Cluster robust models are a kind of statistical models that attempt to estimate parameters considering potential heterogeneity in treatment effects. Absent heterogeneity in treatment effects, the partial and average treatment effect are the same. When heterogeneity in treatment effects occurs, the average treatment effect is a function of the various partial treatment effects and the composition of the population of interest. The first chapter explores the performance of common estimators as a function of the presence of heterogeneity in treatment effects and other characteristics that may influence their performance for estimating average treatment effects. The second chapter examines various approaches to evaluating and improving cluster structures as a way to obtain cluster-robust models. Both chapters are intended to be useful to practitioners as a how-to guide to examine and think about their applications and relevant factors. Empirical examples are provided to illustrate theoretical results, showcase potential tools, and communicate a suggested thought process. The third chapter relates to an open-source statistical software package for the Julia language. The content includes a description for the software functionality and technical elements. In addition, it features a critique and suggestions for statistical software development and the Julia ecosystem. These comments come from my experience throughout the development process of the package and related activities as an open-source and professional software developer. One goal of the paper is to make econometrics more accessible not only through accessibility to functionality, but understanding of the code, mathematics, and transparency in implementations.

To mitigate the loss of experienced administrators, community college leaders create internal grow-your-own (GYO) leadership development programs. Although the purpose of these programs is to ensure institutional efficacy by developing a pipeline of administrative and faculty leaders, little attention has been given to learning about the participants' post-program career advancement over time. Consequently, the value of GYO leadership training for participants is not well known. The purpose of this qualitative case study was to explore the post-program professional development of a group of participants 2 years after completing a rural community college GYO leadership program. The conceptual framework was guided by both the American Association of Community Colleges' Leadership competency model and situated learning theory. Research questions examined the leadership knowledge, skills, and behaviors the participants acquired from their training and how these knowledge, skills, and abilities were being used. Data sources included interviews of a cohort of 10 GYO leadership participants, leadership institute artifacts, and human resources documents, which were used for triangulation of the data. A constant comparative analysis methodology was used to identify themes. Research findings indicated that the participants most often applied their communication, collaboration, and advocacy knowledge and skills in their respective jobs, interests, and professional development over time. Academic and other leaders can use the research results to create positive social change by offering GYO leadership training programs to promote the advancement of the institution's mission, thereby improving the lives of the students and the health of the community.

One rural Midwestern high school discovered a discrepancy among school, state, and national science skill attainment, verified by ACT scores. If students do not acquire vital science skills, they may not perform proficiently on science tests, thus impacting future college options. Inquiry based instruction and constructivism provided the basis for the theoretical framework. This study questioned associations between ACT scores, inquiry science technique usage, and ACT standard usage (Phase 1), and teachers' views on science instruction (Phase 2). This sequential explanatory mixed methods program evaluation included 469 ACT scores, surveys sent to 9 science teachers, and 8 interviews. Phase 1 used the inquiry science implementation scale survey and an ACT college readiness standards workbook to determine proportional associations between datasets. Descriptive statistics, one-sample t tests, and binomial tests were used to analyze Phase 1 data. Phase 2 interviews augmented Phase 1 data and were disassembled, reassembled, and interpreted for parallel viewpoints. Phase 1 data indicated that teachers use a slightly above average amount of inquiry and science ACT standards in the classroom; however, most science students did not test above the curriculum and there were inconsistencies in standards covered. Phase 2 data revealed teachers need time to collaborate and become skilled in inquiry methods to rectify the inconsistencies. The project was an evaluation report. This study will foster positive social change by giving the district a plan: adapt the science curriculum by integrating more ACT and inquiry standards and participate in more professional development that applies inquiry as a tool to increase science skill proficiency, thus generating locally competitive students for college and the workforce.

In 2014, the Federal Aviation Administration (FAA) highlighted to the U.S. Senate the need to focus on air traffic control (ATC) training to meet job qualification and attrition rates within the career field. One U.S. Department of Defense military service assists the FAA in providing worldwide ATC services. This service is referred to as the agency throughout this paper to ensure confidentiality. The agency's ATC career field manager echoed the FAA's call for action in his 2014 Strategic/Action Plan. In August 2013, the agency's ATC trainer program was published. As of December 2015, the program had not been evaluated. The purpose of this study was to ascertain if the program facilitated the learning of critical ATC on-the-job training skills. An ad hoc expertise-oriented evaluation was conducted using the lenses of andragogy, experiential learning, and instructional system design (ISD). Purposeful sampling procedures were used to select 20 participants across the subgroups of supervisors, trainers, managers, and training developers from 7 focus sites. The semi-structured interviews queried 4 topical areas derived from Kirkpatrick's 4 levels of evaluation model. Data collected via documents and interviews were analyzed using descriptive, emotion, eclectic, and pattern coding. Key findings indicated that the program was not developed compliant with ISD principles and did not promote adult learning as endorsed by andragogy and experiential learning theory. The implications for positive social change include providing stakeholders with data needed to make evidence-based decisions regarding the current and future state of the program. The evaluation report project can be shared with the FAA, an agency partner, and has the potential to create a platform for improved training practices focusing on optimum and successful adult learning transactions.

Evidence derived from the 2012 and 2015 College Senior Surveys (CSS) noted showed that college seniors, at a historically Black university, graduated with little to average soft skills. Soft skills, such as personal characteristics and relations with others, are needed for students to succeed in postgraduate careers. The purpose of this study was to assess the level to which performing arts instruction (PAI) courses developed college-level students' soft skills. Kolb's experiential learning theory, which defines the learning process as knowledge and skills developed through experiences, and Stufflebeam's evaluation model, which uses context, input, process, and product, were used to guide this study. A case study design was used to discern students' perceptions of PAI to help develop their soft skills and meet employers' expectations. Maximum variation sampling was used to select 15 participants who met the criteria of being a senior performing arts student at the target site. All 15 participants were interviewed. In addition, the collected data were coded, organized into themes, and then I triangulated the participants' responses with the CSS summary report. Findings indicated that while PAI helped students meet employers' hiring expectations in areas of soft skills, it was also revealed that there is a need for soft skills development to be embedded in other programs of study at the target site. Both a 3-day student and a 1-day faculty professional development) session were developed to instruct both groups on the use of soft skills. Implications for positive social change are that a campus-wide model to improve students' soft skills across all academic disciplines may result in improved employment opportunities, thus contributing to the global economy.

近年來，由於ICT的蓬勃發展，政府逐漸增加投資於電子化計畫的線上服務比例與種類，為了理解投資所能獲得的收益，杜絕資源浪費，並促使政府調整施政方向達到良善治理，針對電子化計畫進行評估是勢然必行的。目前許多評估機構逐漸從衡量政府投資電子化政府服務的客觀產出，轉而討論電子化服務使用者端的影響與成果。本研究，以計畫整體流程中所面臨的內部運作與外部感受、客觀產出與抽象成果的評估構面，提出G2B電子化計畫流程評估架構。 本研究結合Heeks (2006)的價值鏈模型與Millard & Shahin (2006)電子化計畫評估流程建立流程模型，透過文獻整理出評比國家型、網頁感知品質型、滿意模型與利害關係人模型指標，從企業途徑與政府內部途徑角度，探討G2B計畫流程評估架構，並使用層級分析法(analytic hierarchy process)，由專家給予衡量G2B計畫推動時，前端到後端階段中重要構面的權重，並根據分析結果討論現今台灣計劃評估傾向上的差異。 本研究結果發現，在政府計畫生產過程中，應重視政府領導與支持，並發展衡量方式；專家不論在企業途徑與政府內部途徑上，都認為「服務產出評估是必需的過程，但影響與成果階段，才是價值衡量的終點」，因此，電子化計劃評估應從重視成本比的角度，朝向討論「透過政府對計畫服務的投入與投資，所能帶給企業實際或感知的價值與影響」，更能切中利害關係人關注的關鍵利益。 / This study aims at building an e-program evaluation chain model combining output evaluation with impact assessments, in order to assess the comprehensive G2B e-program from back-office process to the impact of online services. This study contributes a G2B evaluation framework with six processes integrated by e-government value chain model and e-program evaluation process. Given the concept of external and internal customer,「Business approach」and「Internal-Government approach」is developed in view of stakeholders. By evaluating subjective expert judgments via analytic hierarchy process method, this G2B evaluation framework is given weights and priorities in each hierarchy under both of approaches. The results indicate that impact and outcome assessments should be emphasized, no matter which approach is. Evaluation toward objective output is only a necessary process, but not the destination. As policy suggestions, this study addresses that government should focus on what kinds of value e-program can bring, and evaluate them.

Background Health education programs that are aimed at improving individuals' skills to self-manage are increasingly recognised as a critical component of chronic disease management. Despite the apparent need for such interventions, current studies show inconsistent results regarding program effectiveness, with meta-analyses indicating only marginal effects for some disease groups. A closer examination of these studies however suggests that the magnitude and inconsistency of the findings may be related to the types of outcomes that were assessed rather than specific disease groups. Where self-report measures were used, results tended to be smaller and inconsistent. It is therefore possible that current studies do not adequately reflect program effects because self-report outcomes have a high risk to be confounded by a range of potential biases. Objective The aim of this thesis was to identify and quantify the potential influence of biases in the measurement of change in chronic disease self-management interventions using self-report. Methods The research design targeted the processes that individuals undergo when filling out questionnaires and whether this has an influence on their self-report outcomes. This was achieved by developing a three-group research design. The Health Education Impact Questionnaire (heiQ) was used to collect outcomes data. While pretest questionnaires were identical across groups, three questionnaire versions were randomly distributed at posttest. One of the groups filled out traditional posttest questions (n=331), whereas the other two groups were asked to provide data in addition to posttest questions, with one group providing transition questions (n=304) and one providing retrospective pretest data (n=314). Resulting datasets were further examined for possible confounding effects through response shift and social desirability bias. Through the random allocation of the heiQs it was ensured that data were not influenced by potential intra-group effects. Results The thesis revealed that the design of the posttest questionnaire significantly influenced people's ratings of their posttest levels. In particular, when participants were asked to provide ratings of their retrospective pretest levels in addition to their posttest levels, the latter scores were significantly higher than those of participants who did not perform this additional task. Subsequent analyses however suggested that these differences could neither be explained by response shift nor by social desirability bias. Conclusions This research has provided important insight into the measurement of outcomes of chronic disease self-management interventions. While the threat to the validity of traditional pretest-posttest data due to confounding effects through response shift and social desirability biases could not be supported, the thesis has highlighted that the cognitive task that subjects are asked to perform when providing data at posttest significantly influenced their self-reported outcomes. Given that previous research has predominantly focused on other aspects of validity - such as applying control group designs to circumvent common threats to internal and external validity - this study suggests that more attention must be paid to the design of questionnaires. The thesis concludes that further research, in particular into the influence of cognitive tasks on obtained scores, is important to improve the interpretation of self-report outcomes data derived from participants of self-management interventions.

<p>Gifted elementary students in a congregated educational program (n = 165) were compared to gifted peers in regular programming (n = 49) in an urban, Western-Canadian, public, school division. Mean scores on measures of self-concept (Multidimensional Self Concept Scale), classroom environment (Classroom Environment Scale), and student life satisfaction (Multidimensional Student Life Satisfaction Scale) were analyzed. MANOVAs revealed main effects of educational programming and no mediating effects of gender or grade level. Students in the congregated program had lower academic self-concept than students in regular programming, replicating the commonly found Big-Fish-Little-Pond Effect (Marsh, 1987). There was also some weak indication that students in the specialized program had lower satisfaction with "self" than those in the regular program. In contrast, students in the specialized program thought their programming to be more innovative relative to how the other group perceived theirs was. However, all differences were of small-to-moderate magnitude (.5 SDs). Moreover, all scores for all measures were at, or slightly above, levels typically found in normally developing peers. </p>

<p>Gifted elementary students in a congregated educational program (n = 165) were compared to gifted peers in regular programming (n = 49) in an urban, Western-Canadian, public, school division. Mean scores on measures of self-concept (Multidimensional Self Concept Scale), classroom environment (Classroom Environment Scale), and student life satisfaction (Multidimensional Student Life Satisfaction Scale) were analyzed. MANOVAs revealed main effects of educational programming and no mediating effects of gender or grade level. Students in the congregated program had lower academic self-concept than students in regular programming, replicating the commonly found Big-Fish-Little-Pond Effect (Marsh, 1987). There was also some weak indication that students in the specialized program had lower satisfaction with "self" than those in the regular program. In contrast, students in the specialized program thought their programming to be more innovative relative to how the other group perceived theirs was. However, all differences were of small-to-moderate magnitude (.5 SDs). Moreover, all scores for all measures were at, or slightly above, levels typically found in normally developing peers. </p>

When confronted with decisions involving the provision of environmental resources, such as clean air, do individuals act cooperatively with societal regulations? This study employs a quasi-experimental design to investigate the cooperative and non-cooperative actions of high-emitting vehicle owners that arose in response to the Inspection and Maintenance (IM) program in the Atlanta airshed. The impact of cooperative and non-cooperative actions of high-emitters on vehicular tail-pipe emissions, such as carbon monoxide (CO), hydrocarbons (HC) and oxides of nitrogen (NOx), is quantified. Furthermore, the hypothesis that high-emitters have significantly higher odds of residing in lower income neighborhoods as compared to the normal emitters is also tested. A remote sensing sample of approximately 1.42 million vehicles observed on-road between 1997 and 2001 is matched with IM program data and vehicle registration data to identify the cooperative and non-cooperative high-emitters. A mixed-pool time-series regression analysis is undertaken to estimate changes in vehicular tail-pipe emissions due to the actions of high-emitters. The information about the socio-economic and demographic contextual conditions of the sampled vehicle owners is ascertained from 2000 census data. Approximately 42% of the high-emitting vehicle owners are found to be cooperative and 58% non-cooperative. The cooperative actions caused a decrease of 47% in HC emission factors during 1997 and 2001. There is no statistical difference between the CO and NO emission factors of vehicles owned by cooperative and non-cooperative high-emitters. Results also suggest that the high-emitters live in 4.4% lower median household income areas as compared to the normal emitters in the Atlanta airshed. Changes in the current IM program rules and vehicle registration laws could improve air quality. Such changes include disallowing IM test failures from registering anywhere in the state of Georgia and requiring an IM test on every change of vehicle ownership inside the IM program area. Better incentive mechanisms for high-emitters can also be designed. The evidence from this study is expected to aid policy-makers to adapt the incentive mechanisms of IM programs, in particular, and environmental regulations, in general, so that public policies are both more effective and equitable in their societal impacts.