Thesis (PhD)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: The practice of psychometric testing of cognitive functioning in South Africa is hampered by the paucity of normative data that adequately characterize our ethnically, linguistically, socioeconomically, and educationally diverse population. The general aim of this study was to ascertain whether cognitive tests developed in settings outside of the Western Cape urbanized area have valid application for clinical and research purposes in that area. Strategies used to achieve that aim included: 1) translation, adaptation, and subsequent administration of a compendium of tests in a sample of typically developing, coloured and white, 12- to 15-yearold, Afrikaans- and English-speaking adolescents; 2) evaluation of the relative impact of sociodemographic factors (age, sex, language, quality of education, and race) on test performance and the consequent derivation of appropriately stratified normative data; and 3) evaluation of the cross-cultural utility of the normative data by comparing data collected from the study sample to norms derived from other populations. Results indicated that sex and language of testing impacted minimally on cognitive functioning. In contrast, the pervasive and deleterious impact of disadvantaged quality of education on cognitive performance within typically developing adolescents was clearly demonstrated. For participants with advantaged quality of education, coloured race was associated with lower performance on measures of intelligence, semantic fluency, and one measure of attention. Furthermore, the results provided evidence of age-related increments in cognitive performance, particularly after the age of 12. For cognitive measures that were significantly affected by language, race, and quality of education, trends of a downward continuum of performance were demonstrated, from highest to lowest, as follows: 1) English-white-advantaged; 2) Afrikaans-white-advantaged; 3) Englishcoloured- advantaged; 4) English-coloured-disadvantaged; 5) Afrikaans-coloured-advantaged; and 6) Afrikaans-coloured-disadvantaged. Cross-cultural comparisons of norms showed that for some tests, norms derived from other populations were suitable for use in the study sample. For other tests, however, results showed that for certain subgroups, it was essential to use the stratified norms derived from the study in order to prevent misdiagnoses. / AFRIKAANSE OPSOMMING: Die psigometriese toetsing van kognitiewe funksionering word in Suid-Afrika gekniehalter deur 'n gebrek aan normatiewe data wat ons etnies, taalkundig, sosio-ekonomies en opvoedkundig diverse bevolking genoegsaam tipeer. Die algemene doel van hierdie studie was om vas te stel of kognitiewe toetse wat in omgewings buite die Wes-Kaapse stedelike gebied ontwikkel is, ook vir kliniese en navorsingsdoeleindes binne hierdie stedelike gebied aangewend kan word. Hiervoor is onder meer die volgende strategieë gevolg: 1) 'n kompendium toetse is vertaal, aangepas en vervolgens afgeneem onder ’n toetsgroep tipies ontwikkelende, bruin en wit, 12- tot 15-jarige, Afrikaans- en Engelssprekende adolessente; 2) die relatiewe impak van sosiodemografiese faktore (ouderdom, geslag, taal, opvoedingsgehalte en ras) op toetsprestasie, en die gevolglike verkryging van toepaslik gestratifiseerde normatiewe data, is beoordeel en 3) die kruiskulturele nut van die normatiewe data is beoordeel deur die data wat van die toetsgroep in hierdie studie verkry is, te vergelyk met norme wat van ander populasies bekom is. Die resultate toon dat geslag en die taal waarin die toets afgeneem word 'n minimale uitwerking op kognitiewe funksionering het. Daarenteen is duidelik bewys dat swakker gehalte opvoeding ’n verreikende en skadelike uitwerking op die kognitiewe funksionering van tipies ontwikkelende adolessente het. By deelnemers met beter gehalte opvoeding blyk daar 'n verband te wees tussen die bruin rassegroep en laer prestasie wat betref maatstawwe van intelligensie en semantiese vaardigheid, asook een maatstaf van konsentrasie. Voorts lewer die resultate bewys van ouderdomsverwante toenames in kognitiewe prestasie, veral ná die ouderdom van 12. Wat betref kognitiewe maatstawwe wat beduidend deur taal, ras en opvoedingsgehalte beïnvloed is, is 'n afwaartse prestasiekontinuum opgemerk wat van hoog na laag soos volg daar uitsien: 1) Engels-wit-bevoordeel, 2) Afrikaans-wit-bevoordeel, 3) Engels-bruin-bevoordeel, 4) Engelsbruin- benadeel, 5) Afrikaans-bruin-bevoordeel en 6) Afrikaans-bruin-benadeel. Kruiskulturele normvergelykings toon dat, wat sommige toetse betref, die norme wat van ander populasies bekom is ook geskik was vir gebruik onder die toetsgroep in hierdie studie. Wat ander toetse betref, het die resultate egter getoon dat dit by bepaalde subgroepe noodsaaklik is om die gestratifiseerde norme wat uit die betrokke studie afgelei is te gebruik ten einde verkeerde diagnoses te voorkom.

The median prevalence of smoking among women of childbearing age in the United States is 22.4%. Of women who identify themselves as smokers in the three months prior to conception, 55% quit during pregnancy; however, 40% of those who quit relapse and return to smoking within six months after delivery. Smoking has been identified as an important means of stress management among smokers in general, and though limited to the perinatal period, pregnancy-specific stress adds to a woman’s typical day-to-day stress burden. Little data exists as to the effect of SHS exposure on smoking status during pregnancy and the impact of SHS exposure on the maternal perception of stress is unknown. Due to limited evidence, a critical need exists to examine the relationships of perceived maternal stress, SHS exposure, and perinatal smoking status in order to better understand perinatal smoking behaviors. The purposes of this dissertation were to: 1) evaluate the literature examining the relationship between the variables of maternal stress, SHS exposure, and perinatal smoking status; 2) determine the reliability and validity of the Everyday Stressors Index (ESI) use in pregnant women; and 3) to investigate the impacts of maternal perception of everyday stress, and SHS exposure on perinatal smoking status. Evidence obtained from the critical review of the literature supported an association between psychosocial stress and smoking during pregnancy or postpartum. Little information regarding the role of SHS exposure on perinatal smoking status was discovered. Psychometric testing of the ESI demonstrated strong internal consistency reliability, and factor analysis yielded three factors capturing three important domains of everyday stress. SHS exposure emerged as the most significant predictor of smoking status. Persistent smokers/relapsers had the highest ESI scores, followed by quitters, and then nonsmokers. While ESI means decreased in all smoking status groups from the first to the third trimester, the magnitude of decrease was not predictive. A significant interaction effect of SHS exposure in the home and decrease in ESI score occurred in the quit group only with quitters 1.14 times more likely to experience a decrease in ESI score compared to smokers/relapsers.

One of the great challenges for positive psychology is that academics, researchers and authors are yet to reach agreement on the terminology, constructs and methodology of the various positive psychological concepts and interventions. One such concept that has been ambiguously and often inconsistently conceptualised is happiness. Interventions that are structured around these conceptualisations have produced mixed results, which are attributable to various aspects such as the unidimensional conceptualisation of the concept, inconsistent measurement, psychometric measuring instruments that have not been validated for the population in question, and fragmented intervention methodologies. Research was, therefore, needed regarding happiness and the development of interventions aimed at increasing happiness. Therefore, this thesis aimed to investigate the concept, manifestation, measurement and development of happiness and positive psychological interventions (PPIs) within a tertiary educational environment. A mixed method research approach was followed to reach the research objectives. The first study aimed to examine the psychometric properties of positive psychological assessment measures, and to determine the relationship between flourishing and academic performance within a tertiary educational institution. A quantitative cross-sectional survey design was used to address this objective. The Mental Health Continuum Long Form, Positive and Negative Affect Scale and Satisfaction with Life Scale were administered and indicated acceptable levels of internal consistency. The MHC-LF would need to be adapted for future studies within the tertiary educational environment. No significant correlations could be established between academic performance and flourishing. However, cross-tabulation indicated some evidence of the impact that flourishing/languishing had on academic performance. Individuals with moderate levels of flourishing typically performed in the „above average‟ to „excelling academically‟ range. Languishing individuals performed at the lower levels of the academic performance spectrum. The results indicated that a large number of individuals within this tertiary educational environment were languishing. Flourishing students experienced higher levels of positive affect and satisfaction with life, as well as lower levels of negative affect than their languishing and moderately flourishing counterparts. The second study aimed to investigate the main streams of research on happiness, the approaches/models flowing from these philosophies and the methodology of happiness interventions. A quantitative meta-analysis of the literature was used to address the aforementioned research objective. Seeing that the literature presents with such incongruent findings regarding the effectiveness of PPIs, research was needed to establish how happiness should be conceptualised, which moderating factors should be addressed in this conceptualisation, and what the content of PPIs should be. A qualitative meta-analysis of the literature indicated that happiness is approached from either a hedonic, eudaimonic or integrated approach. These philosophies gave birth to a stream of scientific literature regarding happiness and its various conceptualisations, namely (a) Subjective well-being; (b) Hedonic well-being; (c) Eudaimonic well-being; (d) Psychological well-being; (e) Flourishing, and (f) Authentic happiness. Furthermore, the results highlighted three causes for failing interventions: (a) unidimensional models/approaches towards happiness, (b) targeting the concept of happiness instead of the mediating factors; and (c) fragmented methodological interventions. The study suggested a multidimensional model for happiness and happiness interventions. Furthermore, the study proposed a multifaceted methodology for happiness interventions, comprising self-administered intentional activities, group-administered interventions and individual coaching. The third study aimed to evaluate a PPI aimed at increasing happiness of students in a tertiary educational institution. A longitudinal pre-experimental research design was used to address the aforementioned research objective. Qualitative data were used to explore the findings further. The results indicated that the overall happiness of a student may be increased through developing individuals on both an emotional and psychological level. The results showed some scientific merit to a multifaceted approach towards PPIs. The PPI affected all the aspects conceptualised in this study except for two components of authenticity, namely authentic behaviour and relational orientations. / PhD, Industrial Psychology, North-West University, Vaal Triangle Campus, 2012

Cross-cultural assessment in South Africa has become more prominent since the first democratic elections held in April 1994, and stronger demands for the cultural appropriateness of psychological tests have arisen. The use of psychometric testing, including personality assessment in the workplace, is now strictly controlled by legislation, among others the Constitution of the Republic of South Africa (Act 108 of 1996), the Labour Relations Act (66 of 1995), and the Employment Equity Act (55 of 1998), and the Health Professions Act (56 of 1974). Much controversy has arisen regarding the relevance and applicability of assessment instruments in South Africa. The majority of assessment procedures still make use of imported instruments that are either used in their original or adapted form. Psychological assessment instruments imported from abroad have an insufficient suitability in the multicultural South African context. There are various perspectives regarding the appropriate measurement of personality across cultures. In this research study implicit perspectives of personality, the lexical approach, indigenous psychology and the emic approach were used to determine the personality perspectives of the Tsonga culture in South Africa. The objectives of this study were to investigate how personality is conceptualised in literature, to identify the problems surrounding personality measurement for the South African context, to explore how personality perspectives could be determined and to investigate the personality descriptive terms in the Tsonga language group. A qualitative research design was used to collect the data of this research. A total of 5 502 personality descriptors were obtained through the 1 0-item interview questionnaires. Content analysis was used to analyse, reduce and interpret the data obtained from the participants. The personality descriptors obtained were reduced by removing superfluous words. These personality descriptors were then interpreted and categorised into a total of 109 personality dimensions. These characteristics were categorised into nine clusters, namely Optimism, Agreeableness, Emotional Stability, Narrow-mindedness, Intelligence, Conscientiousness, Aggressiveness, Dominance and Sociability. The following personality dimensions had the highest frequency: Emotional Stability, Caring, Helpful, Hard working, Advising, Generous, Traditional, Aggression, Recreational, Substance use, Religious, Sociable and Loving. Recommendations for future research were made. / Thesis (M.A. (Industrial Psychology))--North-West University, Potchefstroom Campus, 2007.

Business leaders are increasingly coming to recognise that there is more to business success than technical and cognitive competence. Personnel leadership is proving to be critical for business bottom-line achievements considering that most business outcomes are achieved through human capital. Emotional intelligence can be used to the advantage of organisations by developing an emotional intelligence audit. The objective of this study was to investigate the psychometric properties of the Emotional Intelligence Scale (EIS) developed by Schutte, Malouff, Hall, Haggerty, Cooper, and Golden (1998) for employees in the mining industry in South Africa. The psychometric soundness of the EIS was tested. The general objective of the research was to standardise a psychometric instrument of emotional intelligence and determine the validity of the Schutte Emotional Intelligence Questionnaire (EIS) (Schutte, et al., 1998). A valid and reliable measure of emotional intelligence could be valuable in the organisation to identify specific EI needs that could be developed through the implementation of EI development programmes. In this context a standardised psychometric instrument of EI could be of use in organisations during the training and development of employees. A cross-sectional survey research method with an availability sample (N = 324) from employees in the mining industry was used. Exploratory factor analysis was used to confirm the construct equivalence of the EIS in different racial pups. The results supported a one-factor model of emotional intelligence. ANOVA analysis was conducted to determine the levels of emotional intelligence in terms of various demographic aspects such as gender, racial groups, language groups, and age groups. The results indicated that there are no significant differences between different demographics aspects and the level of emotional intelligence. Recommendations for future research were made. / Thesis (M.A. (Industrial Psychology))--North-West University, Potchefstroom Campus, 2007.

There is a distinct lack of research into the concept of acute mental health crisis. Without investigating the concept of crisis itself, it is not possible to appreciate the attributes of crisis so that it can be measured. This has hampered the development of good psychometric tools for crisis. The aim of this research was to develop the first standardised, valid and reliable measure for the assessment of people presenting to Crisis Resolution and Home Treatment (CRHT) teams. This research utilised qualitative and quantitative research techniques to develop a crisis measure starting with a comprehensive investigation into the concept of acute mental health crisis to identify an item pool and clinically credible item rating scale. A prototype crisis measure was developed and piloted in two NHS CRHTs and data collected. This data was analysed to identify the key areas of crisis assessment (the subscales), a flexible rating scale and scoring system creating a measure named the Crisis Risk and Adaptive Functioning Tool (CRAFT). The CRAFT provides patient crisis profiles highlighting areas of strength, resilience, weakness and vulnerability. A thorough investigation of crisis was completed with CRHT staff and patients through interviews and focus groups. An initial 143 item pool and clinically credible item rating scale were identified and developed into a prototype pilot crisis measure. This measure utilised a flexible rating approach encapsulating both risk and protective factors believed to reflect clinical practice. The measure was piloted and the data analysed to assess the structure of the crisis measure’s item pool using the statistical techniques of Principal Component Analysis (PCA) and Rasch analysis. These analyses resulted in a 66 item measure with 8 unidimensional subscales including; 1) Crisis Recovery Indicators, 2) Adaptive Decision Making, 3) Risk of Harm to Self, 4) Mediating Factors, 5) Daily Structure, 6) Risk of Harm to Others, 7) Feelings and Affect, and 8) Basic Needs. The total variance explained by these 8 subscales was 67.6% with excellent internal reliability as indicated by a Cronbach’s alpha coefficient of 0.98 (p<0.001) and temporal reliability indicated by Spearman’s correlation of 0.971 (p<0.001, one tailed). This suggests that this measure has a strong internal structure and provides stable outcomes over time at both the subscale and global overall measurement levels. Receiver Operator Characteristic curve analysis supported the identification of cut-offs to indicate low, moderate and high levels of crisis and were shown to have good levels of sensitivity and specificity for the crude discrimination between individuals who require CRHT treatment and individuals who do not require CRHT treatment (sensitivity 0.89 and specificity 0.73) and for accurately discriminating between the basic treatment levels of low, moderate and high (sensitivity 0.80; specificity 0.69). One of the great advantages of utilising the Rasch model is that it supports the identification of key characteristics from an item pool. Application of the subscales and the overall measure to the Rasch model identified items that were most representative of underlying constructs and risk, highlighting items of essential essence for assessing crisis in the context of community treatment. These items may act as useful clinical and risk indicators for community assessment. After considering the evidence from the PCA and Rasch analysis for the underpinning construct, the measure was named the Crisis Risk and Adaptive Functioning Tool (CRAFT) to encapsulate both the risk and adaptive functioning (coping and management) aspects measured by the tool. There are a number of clinical implications resulting from the development of CRAFT for the assessment of crisis. This research clearly identifies 8 key areas for crisis assessment and the specific items that describe them. This promises to be a powerful clinical tool as it clarifies the main areas of concern and importance for crisis assessment and provides mental health professionals with a means of assessing and monitoring patients experiencing crisis. In addition to the clinical benefits offered by the CRAFT, it provides an approach to assessing and monitoring crisis to support further research in the area of acute mental health crisis. This research offers significant steps towards the development of a quality measure for crisis assessment. However, it is acknowledged that the process of measurement development is never complete. It simply evolves over time with the aim of coming closer to the valued direction.

Although most people have experienced word-finding difficulty at one time or another, there are no clinical instruments able to reliably distinguish normal age-related effects from pathology in word-finding impairment. Two experiments were conducted to establish a modified version of the Boston Naming Test (BNT) that includes latency times, the Boston Naming Test of Latencies (BNT-L), in order to improve the instrument's sensitivity to mild to moderate word-finding impairment. Experiment 1: Latency times on the 60-item BNT (Goodglass et al., 2001) for 235 healthy adults' ages 18-89 years were collected on a representative sample. Qualitative features of the BNT items, statistical analyses, IRT, and demographic considerations of age, gender, education, vocabulary, race and culture, helped create a reduced BNT-L version with 15 of the most discriminating items. Statistically sound and sophisticated normative tables are provided that adjust for unseen covariates. Response latencies did not indicate earlier age-related decline in an optimally healthy sample. Experiment 2: Twenty-three patients referred for neuropsychological testing were administered the BNT-L. Patients referred for evaluation of mild cognitive impairment or possible dementia produced significantly different response BNT-L latencies from the healthy sample whereas patients referred for mild brain injury evaluation did not. Normal word-finding problems were discussed in terms of serial stage models of lexical access, as well as in terms of automatic and controlled cognitive processes in younger and older adults. Statistical process for creating a psychometric instrument using latencies is illustrated.

Background: Recently, health plans have initiated pay-for-performance (P4P) programs to incentivize pharmacies to focus on medication-related quality measures. Little research exists as to what attributes help a pharmacy to perform well in a P4P program. Objectives: This study aimed to: (1) propose a conceptual framework evolved from theories pertaining to economic, psychology, and organizational behavior disciplines that may contribute to a pharmacy's performance in a financial incentive program; (2) develop theoretically-derived questionnaires designed to elicit from pharmacists and pharmacy management pharmacy characteristics and pharmacy personnel factors thought to be associated with achieving quality measures specified for a pharmacy P4P program; (3) psychometrically assess these two newly developed questionnaires designed to collect information from community pharmacies eligible to participate in a pharmacy P4P program; (4) assess the magnitude of the relationship between pharmacy-related and construct redundancy of latent variables identified in objective 3; and (5) examine pharmacy factors' association with community pharmacies achieving quality measures in a pharmacy P4P program. Methods: Two survey instruments were created from a theoretically-derived conceptual framework to measure pertinent pharmacy characteristics and pharmacy personnel factors using a multi-step, mixed-methods process. Questionnaire development entailed semi-structured interviews, item generation, expert content validation, and cognitive debriefings. Developed questionnaires were evaluated in a non-experimental, cross-sectional survey of pharmacists and pharmacy management. Pharmacy personnel surveyed for this study were affiliated with pharmacies in Inland Empire Health Plan's (IEHP's) community pharmacy network. Using items measured in each questionnaire, specified groups of items were hypothesized to have a certain underlying latent variable. Latent variables hypothesized for items measured in the pharmacist questionnaire included: (1)"adherence management -services," (2)"adherence management-counseling," (3) "asthma management," (4) "pharmacist-prescriber relationship," (5) "chronic disease management," (6) "non-dispensing ability," and (7) "workload impact". For items measured in the pharmacy management questionnaire, eight latent variables were hypothesized: (1) "program understanding," (2) "program financial salience," (3) "program involvement," (4) "organization's adaptability," (5) "organization's innovativeness," (6) "organization's proactiveness," (7) "organization's risk-taking." and (8) "organization's focal emphasis." For hypothesized latent variables with four or more items, the value thresholds of three assessment indexes were employed as criteria to evaluate each measurement model's goodness-of-fit. Indexes included: (1) 2 test statistic (i.e.,>0.05), (2) root mean square error of approximation (RMSEA) (i.e., 0.05), and (3) comparative fit index (CFI) (i.e.,>0.95). The statistical significance of item factor loadings was assessed for latent variables with (1) four or more items deemed to have adequate model fit per index criteria and (2) only three items, for which goodness-of-fit may not be assessed. Structural modeling was used to explore associations between (1) pairs of latent variables with adequate model fit and/or statistically significant factor loadings; and (2) latent variables with medication-related quality measures. The strength of associations among latent variable pairs was measured with correlation coefficients. Latent variable pairs with significant correlations greater than 0.50 were evaluated for construct redundancy, for which a change in CFI was used as the comparison criteria to determine whether the latent variables are better represented as one (i.e., (CFI<-0.01) or two constructs (i.e., (CFI ≥ -0.01). Mediation-related quality measures specified as dependent variables in structural models included: diabetes medication adherence, hypertension medication adherence, hypercholesterolemia medication adherence, absence of controller therapy in patients with asthma, use of high-risk medications in elderly, and generic dispensing rate. Results: The empirically derived conceptual framework outlined four major domains proposed for consideration when evaluating pharmacy P4P programs: (1) incentive; (2) pharmacy; (3) other influencing factors; and (4) P4P program measures. From this framework, two questionnaires were developed with the aim of measuring pharmacists' and pharmacy management's pharmacy attitudes and perspectives of a community pharmacy P4P program. The pharmacist survey instrument consisted of five sections: (i) community pharmacy P4P program; (ii) pharmacy services; (iii) pharmacist workload; (iv) pharmacy practice site; and (v) pharmacist's background. The pharmacy management questionnaire consisted of four sections: (i) pharmacy P4P program; (ii) pharmacy organization's characteristics; (iii) pharmacy practice site; and (iv) respondent's background information. Among the 604 pharmacies invited to participate, pharmacists from 114 unique pharmacies (19%) and pharmacy management from 100 unique pharmacies (17%) participated in the survey. This study identified psychometrically validated measurement models for latent constructs with items measured in the pharmacist ("adherence management-counseling," "asthma management," "workload impact" ) and pharmacy management ("program involvement," "organization's innovativeness," "organization's proactiveness," "organization's risk-taking" ) questionnaires). With the exception of "program involvement," the latent constructs measured in the pharmacy management questionnaire were identified to be significantly correlated (𝛹>0.50, p<0.001) and more parsimoniously represented as one factor rather than two (∆CFI<-0.01). No significant associations were identified in exploratory analyses of validated measurement models from the pharmacist questionnaire with targeted medication-related quality measures in IEHP's community pharmacy P4P program. Conclusion: This study collected data for two questionnaires evaluating (1) pharmaceutical care services conceptualized to be associated with targeting medication-related quality measures and (2) pharmacy management's attitudes toward a pharmacy P4P program. Furthermore, psychometric assessment of each questionnaire supplied initial validity evidence for three constructs for the pharmacist questionnaire ("adherence management-counseling," "asthma management," and "workload impact" ) and four constructs for the pharmacy management questionnaire ("program involvement," "organization's innovativeness," "organization's proactiveness," and "organization's risk-taking" ). Subsequent validation of these latent constructs in larger sample sizes is required. Continued study of pharmacy factors and their association with medication-related quality measures is needed to improve our understanding of pharmacies' performance in financial incentive programs. As pharmacy organizations are increasingly integrated into more financial incentive programs aimed at targeting process- and outcomes-related quality measures, it is essential to be able to measure pharmacy and management factors in order to determine which factor(s) impact pharmacy performance.

There are several conditions for applying equal weights as an alternative to least squares weights. Psychometric parallelism, one of the conditions, has been suggested as a necessary and sufficient condition for equal-weights aggregation. The purpose of this study is to investigate the effect of psychometric parallelism among predictors on the efficiency of equal weights and least squares weights. Target correlation matrices with 10,000 cases were simulated so that the matrices had varying degrees of psychometric parallelism. Five hundred samples with six ratios of observation to predictor = 5/1, 10/1, 20/1, 30/1, 40/1, and 50/1 were drawn from each population. The efficiency is interpreted as the accuracy and the predictive power estimated by the weighting methods. The accuracy is defined by the deviation between the population R² and the sample R² . The predictive power is referred to as the population cross-validated R² and the population mean square error of prediction. The findings indicate there is no statistically significant relationship between the level of psychometric parallelism and the accuracy of least squares weights. In contrast, the correlation between the level of psychometric parallelism and the accuracy of equal weights is significantly negative. Under different conditions, the minimum p value of χ² for testing psychometric parallelism among predictors is also different in order to prove equal weights more powerful than least squares weights. The higher the number of predictors is, the higher the minimum p value. The higher the ratio of observation to predictor is, the higher the minimum p value. The higher the magnitude of intercorrelations among predictors is, the lower the minimum p value. This study demonstrates that the most frequently used levels of significance, 0.05 and 0.01, are no longer the only p values for testing the null hypotheses of psychometric parallelism among predictors when replacing least squares weights with equal weights.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.