easily adapted in most clinical conditions for detecting major depression and recommending an appropriate intervention. Although this scale represents a sound path for detecting depression in patients with medical conditions, the clinician should seek evidence for how to interpret the score before using the Beck Depression Inventory-II to make clinical decisions

(Table 1) Circulating levels of TT3 and TT4, and age, length, girth, and estimated body mass in polar bears (Ursus maritimus) from East Greenland 1999-2001

We investigated the multivariate relationships between adipose tissue residue levels of 48 individual organohalogen contaminants (OHCs) and circulating thyroid hormone (TH) levels in polar bears (Ursus maritimus) from East Greenland (1999-2001, n = 62), using projection to latent structure (PLS) regression for four groupings of polar bears; subadults (SubA), adult females with cubs (AdF_N), adult females without cubs (AdF_S) and adult males (AdM). In the resulting significant PLS models for SubA, AdF_N and AdF_S, some OHCs were especially important in explaining variations in circulating TH levels: polybrominated diphenylether (PBDE)-99, PBDE-100, PBDE-153, polychlorinated biphenyl (PCB)-52, PCB-118, cis-nonachlor, trans-nonachlor, trichlorobenzene (TCB) and pentachlorobenzene (QCB), and both negative and positive relationships with THs were found. In addition, the models revealed that DDTs had a positive influence on total 3,5,3'-triiodothyronine (TT3) in AdF_S, and that a group of 17 higher chlorinated ortho-PCBs had a positive influence on total 3,5,3',5'-tetraiodothyronine (thyroxine, TT4) in AdF_N. TH levels in AdM seemed less influenced by OHCs because of non-significant PLS models. TH levels were also influenced by biological factors such as age, sex, body size, lipid content of adipose tissue and sampling date. When controlling for biological variables, the major relationships from the PLS models for SubA, AdF_N and AdF_S were found significant in partial correlations. The most important OHCs that influenced TH levels in the significant PLS models may potentially act through similar mechanisms on the hypothalamic-pituitary-thyroid (HPT) axis, suggesting that both combined effects by dose and response addition and perhaps synergistic potentiation may be a possibility in these polar bears. Statistical associations are not evidence per se of biological cause-effect relationships. Still, the results of the present study indicate that OHCs may affect circulating TH levels in East Greenland polar bears, adding to the "weight of evidence" suggesting that OHCs might interfere with thyroid homeostasis in polar bears.

Bayesian Inference in Large-scale Problems

Many modern applications fall into the category of "large-scale" statistical problems, in which both the number of observations n and the number of features or parameters p may be large. Many existing methods focus on point estimation, despite the continued relevance of uncertainty quantification in the sciences, where the number of parameters to estimate often exceeds the sample size, despite huge increases in the value of n typically seen in many fields. Thus, the tendency in some areas of industry to dispense with traditional statistical analysis on the basis that "n=all" is of little relevance outside of certain narrow applications. The main result of the Big Data revolution in most fields has instead been to make computation much harder without reducing the importance of uncertainty quantification. Bayesian methods excel at uncertainty quantification, but often scale poorly relative to alternatives. This conflict between the statistical advantages of Bayesian procedures and their substantial computational disadvantages is perhaps the greatest challenge facing modern Bayesian statistics, and is the primary motivation for the work presented here.

Two general strategies for scaling Bayesian inference are considered. The first is the development of methods that lend themselves to faster computation, and the second is design and characterization of computational algorithms that scale better in n or p. In the first instance, the focus is on joint inference outside of the standard problem of multivariate continuous data that has been a major focus of previous theoretical work in this area. In the second area, we pursue strategies for improving the speed of Markov chain Monte Carlo algorithms, and characterizing their performance in large-scale settings. Throughout, the focus is on rigorous theoretical evaluation combined with empirical demonstrations of performance and concordance with the theory.

One topic we consider is modeling the joint distribution of multivariate categorical data, often summarized in a contingency table. Contingency table analysis routinely relies on log-linear models, with latent structure analysis providing a common alternative. Latent structure models lead to a reduced rank tensor factorization of the probability mass function for multivariate categorical data, while log-linear models achieve dimensionality reduction through sparsity. Little is known about the relationship between these notions of dimensionality reduction in the two paradigms. In Chapter 2, we derive several results relating the support of a log-linear model to nonnegative ranks of the associated probability tensor. Motivated by these findings, we propose a new collapsed Tucker class of tensor decompositions, which bridge existing PARAFAC and Tucker decompositions, providing a more flexible framework for parsimoniously characterizing multivariate categorical data. Taking a Bayesian approach to inference, we illustrate empirical advantages of the new decompositions.

Latent class models for the joint distribution of multivariate categorical, such as the PARAFAC decomposition, data play an important role in the analysis of population structure. In this context, the number of latent classes is interpreted as the number of genetically distinct subpopulations of an organism, an important factor in the analysis of evolutionary processes and conservation status. Existing methods focus on point estimates of the number of subpopulations, and lack robust uncertainty quantification. Moreover, whether the number of latent classes in these models is even an identified parameter is an open question. In Chapter 3, we show that when the model is properly specified, the correct number of subpopulations can be recovered almost surely. We then propose an alternative method for estimating the number of latent subpopulations that provides good quantification of uncertainty, and provide a simple procedure for verifying that the proposed method is consistent for the number of subpopulations. The performance of the model in estimating the number of subpopulations and other common population structure inference problems is assessed in simulations and a real data application.

In contingency table analysis, sparse data is frequently encountered for even modest numbers of variables, resulting in non-existence of maximum likelihood estimates. A common solution is to obtain regularized estimates of the parameters of a log-linear model. Bayesian methods provide a coherent approach to regularization, but are often computationally intensive. Conjugate priors ease computational demands, but the conjugate Diaconis--Ylvisaker priors for the parameters of log-linear models do not give rise to closed form credible regions, complicating posterior inference. In Chapter 4 we derive the optimal Gaussian approximation to the posterior for log-linear models with Diaconis--Ylvisaker priors, and provide convergence rate and finite-sample bounds for the Kullback-Leibler divergence between the exact posterior and the optimal Gaussian approximation. We demonstrate empirically in simulations and a real data application that the approximation is highly accurate, even in relatively small samples. The proposed approximation provides a computationally scalable and principled approach to regularized estimation and approximate Bayesian inference for log-linear models.

Another challenging and somewhat non-standard joint modeling problem is inference on tail dependence in stochastic processes. In applications where extreme dependence is of interest, data are almost always time-indexed. Existing methods for inference and modeling in this setting often cluster extreme events or choose window sizes with the goal of preserving temporal information. In Chapter 5, we propose an alternative paradigm for inference on tail dependence in stochastic processes with arbitrary temporal dependence structure in the extremes, based on the idea that the information on strength of tail dependence and the temporal structure in this dependence are both encoded in waiting times between exceedances of high thresholds. We construct a class of time-indexed stochastic processes with tail dependence obtained by endowing the support points in de Haan's spectral representation of max-stable processes with velocities and lifetimes. We extend Smith's model to these max-stable velocity processes and obtain the distribution of waiting times between extreme events at multiple locations. Motivated by this result, a new definition of tail dependence is proposed that is a function of the distribution of waiting times between threshold exceedances, and an inferential framework is constructed for estimating the strength of extremal dependence and quantifying uncertainty in this paradigm. The method is applied to climatological, financial, and electrophysiology data.

The remainder of this thesis focuses on posterior computation by Markov chain Monte Carlo. The Markov Chain Monte Carlo method is the dominant paradigm for posterior computation in Bayesian analysis. It has long been common to control computation time by making approximations to the Markov transition kernel. Comparatively little attention has been paid to convergence and estimation error in these approximating Markov Chains. In Chapter 6, we propose a framework for assessing when to use approximations in MCMC algorithms, and how much error in the transition kernel should be tolerated to obtain optimal estimation performance with respect to a specified loss function and computational budget. The results require only ergodicity of the exact kernel and control of the kernel approximation accuracy. The theoretical framework is applied to approximations based on random subsets of data, low-rank approximations of Gaussian processes, and a novel approximating Markov chain for discrete mixture models.

Data augmentation Gibbs samplers are arguably the most popular class of algorithm for approximately sampling from the posterior distribution for the parameters of generalized linear models. The truncated Normal and Polya-Gamma data augmentation samplers are standard examples for probit and logit links, respectively. Motivated by an important problem in quantitative advertising, in Chapter 7 we consider the application of these algorithms to modeling rare events. We show that when the sample size is large but the observed number of successes is small, these data augmentation samplers mix very slowly, with a spectral gap that converges to zero at a rate at least proportional to the reciprocal of the square root of the sample size up to a log factor. In simulation studies, moderate sample sizes result in high autocorrelations and small effective sample sizes. Similar empirical results are observed for related data augmentation samplers for multinomial logit and probit models. When applied to a real quantitative advertising dataset, the data augmentation samplers mix very poorly. Conversely, Hamiltonian Monte Carlo and a type of independence chain Metropolis algorithm show good mixing on the same dataset.

Don’t expect me to vote for you just because I like you, even if you do make me feel warm inside: a comparison of the validity of non-ipsative measures of party support

The inclusion of non-ipsative measures of party preference (in essence ratings for each of the parties of a political system) has become established practice in mass surveys conducted for election studies. They exist in different forms, known as thermometer ratings or feeling scores, likes and dislikes scores, or support propensities. Usually only one of these is included in a single survey, which makes it difficult to assess the relative merits of each. The questionnaire of the Irish National Election Study 2002 (INES2002) contained three different batteries of non-ipsative party preferences. This paper investigates some of the properties of these different indicators. We focus in particular on two phenomena. First, the relationship between non-ipsative preferences and the choices actually made on the ballot. In Ireland this relationship is more revealing than in most other countries owing to the electoral system (STV) which allows voters to cast multiple ordered votes for candidates from different parties. Second, we investigate the latent structure of each of the batteries of party preferences and the relationships between them. We conclude that the three instruments are not interchangeable, that they measure different orientations, and that one –the propensity to vote for a party– is by far preferable if the purpose of the study is the explanation of voters’ actual choice behaviour. This finding has important ramifications for the design of election study questionnaires.