Search

The structure of many complex social networks is determined by nodal and dyadic covariates that are endogenous to the tie variables. While exponential-family random graph models (ERGMs) have been very successful in modeling social networks with exogenous covariates, they are often misspecified for networks where some covariates are stochastic. Exponential-family random network models (ERNMs) are an extension of ERGM that retain the desirable properties of ERGM, but allow the joint modeling of tie variables and covariates. We compare ERGM to ERNM to show how conclusions of ERGM modeling are improved by consideration of the ERNM framework. In particular, ERNM simultaneously represents the effects of social influence and social selection processes while commonly used models do not. We also look into the latent class models for group clustering problems in social network. The random nodal attributes can be observed or latent in ERNM. When the attributes is treated as latent, we can investigate the probability of cluster group. Stochastic Block Models are a well-known statistical models for group classification, however, they rely on the relational data and omit the nodal characteristics. We compare these two models and provide a case study to illustrate the main points.

Logistic regression is a commonly used statistical technique in business and the sciences when an outcome is binary. For example, clinical trials may employ a logistic regression model when an outcome is presence or absence of disease, or a business may use such a model when the outcome is the presence or absence of a customer’s purchase of a product. An ideal logistic regression model both discriminates well and is well-calibrated. A well-calibrated model is one where the predicted percentages of success are close to the observed percentages. The Hosmer-Lemeshow test is a commonly used goodness of fit test that is used to test the calibration of a logistic regression model. The Hosmer-Lemeshow test becomes too powerful as the sample size increases, and an adaptive equation was recently proposed by Paul et al. (2013) to recommend the number of groups to use as the sample size increases. A new method to test the calibration of a logistic regression model, the calibration belt, was recently proposed by Nattino et al. (2014). The purpose of this study is to compare the power of the calibration belt with the Hosmer-Lemeshow test through simulations of several models with differing deviations from the true model and various probabilities of success. The Hosmer-Lemeshow test is applied to the models with varying number of groups (from g=6 to g= 5000), including the number of groups recommended through the adaptive equation proposed by Paul et al. (2013). The type 1 error rate of the calibration belt and the Hosmer-Lemeshow test is also assessed in all of these models. The simulations show that the calibration belt is nearly always the most powerful test, but the type 1 error rate of the calibration belt is often significantly below the nominal rate of 5%. The Hosmer-Lemeshow test does not suffer from this problem. It is also shown that the adaptive group equation proposed by Paul et al. (2013) depends largely on the probability of success of each of the models.

In our study, we have extended the Neyman smooth-type goodness of fit tests by Eubank (1997) from simple random sample to complex surveys (Methodologies have been provided for complex surveys, and theorems have been provided only for stratified random samples.) by incorporating consistent estimators under the survey design, which is accomplished by a data-driven nonparametric order selection method. Simulation results show that these proposed methods potentially improve the statistical power while controlling the type I error very well compared to those commonly used existing test procedures, especially for the cases with slow-varying probabilities. We also derived the large sample properties of the test statistics in stratified sampling.

An important aspect of empirical research is the construction of a model that represents the data. In psychological and educational measurement, models are typically evaluated regarding their ability to fit well to the observed data. Philosophers of science have long recognized that goodness-of-fit to the realized data is an insufficient metric of a model’s usefulness; models should also be appraised regarding their generalizability to unseen data. Frequentist statistics, Bayesian inference, and information theory seem to offer philosophically and methodologically dissimilar perspectives on model evaluation. However, this dissertation develops a simple theoretical framework that integrates these three perspectives. Within this framework, the information-theoretic principle of minimum description length is explored in the context of item response theory modeling. The findings reveal that complexity in item response theory is defined not by the number of freely estimated parameters in a model, but by its functional form. The frequentist, Bayesian, and information-theoretic approaches are then utilized in evaluating the usefulness of a unidimensional 3-parameter logistic model of item response data from the Program for International Student Assessment. Philosophical ramifications, future research directions, and implications for educational and psychological measurement are discussed.

The purpose of this phenomenological study was to understand the socialization (Weidman, Twale, & Stein, 2001) experiences of part-time doctoral students as a result of peer mentorship in one college. Part-time doctoral students are identified as students who are maintaining full-time employment or obligations outside of the university. The research questions for this study include: How do part-time doctoral students experience socialization through peer mentorship? How do part-time doctoral students experience spontaneous peer mentorship? How do part-time doctoral students experience peer mentorship as a result of their involvement in formal socialization experiences? How can the experiences of these part-time doctoral students be used to create implications for practice in the socialization by peer mentorship? The 21 participants in this study were all enrolled in a large university at the time of the study. A composite description including themes of collegiality and isolation in the participants’ experiences with peer mentorship was developed. Participants experienced collegiality through nurturing mentoring, educative mentoring, dyadic co-mentoring with peers, workplace colleagues and personal friends who provided efficacious experiences as they negotiated their scholarly identity. Isolation includes feelings of being second-tier doctoral students and life getting in the way combined with feelings of negative self-efficacy as they negotiated their scholarly identity. Interrelationships between the themes and subthemes are delineated. Implications for practice to encourage peer mentorship for part-time doctoral students were created. Suggestions are provided for both part-time doctoral students and their faculty in order to facilitate the process of socialization to the academy through peer mentorship.

The most commonly used method of determining calibration of logistic models is through the use of the Hosmer-Lemeshow goodness-of fit test (H-L test). However, a new method of calibration has been presented, known as the GiViTI Calibration Belt (GCB). Through the use of artificial and observational datasets from the GiViTI ICU study and from the Ohio State University Medical Center (OSUMC) patient readmission study, we present a comparison of the two methods of determining calibration. We have found that the GCB is most useful for identifying specific regions of miscalibration rather than an overall test of calibration. It is of particular importance in models where miscalibration is to be expected as it can identify regions of poor calibration due to insufficient numbers of observations. In models that are properly calibrated, the GCB matches the H-L test in concluding that there are no major regions of miscalibration regardless of the number of observations the model is based upon. The GCB shows promise as a tool for recalibrating models to adjust for changes in hospital practices that may have occurred after the model’s development.

Research has shown that effective school leadership has a positive influence on school effectiveness and student achievement. Current reform efforts include teachers, both formally and informally, as leaders of schools. However, there are currently no widely-accepted measurements or models to assess both formal and informal teacher leadership in schools. The purpose of this study was to compare model fit for the four-factor model of teacher leadership to model fit for three alternative models. The four-factor model was developed during the second administration of the Teacher Leadership Inventory (TLI), and the three alternative models were developed from the results and recommendations from the confirmatory factor analysis of that administration. Teacher responses to the second administration of the TLI constituted the data set for this study. Participants included 421 teachers from 23 schools in three East Tennessee school districts. Confirmatory factor analyses were conducted for each of the measurement models under investigation, and model fit indices and parameter estimates of all four models were used for comparison. Model fit indices indicate better model fit for the four-factor model over both the two-factor and five-factor models but not over the three-factor model. However, further evaluation of both parameter estimates and prior research provide support for the acceptability of the four-factor model over the three-factor model.

Abstract: Exponential Random Graph Models (ERGMs) have been developed for fitting social network data on both static and dynamic levels. However, the lack of large sample asymptotic properties makes it inadequate in assessing the goodness-of-fit of these ERGMs. Simulation-based goodness-of-fit plots were proposed by Hunter et al (2006), comparing the structured statistics of observed network with those of corresponding simulated networks. In this research, we propose an improved approach to assess the goodness of fit of ERGMs. Our method is shown to improve the existing graphical techniques. We also propose a simulation based test statistic with which the model comparison can be easily achieved.

Biological invasions are major threats to any nation’s economy and biodiversity. To detect new biological incursions of some species biosecurity agencies deploy pheromone sentinel traps for targeted species at high risk sites such as airports, seaports and transitional facilities. A good example is the gypsy moth surveillance program in New Zealand. Following the detection of an incursion by an unwanted organism, ground-based searches to locate the source can be very expensive, but are essential to identify the introduction pathway and to increase the chances of success eradicating the unwanted organism. In such circumstances, the possibility of better targeting the search for the source of the incursion using a modelling approach is worthy of investigation A stochastic mechanistic model to hindcast moth flight from a recapture location to the release location was developed based on insect behaviour in response to wind and pheromones. The model was composed of two main processes, 1) downwind dispersal, assumed to result from an appetitive behaviour, indicated by an analysis of a previous mark-release-recapture experiment on painted apple moth (Teia anartoides, Walker) and, 2) anemotaxic dispersal inspired by pheromone anemotaxis theory but up-scaled from a fine-scaled behaviour model to a 2 m scale. A genetic algorithm was used to fit some model parameters. A specialised fitness function was developed to allow the genetic algorithm to identify parameters that resulted in models that reflected both the spread and density patterns in the trapping data. The resulting function allowed the stochastic model results to be compared with the inherently stochastic trapping data. The resulting individual based model simulates the spatio-temporal dispersal pattern of painted apple moth recorded during a previous mark-release-recapture experiment. While the proposed model is shown to have limitations with respect to accuracy and precision it is also demonstrated to greatly improve biosecurity incursion response capability, by more efficient targeting of search effort for the proximal source of an incursion.

The purpose of this dissertation is to understand the managerial "action" of public administrators in the management of their organizations within the Brazilian context. The research seeks to understand the relationships between managers and formal management mechanisms by exploring the complementary nature of the effective managerial action in the face of structural deficiencies and flaws, considering the possibility of overcoming the structuralism-subjectivism dichotomy present in the construction of the Theory of Organizations. Initially, the study provides a review of the literature on organizational design. It highlights the "goodness of fit" proposition on strategic choice issues concerning the main organizational variables design and organizational goal attainment. It also calls special attention to the emerging interest of designing theorists on interpretivist approaches to the matter, such that of Karl Weick. A review of the the administrative reforms in Brazil is made from the perspective of the main stream organizational design conceptual framework. It highlights the complex dynamics of a constant search for differentiation and flexibilization subject to patterns of advances and reversals, due to the centrality, strength and pervasiveness of the bureaucratic model. It is concluded that in no single given moment, a public manager and his team, may count on a formal organizational design which attends the"congruency" criteria, devised by organizational design conceptual frameworks, to explain organizational results in different environmental sets. Although this conclusion may explain failure at the public sector, it can not provide understanding on the many instances of significative success attained by government operations in spite of inadequate formal administrative structures. This point calls for a better understanding from the interpretivist approach, on how public administrators, strongly associated with good organizational results, engage into transformative action, in order to superate administrative structures flaws and dysfunctional cultural patterns of conduct, structurally present and constantly reproduced, in vigorous developing countries, such as Brazil. The dissertation transcribes the testimony of four outstanding public administrators, doing a deep incursion in the managerial real world of public administration, as subjectively defined by them and transformed by their engagement into action.Through the thematic version of the Oral History methodology, full segments of the complete interviews are categorized into the thirty two managerial strategies captured which are presented on a recategorized manner under eight main strategies: (1) Interchanging Frames of Reference; (2) Exploring the Formal Limits; (3) Playing the Bureaucracy Game; (4) Inducing the Inclusion of Others (5)Promoting Internal Cohesion; (6) Creating Shields against Transgressions; (7) Overcoming Internal Restrictions; (8) Letting the Structures Blossom. Each one of these eight blocks of strategies presented, deserves further reflexive interpretation by the author, on the light of the interpretivist approach to organizational design. A final effort is made, now on theory building, for improving understanding on the matter. In order to find a significant meaning underlining all the strategies extracted from the "practical consciousness" of the interviewers as revealed in their report, the author resort to a metaphor. This metaphor helps to: (1) better describe and understand a not adequately treated phenomenon, namely, good results under inadequate structural social and organizational conditions; (2) reveal the logic and the meaning underlining all the strategies adopted to generate results under these unfaithful conditions; (3) name, accordingly to the nature of the managerial transformative social action involved, an open ended class of managerial interventions of a pragmatic sort driven by an ethics of results much common to good managers, that is, the concept of "managerial equalization"; and (4) give back to public administrators, represented by the interviewees, to be incorporated in their "discursive consciousness", something the most effective and experienced public managers already have as tacit knowledge built in their "practical consciousness", and so, help the education and development of new talents.

The publisher of the Degrees of Reading Power test of reading comprehension (DRP) calibrate their test using an item response model called the Rasch or one-parameter logistic model. The relationship between the use of the Rasch model in calibration of the DRP and the use of the DRP as a component of the Virginia Literacy Passport Testing Program (LPT) is addressed. Analyses concentrate on sixth grade students who were administered the DRP in 1991. The question that arises is whether the Rasch model is the appropriate model to use to calibrate the DRP in this high-stakes setting. The majority of research that has been reported by the publisher of the DRP to assess the adequacy of the Rasch model have not included direct checks on model assumptions, model features or model predictions. Instead, they have relied almost exclusively on statistical tests in assessment of model fit. This study will assess the adequacy of fitting DRP test data to the Rasch model through direct examination of the assumptions, features and predictions of the IRT model. This is accomplished by comparing the Rasch model to the less restrictive two- and three-parameter logistic models. Robust IRT-based goodness-of-fit techniques are conducted. When the DRP is used in a high stakes setting, guessing is likely for those in jeopardy of failing. Under these circumstances, we must attend to the possibility that guessing may be a factor and thereby calibrate the DRP with the three-parameter model, as this model takes guessing into account.