Statistical analysis and meta-analysis of microarray data

The microarray technology provides a high-throughput technique to study gene expression. Microarrays can help us diagnose different types of cancers, understand biological processes, assess host responses to drugs and pathogens, find markers for specific diseases, and much more. Microarray experiments generate large amounts of data. Thus, effective data processing and analysis are critical for making reliable inferences from the data. ^ The first part of dissertation addresses the problem of finding an optimal set of genes (biomarkers) to classify a set of samples as diseased or normal. Three statistical gene selection methods (GS, GS-NR, and GS-PCA) were developed to identify a set of genes that best differentiate between samples. A comparative study on different classification tools was performed and the best combinations of gene selection and classifiers for multi-class cancer classification were identified. For most of the benchmarking cancer data sets, the gene selection method proposed in this dissertation, GS, outperformed other gene selection methods. The classifiers based on Random Forests, neural network ensembles, and K-nearest neighbor (KNN) showed consistently god performance. A striking commonality among these classifiers is that they all use a committee-based approach, suggesting that ensemble classification methods are superior. ^ The same biological problem may be studied at different research labs and/or performed using different lab protocols or samples. In such situations, it is important to combine results from these efforts. The second part of the dissertation addresses the problem of pooling the results from different independent experiments to obtain improved results. Four statistical pooling techniques (Fisher inverse chi-square method, Logit method. Stouffer's Z transform method, and Liptak-Stouffer weighted Z-method) were investigated in this dissertation. These pooling techniques were applied to the problem of identifying cell cycle-regulated genes in two different yeast species. As a result, improved sets of cell cycle-regulated genes were identified. The last part of dissertation explores the effectiveness of wavelet data transforms for the task of clustering. Discrete wavelet transforms, with an appropriate choice of wavelet bases, were shown to be effective in producing clusters that were biologically more meaningful. ^

Financial aid and the persistence of associate of arts graduates transferring to a senior university

This study examined the effects of financial aid on the persistence of associate of arts graduates transferring to a senior university in one of four consecutive fall semesters (1998-2001). Situated in an international metropolitan area in the southeastern United States, the institution where the study was conducted is a large public research university identified as a Hispanic Serving Institution. Archival databases served as the source of information on the academic and social background of the 4,669 participants in the study. Data from institutional financial aid records were pooled with the data in the student administrative system.^ For purposes of this study, persistence was defined as ongoing progress until completing the baccalaureate degree. Student social background variables used in the study were gender, ethnicity, age, and income, with GPA and part-time or full-time enrollment status being the academic variables. Amount and type of aid, including grants, loans, scholarships, and work study were incorporated in the models to determine the effect of financial aid on the persistence of these transfer students. Because the dependent variable persistence had three possible outcomes (graduated, still enrolled, dropped out) multinomial logistic regression was the appropriate technique for analyzing the data; four multinomial models were employed in the analysis.^ Findings suggest that grants awarded based on the financial need of students and loans were effective in encouraging the persistence of students, but scholarships and work study were not effective.^

Does hope matter? The influence of dispositional hope on persistence in a developmental writing course

The purpose of this study was to determine hope’s unique role, if any, in predicting persistence in a developmental writing course. Perceived academic self-efficacy was also included as a variable of interest for comparison because self-efficacy has been more widely studied than hope in terms of its non-cognitive role in predicting academic outcomes. A significant body of research indicates that self-efficacy influences academic motivation to persist and academic performance. Hope, however, is an emerging psychological construct in the study of non-cognitive factors that influence college outcomes and warrants further exploration in higher education. This study examined the predictive value of hope and self-efficacy on persistence in a developmental writing course. The research sample was obtained from a community college in the southeastern United States. Participants were 238 students enrolled in developmental writing courses during their first year of college. Participants were given a questionnaire that included measures for perceived academic self-efficacy and hope. The self-efficacy scale asked participants to self-report on their beliefs about how they cope with different academic tasks in order to be successful. The hope scale asked students to self-report on their beliefs about their capability to initiate action towards a goal (“agency”) and create a plan to attain these goals (“pathways”). This study utilized a correlational research design. A statistical association was estimated between hope and self-efficacy as well as the unique variance contributed by each on course persistence. Correlational analysis confirmed a significant relationship between hope and perceived academic self-efficacy, and a Fisher’s z-transformation confirmed a stronger relationship between the agency component of hope and perceived academic self-efficacy than for the pathways component. A series of multinomial logistic regression analyses were conducted to assess if (a) perceived self-efficacy and hope predict course persistence, (b) hope independent of self-efficacy predicts course persistence, and (c) if including the interaction of perceived self-efficacy and hope predicts course persistence. It was found that hope was only significant independent of self-efficacy. Some implications for future research are drawn for those who lead and coordinate academic support initiatives in student and academic affairs.

Glycemic control: Risk factors, quality of life, workforce participation, and mortality among US adults with type 2 diabetes

Despite research showing the benefits of glycemic control, it remains suboptimal among adults with diabetes in the United States. Possible reasons include unaddressed risk factors as well as lack of awareness of its immediate and long term consequences. The objectives of this study were to, using cross-sectional data, (1) ascertain the association between suboptimal (Hemoglobin A1c (HbA1c) .7%), borderline (HbA1c 7-8.9%), and poor (HbA1c .9%) glycemic control and potentially new risk factors (e.g. work characteristics), and (2) assess whether aspects of poor health and well-being such as poor health related quality of life (HRQOL), unemployment, and missed-work are associated with glycemic control; and (3) using prospective data, assess the relationship between mortality risk and glycemic control in US adults with type 2 diabetes. Data from the 1988-1994 and 1999-2004 National Health and Nutrition Examination Surveys were used. HbA1c values were used to create dichotomous glycemic control indicators. Binary logistic regression models were used to assess relationships between risk factors, employment status and glycemic control. Multinomial logistic regression analyses were conducted to assess relationships between glycemic control and HRQOL variables. Zero-inflated Poisson regression models were used to assess relationships between missed work days and glycemic control. Cox-proportional hazard models were used to assess effects of glycemic control on mortality risk. Using STATA software, analyses were weighted to account for complex survey design and non-response. Multivariable models adjusted for socio-demographics, body mass index, among other variables. Results revealed that being a farm worker and working over 40 hours/week were risk factors for suboptimal glycemic control. Having greater days of poor mental was associated with suboptimal, borderline, and poor glycemic control. Having greater days of inactivity was associated with poor glycemic control while having greater days of poor physical health was associated with borderline glycemic control. There were no statistically significant relationships between glycemic control, self-reported general health, employment, and missed work. Finally, having an HbA1c value less than 6.5% was protective against mortality. The findings suggest that work-related factors are important in a person’s ability to reach optimal diabetes management levels. Poor glycemic control appears to have significant detrimental effects on HRQOL.^

A methodology to estimate time varying user responses to travel time and travel time reliability in a road pricing environment

Road pricing has emerged as an effective means of managing road traffic demand while simultaneously raising additional revenues to transportation agencies. Research on the factors that govern travel decisions has shown that user preferences may be a function of the demographic characteristics of the individuals and the perceived trip attributes. However, it is not clear what are the actual trip attributes considered in the travel decision- making process, how these attributes are perceived by travelers, and how the set of trip attributes change as a function of the time of the day or from day to day. In this study, operational Intelligent Transportation Systems (ITS) archives are mined and the aggregated preferences for a priced system are extracted at a fine time aggregation level for an extended number of days. The resulting information is related to corresponding time-varying trip attributes such as travel time, travel time reliability, charged toll, and other parameters. The time-varying user preferences and trip attributes are linked together by means of a binary choice model (Logit) with a linear utility function on trip attributes. The trip attributes weights in the utility function are then dynamically estimated for each time of day by means of an adaptive, limited-memory discrete Kalman filter (ALMF). The relationship between traveler choices and travel time is assessed using different rules to capture the logic that best represents the traveler perception and the effect of the real-time information on the observed preferences. The impact of travel time reliability on traveler choices is investigated considering its multiple definitions. It can be concluded based on the results that using the ALMF algorithm allows a robust estimation of time-varying weights in the utility function at fine time aggregation levels. The high correlations among the trip attributes severely constrain the simultaneous estimation of their weights in the utility function. Despite the data limitations, it is found that, the ALMF algorithm can provide stable estimates of the choice parameters for some periods of the day. Finally, it is found that the daily variation of the user sensitivities for different periods of the day resembles a well-defined normal distribution.

Variables affecting the success of older applicants for employment in a public organization: An age -cued cognitive model of administrative decision-making

The dissertation takes a multivariate approach to answer the question of how applicant age, after controlling for other variables, affects employment success in a public organization. In addition to applicant age, there are five other categories of variables examined: organization/applicant variables describing the relationship of the applicant to the organization; organization/position variables describing the target position as it relates to the organization; episodic variables such as applicant age relative to the ages of competing applicants; economic variables relating to the salary needs of older applicants; and cognitive variables that may affect the decision maker's evaluation of the applicant. ^ An exploratory phase of research employs archival data from approximately 500 decisions made in the past three years to hire or promote applicants for positions in one public health administration organization. A logit regression model is employed to examine the probability that the variables modify the effect of applicant age on employment success. A confirmatory phase of the dissertation is a controlled experiment in which hiring decision makers from the same public organization perform a simulated hiring decision exercise to evaluate hypothetical applicants of similar qualifications but of different ages. The responses of the decision makers to a series of bipolar adjective scales add support to the cognitive component of the theoretical model of the hiring decision. A final section contains information gathered from interviews with key informants. ^ Applicant age has tended to have a curvilinear relationship with employment success. For some positions, the mean age of the applicants most likely to succeed varies with the values of the five groups of moderating variables. The research contributes not only to the practice of public personnel administration, but is useful in examining larger public policy issues associated with an aging workforce. ^

Treatment attrition among racial /ethnic minority adolescents participating in a brief motivational intervention for substance use problems: The influence of individual, social, cultural, and treatment factors

The present study has the primary aim of examining the predictors of treatment attrition among racial/ethnic minority adolescents with substance use problems. This study explores the potential differential influence of specific individual, social, cultural, and treatment factors on treatment attrition within three racial/ethnic subgroups of adolescents. Participants: A unique feature of the study is the use of a racial/ethnic minority sample (N=453), [U.S.-born Hispanics (n = 262), Foreign-born Hispanics (n = 117), and African-Americans (n = 74)]. Multivariate logit analyses were used to examine the influence of specific factors on treatment attrition among the full sample of adolescents, as well as within each racial/ethnic subgroup. Consistent with expectations, multivariate logit analyses reveal that, the specific factors associated with attrition varied across the three racial/ethnic subgroups. Having parents with problem substance use, being on the waitlist, and being court mandated to treatment emerged as predictors of attrition among the US-born Hispanics, while only Conduct Disorder was significantly associated with greater attrition among foreign-born Hispanics. Finally, among African-Americans, parental crack/cocaine use, parental abstinence from alcohol, and being on the waitlist were predictive of attrition. Multiple factors were associated with treatment attrition among racial/ethnic minority adolescents with specific factors differentially predicting attrition within each racial/ethnic subgroup. African-American youth were more than twice as likely as their Hispanic counterparts to leave treatment prematurely. It is critically important to understand the predictors of attrition among racial/ethnic minority youth in order to better meet the needs of youth most at risk of dropping out. ^

Treatment attrition among racial/ethnic minority adolescents participating in a brief motivational intervention for substance use problems : the influence of individual, social, cultural, and treatment factors

The present study has the primary aim of examining the predictors of treatment attrition among racial/ethnic minority adolescents with substance use problems. This study explores the potential differential influence of specific individual, social, cultural, and treatment factors on treatment attrition within three racial/ethnic subgroups of adolescents. Participants: A unique feature of the study is the use of a racial/ethnic minority sample (N=453), [U.S.-born Hispanics (n = 262), Foreign-born Hispanics (n = 117), and African- Americans (n = 74)]. Multivariate logit analyses were used to examine the influence of specific factors on treatment attrition among the full sample of adolescents, as well as within each racial/ethnic subgroup. Consistent with expectations, multivariate logit analyses reveal that, the specific factors associated with attrition varied across the three racial/ethnic subgroups. Having parents with problem substance use, being on the waitlist, and being court mandated to treatment emerged as predictors of attrition among the US-born Hispanics, while only Conduct Disorder was significantly associated with greater attrition among foreign-born Hispanics. Finally, among African-Americans, parental crack/cocaine use, parental abstinence from alcohol, and being on the waitlist were predictive of attrition. Multiple factors were associated with treatment attrition among racial/ethnic minority adolescents with specific factors differentially predicting attrition within each racial/ethnic subgroup. African-American youth were more than twice as likely as their Hispanic counterparts to leave treatment prematurely. It is critically important to understand the predictors of attrition among racial/ethnic minority youth in order to better meet the needs of youth most at risk of dropping out.

Access to better education: The school choice experience of families served by low-performing elementary public schools in Miami-Dade County

Public school choice education policy attempts to create an education marketplace. Although school choice research has focused on the parent role in the school choice process, little is known about parents served by low-performing schools. Following market theory, students attending low-performing schools should be the primary students attempting to use school choice policy to access high performing schools rather than moving to a better school. However, students remain in these low-performing schools. This study took place in Miami-Dade County, which offers a wide variety of school choice options through charter schools, magnet schools, and open-choice schools. ^ This dissertation utilized a mixed-methods design to examine the decision-making process and school choice options utilized by the parents of students served by low-performing elementary schools in Miami-Dade County. Twenty-two semi-structured interviews were conducted with the parents of students served by low-performing schools. Binary logistic regression models were fitted to the data to compare the demographic characteristics, academic achievement and distance from alternative schooling options between transfers and non-transfers. Multinomial logistic regression models were fitted to the data to evaluate how demographic characteristics, distance to transfer school, and transfer school grade influenced the type of school a transfer student chose. A geographic analysis was conducted to determine how many miles students lived from alternative schooling options and the miles transfer students lived away from their transfer school. ^ The findings of the interview data illustrated that parents’ perceived needs are not being adequately addressed by state policy and county programs. The statistical analysis found that students from higher socioeconomic social groups were not more likely to transfer than students from lower socioeconomic social groups. Additionally, students who did transfer were not likely to end up at a high achieving school. The findings of the binary logistic regression demonstrated that transfer students were significantly more likely to live near alternative school options.^

The effects of government disability benefits, accessibility laws and rehabilitation on employment choices by individuals with spinal cord injuries

Greater inclusion of individuals with disabilities into mainstream society is an important goal for society. One of the best ways to include individuals is to actively promote and encourage their participation in the labor force. Of all disabilities, it is feasible to assume that individual with spinal cord injuries can be among the most easily mainstreamed into the labor force. However, less that fifty percent of individuals with spinal cord injuries work. This study focuses on how disability benefit programs, such as Social Security Disability Insurance, and Worker's Compensation, the Americans with Disabilities Act and rehabilitation programs affect employment decisions. The questions were modeled using utility theory with an augmented expenditure function and indifference theory. Statically, Probit, Logit, predicted probability, and linear regressions were used to analyze these questions. Statistical analysis was done on the probability of working, ever attempting to work after injury, and on the number of years after injury that work was first attempted and the number of hours worked per week. The data utilized were from the National Spinal Cord Injury Database and the Spinal Cord Injuries and Labor Database. The Spinal Cord Injuries and Labor Database was created specifically for this study by the author. Receiving disability benefits decreased the probability of working, of ever attempting to work, increased the number of years after injury before the first work attempt was made, and decreased the number of hours worked per week for those individuals working. These results were all statistically significant. The Americans with Disabilities Act decrease the number of years before an individual made a work attempt. The decrease is statistically significant. The amount of rehabilitation had a significant positive effect for male individuals with low paraplegia, and significant negative effect for individuals with high tetraplegia. For women, there were significant negative effects for high tetraplegia and high paraplegia. This study finds that the financial disincentives of receiving benefits are the major determinants of whether an individual with a spinal cord injury returns to the labor force. Policies are recommended that would decrease the disincentive.

Does Hope Matter? The Influence of Dispositional Hope on Persistence in a Developmental Writing Course

The purpose of this study was to determine hope’s unique role, if any, in predicting persistence in a developmental writing course. Perceived academic self-efficacy was also included as a variable of interest for comparison because self-efficacy has been more widely studied than hope in terms of its non-cognitive role in predicting academic outcomes. A significant body of research indicates that self-efficacy influences academic motivation to persist and academic performance. Hope, however, is an emerging psychological construct in the study of non-cognitive factors that influence college outcomes and warrants further exploration in higher education. This study examined the predictive value of hope and self-efficacy on persistence in a developmental writing course. The research sample was obtained from a community college in the southeastern United States. Participants were 238 students enrolled in developmental writing courses during their first year of college. Participants were given a questionnaire that included measures for perceived academic self-efficacy and hope. The self-efficacy scale asked participants to self-report on their beliefs about how they cope with different academic tasks in order to be successful. The hope scale asked students to self-report on their beliefs about their capability to initiate action towards a goal (“agency”) and create a plan to attain these goals (“pathways”). This study utilized a correlational research design. A statistical association was estimated between hope and self-efficacy as well as the unique variance contributed by each on course persistence. Correlational analysis confirmed a significant relationship between hope and perceived academic self-efficacy, and a Fisher’s z-transformation confirmed a stronger relationship between the agency component of hope and perceived academic self-efficacy than for the pathways component. A series of multinomial logistic regression analyses were conducted to assess if (a) perceived self-efficacy and hope predict course persistence, (b) hope independent of self-efficacy predicts course persistence, and (c) if including the interaction of perceived self-efficacy and hope predicts course persistence. It was found that hope was only significant independent of self-efficacy. Some implications for future research are drawn for those who lead and coordinate academic support initiatives in student and academic affairs.

Glycemic Control: Risk Factors, Quality of Life, Workforce Participation, and Mortality Among US Adults with Type 2 Diabetes

Despite research showing the benefits of glycemic control, it remains suboptimal among adults with diabetes in the United States. Possible reasons include unaddressed risk factors as well as lack of awareness of its immediate and long term consequences. The objectives of this study were to, using cross-sectional data, 1) ascertain the association between suboptimal (Hemoglobin A1c (HbA1c) ≥7%), borderline (HbA1c 7-8.9%), and poor (HbA1c ≥9%) glycemic control and potentially new risk factors (e.g. work characteristics), and 2) assess whether aspects of poor health and well-being such as poor health related quality of life (HRQOL), unemployment, and missed-work are associated with glycemic control; and 3) using prospective data, assess the relationship between mortality risk and glycemic control in US adults with type 2 diabetes. Data from the 1988-1994 and 1999-2004 National Health and Nutrition Examination Surveys were used. HbA1c values were used to create dichotomous glycemic control indicators. Binary logistic regression models were used to assess relationships between risk factors, employment status and glycemic control. Multinomial logistic regression analyses were conducted to assess relationships between glycemic control and HRQOL variables. Zero-inflated Poisson regression models were used to assess relationships between missed work days and glycemic control. Cox-proportional hazard models were used to assess effects of glycemic control on mortality risk. Using STATA software, analyses were weighted to account for complex survey design and non-response. Multivariable models adjusted for socio-demographics, body mass index, among other variables. Results revealed that being a farm worker and working over 40 hours/week were risk factors for suboptimal glycemic control. Having greater days of poor mental was associated with suboptimal, borderline, and poor glycemic control. Having greater days of inactivity was associated with poor glycemic control while having greater days of poor physical health was associated with borderline glycemic control. There were no statistically significant relationships between glycemic control, self-reported general health, employment, and missed work. Finally, having an HbA1c value less than 6.5% was protective against mortality. The findings suggest that work-related factors are important in a person’s ability to reach optimal diabetes management levels. Poor glycemic control appears to have significant detrimental effects on HRQOL.

Linking labour force participation, education, labour wages and household consumption in developing countries:push and pull effects in self-employment - evidence from Nigeria

The contribution of this thesis is in understanding the origins in developing countries of differences in labour wage and household consumption vis-à-vis educational abilities (and by extension employment statuses). This thesis adds to the labour market literature in developing countries by investigating the nature of employment and its consequences for labour wage and household consumption in a developing country. It utilizes multinomial probit, blinder-oaxaca, Heckman and quantile regressions to examine one human capital indicator: educational attainment; and two welfare proxies: labour wage and household consumption, in a developing country, Nigeria. It finds that, empirically, the self-employed are a heterogeneous group of individuals made up of a few highly educated individuals, and a significant majority of ‘not so educated’ individuals who mostly earn less than paid workers. It also finds that a significant number of employers enjoy labour wage premiums; and having a higher proportion of employers in the household has a positive relationship with household consumption. The thesis furthermore discovers an upper educational threshold for women employers not found for men. Interestingly, the thesis also finds that there is indeed an ordering of labour wages into low-income self-employment (which seems to be found mainly in “own account” self-employment), medium-income paid employment, and high-income self-employment (which seems to be found mainly among employers), and that this corresponds to a similar ordering of low human capital, medium human capital and high human capital among labour market participants, as expressed through educational attainments. These show that as a whole, employers can largely be classed as experiencing pulled self-employment, as they appear to be advantaged in all three criteria (educational attainments, labour wage and household consumption). A minority of self-employed “own account” workers (specifically those at the upper end of the income distribution who are well educated), can also be classed as experiencing pulled self-employment. The rest of the significant majority of self-employed “own account” workers in this study can be classed as experiencing pushed self-employment in terms of the indicators used.

Measured parental weight status and familial socio-economic status correlates with childhood overweight and obesity at age 9

Background: Parental obesity is a predominant risk factor for childhood obesity. Family factors including socio-economic status (SES) play a role in determining parent weight. It is essential to unpick how shared family factors impact on child weight. This study aims to investigate the association between measured parent weight status, familial socio-economic factors and the risk of childhood obesity at age 9. Methodology/Principal Findings: Cross sectional analysis of the first wave (2008) of the Growing Up in Ireland (GUI) study. GUI is a nationally representative study of 9-year-old children (N = 8,568). Schools were selected from the national total (response rate 82%) and age eligible children (response rate 57%) were invited to participate. Children and their parents had height and weight measurements taken using standard methods. Data were reweighted to account for the sampling design. Childhood overweight and obesity prevalence were calculated using International Obesity Taskforce definitions. Multinomial logistic regression examined the association between parent weight status, indicators of SES and child weight. Overall, 25% of children were either overweight (19.3%) or obese (6.6%). Parental obesity was a significant predictor of child obesity. Of children with normal weight parents, 14.4% were overweight or obese whereas 46.2% of children with obese parents were overweight or obese. Maternal education and household class were more consistently associated with a child being in a higher body mass index category than household income. Adjusted regression indicated that female gender, one parent family type, lower maternal education, lower household class and a heavier parent weight status significantly increased the odds of childhood obesity. Conclusions/Significance: Parental weight appears to be the most influential factor driving the childhood obesity epidemic in Ireland and is an independent predictor of child obesity across SES groups. Due to the high prevalence of obesity in parents and children, population based interventions are required.

Advances in Dynamic Modeling and Computation for Count Data

A class of multi-process models is developed for collections of time indexed count data. Autocorrelation in counts is achieved with dynamic models for the natural parameter of the binomial distribution. In addition to modeling binomial time series, the framework includes dynamic models for multinomial and Poisson time series. Markov chain Monte Carlo (MCMC) and Po ́lya-Gamma data augmentation (Polson et al., 2013) are critical for fitting multi-process models of counts. To facilitate computation when the counts are high, a Gaussian approximation to the P ́olya- Gamma random variable is developed.

Three applied analyses are presented to explore the utility and versatility of the framework. The first analysis develops a model for complex dynamic behavior of themes in collections of text documents. Documents are modeled as a “bag of words”, and the multinomial distribution is used to characterize uncertainty in the vocabulary terms appearing in each document. State-space models for the natural parameters of the multinomial distribution induce autocorrelation in themes and their proportional representation in the corpus over time.

The second analysis develops a dynamic mixed membership model for Poisson counts. The model is applied to a collection of time series which record neuron level firing patterns in rhesus monkeys. The monkey is exposed to two sounds simultaneously, and Gaussian processes are used to smoothly model the time-varying rate at which the neuron’s firing pattern fluctuates between features associated with each sound in isolation.

The third analysis presents a switching dynamic generalized linear model for the time-varying home run totals of professional baseball players. The model endows each player with an age specific latent natural ability class and a performance enhancing drug (PED) use indicator. As players age, they randomly transition through a sequence of ability classes in a manner consistent with traditional aging patterns. When the performance of the player significantly deviates from the expected aging pattern, he is identified as a player whose performance is consistent with PED use.

All three models provide a mechanism for sharing information across related series locally in time. The models are fit with variations on the P ́olya-Gamma Gibbs sampler, MCMC convergence diagnostics are developed, and reproducible inference is emphasized throughout the dissertation.