Self-rated health (SRH) is an outcome commonly studied by demographers, epidemiologists, and sociologists of health, typically measured using an ordinal scale. SRH is analyzed in cross-sectional and longitudinal studies for both descriptive and inferential purposes, and has been shown to have significant validity with regard to predicting mortality. Despite the wide spread use of this measure, only limited attention is explicitly given to its unique attributes in the case of longitudinal studies. While self-rated health is assumed to represent a latent continuous and dynamic process, SRH is actually measured discretely and asymmetrically. Thus, the validity of methods ignoring the scale of measurement remains questionable. We compare three approaches to modeling SRH with repeated measures over time: linear multilevel models (MLM or LGM), including corrections for non-normality; and marginal and conditional ordered-logit models for longitudinal data. The models are compared using simulated data and illustrated with results from the Health and Retirement Study. We find that marginal and conditional models result in very different interpretations, but that conditional linear and non-linear models result in similar substantive conclusions, albeit with some loss of power in the linear case. In conclusion, we suggest guidelines for modeling self-rated health and similar ordinal outcomes in longitudinal studies. / text

Emotion language use provides insight into a person's emotional landscape. However, little is known about how preschool aged children and their parents use emotion language in their real world interactions. To address the shortcomings of the current body of empirical work on naturally occurring emotion language, this dissertation asks the following four research questions: 1) How do children and parents use emotion words in their daily lives?; 2) How is children's emotion language related to parents' emotion language?; 3) How is emotion language use related to emotional functioning?; and 4) How does emotion language in a lab setting compare to a natural setting? This dissertation implements a naturalistic methodology tool to answer the above questions. Thirty-five preschool aged children and their parents were recruited to participate in a two-wave longitudinal study in which the children wore a digital recording device for one day at each of the time points to capture acoustic information about the emotion language and behaviors they and their parents use in their daily lives. Additionally, participants completed a traditional laboratory based paradigm used to study emotion language within families. Parents also completed self-report measures related to emotion functioning for themselves and their child. Results indicate that children and their parents use high rates of positive emotion but very low rates of negative emotion in their naturally occurring interactions. This is different from lab based paradigms that elicit high rates of both positive and negative emotion language from children and parents. Next, children's use of emotion words tends to match the emotion language of their mothers more than their fathers but gender of the child also plays an important role. Very little support emerged for the emotion regulation model, as evidenced by children who cry and whine the most and who have the most behavioral problems tending to use negative emotions the most. The preponderance of the evidence suggests that emotion language reflects emotional state rather than regulates it. And, finally, the ecological validity of laboratory studies of emotion word usage is called into question by the independence of emotion language elicited in the lab and the natural expression of emotion words in a natural setting. Implications for researchers conducting work in the area of emotion language and emotional development are discussed. / text

Land use and transportation researchers have conducted numerous studies about land use effects on travel mode choice, and probed for effective policies to reduce driving, since less driving and more non-driving are widely recognized as more sustainable travel behaviors to resolve many environmental, energy and social equity issues. However, most of the previous studies rely on methodologies developed by cross-sectional data; only limited attention is explicitly given to explore the statistical techniques for longitudinal design and analysis. Using the neighborhood-level land use and persona-level travel mode choice data of 1997 and 2006 in the city of Austin, this paper attempts to establish and compare three distinct modeling approaches to analyze the trends of land use effects on people’s choice behavior of non-driving travel mode. The three modeling approaches are: a comparison approach with two cross-sectional multilevel Logit models using single-year data, a pooling approach by building one multilevel model with two-year data, and a longitudinal multilevel model. Empirical modeling results indicate that the longitudinal multilevel model is the most reasonable model for analyzing the longitudinal and multilevel datasets, since it is capable of estimating both time-invariant and time-variant land use effects, and internalizes time-variant random effects. The other two approaches may have several shortcomings. For example, the comparison approach fails to distinguish the time-variant and time-invariant effects; while the pooling model may lead to underestimated standard errors and t-statistics, and thus overestimate the significance of variables. / text

published_or_final_version / Paediatrics / Doctoral / Doctor of Philosophy

published_or_final_version / Medical Sciences / Master / Master of Medical Sciences

 Much of the research on parental involvement in education has focused on its positive influence on school-aged students in the western culture. This thesis examined Chinese parental involvement in the early years by conducting three empirical studies in Hong Kong and Shenzhen. Study 1 was designed to be qualitative in nature to examine kindergarten teachers’ and parents’ perceptions and practices of parental involvement by conducting focus group interviews with 35 teachers and 41 parents. Responses revealed that teachers and parents defined parental involvement differently, in which teachers defined it as parents assisting the teachers or the school, while parents defined it as the support they provide for their child’s learning and development. Respondents also had varying opinions about the current parental involvement practices and discussed different family, school and child factors that influence parental involvement. Study 2 investigated the association between parental involvement and children’s readiness for school using a sample of 431 children. Parents’ involvement was examined using parents’ self-report, whereas readiness for school was assessed using two child tests and parents’ report. Results indicated that parents practiced more home-based involvement than school-based involvement. All parental involvement dimensions significantly predicted overall readiness for school; however, only the dimensions of Language and Cognitive Activities and Home-School Conferencing were found the significantly predictors of children’s Chinese literacy and cognitive development. Study 3 was a follow-up investigation of the changes in parents’ perceptions and practices of parental involvement during the transition from kindergarten to primary school. Individual interviews were conducted with 18 parents three months after their child’s school entrance. Compared to their involvement during the kindergarten years, most parents indicated a shift to become more academic oriented in their involvement and expressed their dissatisfaction in the distant parent-teacher relationship in the primary school. Implications, limitations and future research directions on this topic are discussed. / published_or_final_version / Education / Doctoral / Doctor of Philosophy

Rumination has been consistently found to be a risk factor of depression. However, few studies examined rumination and its relationship with depression, and the protective role of such individual positive traits as hope, in a Hong Kong Chinese context. As opposed to western findings, a recent local study found that the two components of rumination (Treynor, Gonzalez, & Nolen-Hoeksema, 2003), namely brooding and reflective pondering, were not correlated, and the latter was rather adaptive in predicting depressive symptoms (Lo, Ho, & Hollen, 2008). Together with some psychometric issues revealed in the same study, it is unclear if the two-factor model of rumination is applicable in the Hong Kong Chinese context. In addition, given that hope only buffers against negative impact of risk factors of depression, whether the moderating effect of hope on the relationship of brooding and reflective pondering with depressive symptoms in Geiger and Kwon’s (2010) western sample can be replicated in Hong Kong Chinese populations is not clear. Therefore, the present study re-examines the two-factor model of rumination and its relationship with depressive symptoms, and the moderating role of hope in a Hong Kong Chinese sample using a longitudinal design. Adult Trait Hope Scale, Ruminative Response Scale, and Center for Epidemiologic Studies Depression Scale were administered to 189 Hong Kong Chinese university students at lectures with a four-week interval between the two time-points. Analyses using hierarchical linear regression were conducted to examine the relationship of brooding and reflective pondering, and their relationship with depressive symptoms and the role of hope, both concurrently and prospectively. Results confirmed the applicability of the two-factor model of rumination in the Hong Kong Chinese college sample. Rumination (total) and brooding consistently predicted depressive symptoms. Also, moderating effect of hope on the relationship of brooding and depressive symptoms was replicated using the time 1 data in the present study. Nonetheless, contrary to the recent local finding by Lo et al. (2008), brooding and reflective pondering were positively correlated as in western samples, and there was some evidence of one-way relationship from reflective pondering to brooding as shown by the longitudinal data. Reflective pondering itself was neither adaptive nor maladaptive, because there was no significant relationship between reflective pondering and depressive symptoms after controlling for level of brooding. Hence, hope had no interaction with reflective pondering in predicting depressive symptoms. For longitudinal data, after controlling for baseline level of the dependent variables, the one-way relationship from reflective pondering to brooding, the relationship of rumination (total) and brooding with depressive symptoms, and the moderating effect of hope on the relationship between brooding and depressive symptoms, all became nonsignificant. The findings of the present study support the use of the two-factor model of rumination in Hong Kong Chinese context, and suggest that the two components of rumination were correlated probably because reflective pondering tended to lead to brooding. This study also offers further empirical support for brooding being a more robust risk factor of depression than reflective pondering and the buffering impact of hope on psychological well-being. Implications of the results, limitations of this study and recommendations for further research were also addressed. / published_or_final_version / Psychology / Master / Master of Philosophy

Not available / text

<p>Research into how different life events shape individual religiosity has a long history within sociology. However, some scholars have begun to question whether research in this area has methodologically justified making strong causal claims. In an effort to re-center religion within the field of sociological concerns, quantitative sociologists of religion have tended to over-state the meaning of their statistical relationships and this has led to many of their causal assumptions being unstated and/or untested in their analyses. The advances in causal statistical modeling and counterfactually grounded analyses has led to the development of statistical models that are better able to establish causal relationships. It is time to begin implementing these approaches within the sociology of religion. This more rigorous statistical approach runs the risk of demonstrating that social life’s influence on religion may be less impressive than was previously thought. But researchers in this area must take this risk to develop a better sense of the real effects of society on religion. This in turn will provide a better foundation for developing theories of religion’s role in our modern world. </p><p> One way in which sociologists of religion can improve their causal modeling strategies is through the use of longitudinal data and methods. In recent years there has been a significant increase in the availability of large-scale longitudinal data that collects information on respondents’ religious belief, practice, and belonging. With these data, scholars interested in religious change can move away from their reliance on comparing individuals to one another – a constraint of cross-sectional data – and begin to analyze how certain life course events may lead to change in individual religiosity. I revisit two important areas within the sociology of religion –the relationship between family formation and religious service attendance and the effects of educational achievement on religious beliefs and practices – to assess whether these relationships can be considered causal in light of results from longitudinal statistical models. By implementing longitudinal models, I am able to directly assess whether between-individual differences or individual change over time is driving the statistical relationships found in my analyses. I will show that the story we thought we knew about how religion responds to family formation and educational attainment changes when these additional statistical tests are brought to bear on the data.</p> / Dissertation

This thesis aimed to evaluate the feasibility and utility of a sickle cell disease registry for clinical patient management and research. Five hospitals out of nine in the North West London health region participated in the registry, with 78 percent coverage of the sickle cell disease population. There was 80% case ascertainment in participating hospitals. Aggregated anonymised demographic and diagnostic data was collected for all haemoglobinopathy patients. This provided the core dataset for quantifying prevalence of sickle cell and thalassaemia and mapping local hospital workloads and service requirements. Thirteen percent of HbSS adult patients were taking hydroxycarbamide. The cohort of patients treated with hydroxycarbamide was evaluated. Sixty two of the 80 patients started on treatment were included. Follow-up was censored after 9 years, totalling 249 person-years of data with a median follow-up of three years (IQR, 1-6). Results showed that haematological benefits were maintained in the long-term with treatment, but evidence of long-term clinical effectiveness was less strong. This appeared to be due to the patterns of clinical management in everyday practice. Patients tend to be treated with modest doses of hydroxycarbamide due to intolerance or inability to attain or maintain maximum tolerated dose. For example maximum tolerated dose was the aim of treatment for 91% of patients but it was achieved for 65% of participants. Non- compliance with treatment and monitoring schedule was the main reason for non- attainment. Results suggest that it is sensible to strive for maximum tolerated dose to ensure therapy remains effective, but with more realistic expectations of the dose patients can attain and maintain. Doses in adult patients average 20mg/kg/day and 25mg/kg/day in children. Adult patients may be able to achieve a higher dose, if there was more stringent monitoring and improved management of non-compliance. The North West London HU Sub-Registry proved useful for measuring long-term effectiveness and tolerability of hydroxycarbamide. Routinely collected data was utilized for both clinical management and research purposes. The novelty lay in examination of the nuances of routine clinical practice. An electronic patient record was developed as a clinical management tool. It is the first study reporting long-term outcomes for UK sickle cell disease patients on hydroxycarbamide. Findings should help clinicians devise effective treatment protocols and strategies for managing patients commenced on this therapy. Interventions need to be targeted at increasing utilisation, patient adherence and persistence with treatment. The electronic patient record could be used to maximise treatment benefit and improve adherence. More effective involvement of the multidisciplinary team and primary care colleagues in patient education and management should improve usage. Patients and carers need up to date and easy to assimilate information to make informed decisions about treatment options. Maintaining a SCD registry is challenging. Models which operate as clinical information systems provide an incentive for participation. These enable active involvement of local care providers in registry management and the ability to keep and utilize their own data. Clinicians require accurate and current data for patient management and to enable them to benchmark their local outcomes against national outcomes and care standards.