<p>In an era of widespread automation—from grocery store self-checkout machines to self-driving cars—it is not outrageous to wonder: can teachers be automated? And more specifically, can automated computer teachers instruct students how to write? Automated computer programs have long been used in summative writing evaluation efforts, such as scoring standardized essay exams, ranking placement essays, or facilitating programmatic outcomes assessments. However, new claims about automated writing evaluation’s (AWE) formative educational potential mark a significant shift. My project questions the effectiveness of using AWE technology for formative educational efforts such as improving and teaching writing. Taken seriously, these efforts portend a future embrace of semi, or even fully, automated writing classes, an unprecedented development in writing pedagogy.</p><p>Supported by a summer-long grant from the Purdue Research Foundation, I conducted a small-<i>n </i>quasi-experiment to test claims by online college tutoring site Chegg.com that its EasyBib Plus AWE tool can improve both writing and writers. The experiment involved four college English instructors reading pairs of essays comprising one AWE-treated and untreated version per pair. Using a comparative judgment model, a rubric-free method of writing assessment based on Thurstone’s law, raters read and designated one of each pair “better.” Across four raters and 160 essays, I found that AWE-treated essays were designated better only 30% of the time (95% confidence interval: 20-40%), a statistically significant difference from the null hypothesis of 50%. The results suggest that Chegg’s EasyBib Plus tool offers no discernible improvement to student writing, and potentially even worsens it.</p><p>Finally, I analyze Chegg’s recent partnership with the Purdue Writing Lab and Online Writing Lab (OWL). The Purdue-Chegg partnership offers a useful test case for anticipating the effects of higher education’s embrace of automated educational technology going forward. Drawing on the history of writing assessment and the results of the experiment, I argue against using AWE for formative writing instruction. In an era of growing automation, I maintain that a human-centered pedagogy remains one of the most durable, important, effective, and transformative ingredients of a quality education.</p>

In this study, the transfer of learning criteria that can be implemented before, during and after a learning programme was investigated. The transfer of learning criteria was identified, after which the Station Management Learning Programme was evaluated to see whether transfer of learning criteria was used during the facilitation of the programme. The study population for the research was comprised of facilitators and station commanders, who facilitated and attended the Station Management Learning Programme in Gauteng as part of their development as Station Commanders. The study methodology involved qualitative and quantitative approaches to data collection, with questionnaires and one-on-one interviews. Descriptive statistics were produced and literature, questionnaires and interviews were examined to establish whether transfer of learning took place. The findings of the study reflect that various learning transfer strategies exist and that these can be used to transfer learning from the classroom to the work environment. The study concludes that a significant number of transfer of learning strategies are already implemented in the South African Police Service, in the presentation of the Station Management Learning Programme. / Adult Basic Education (ABET) / M. Ed. (Adult Education)

Analyzing the structure of a dataset is a challenging problem in high-dimensions as the volume of the space increases at an exponential rate and typically, data becomes sparse in this high-dimensional space. This poses a significant challenge to machine learning methods which rely on exploiting structures underlying data to make meaningful inferences. This dissertation proposes the <i>n</i>-TARP method as a building block for high-dimensional data analysis, in both supervised and unsupervised scenarios.<div><br></div><div>The basic element, <i>n</i>-TARP, consists of a random projection framework to transform high-dimensional data to one-dimensional data in a manner that yields point separations in the projected space. The point separation can be tuned to reflect classes in supervised scenarios and clusters in unsupervised scenarios. The <i>n</i>-TARP method finds linear separations in high-dimensional data. This basic unit can be used repeatedly to find a variety of structures. It can be arranged in a hierarchical structure like a tree, which increases the model complexity, flexibility and discriminating power. Feature space extensions combined with <i>n</i>-TARP can also be used to investigate non-linear separations in high-dimensional data.<br></div><div><br></div><div>The application of <i>n</i>-TARP to both supervised and unsupervised problems is investigated in this dissertation. In the supervised scenario, a sequence of <i>n</i>-TARP based classifiers with increasing complexity is considered. The point separations are measured by classification metrics like accuracy, Gini impurity or entropy. The performance of these classifiers on image classification tasks is studied. This study provides an interesting insight into the working of classification methods. The sequence of <i>n</i>-TARP classifiers yields benchmark curves that put in context the accuracy and complexity of other classification methods for a given dataset. The benchmark curves are parameterized by classification error and computational cost to define a benchmarking plane. This framework splits this plane into regions of "positive-gain" and "negative-gain" which provide context for the performance and effectiveness of other classification methods. The asymptotes of benchmark curves are shown to be optimal (i.e. at Bayes Error) in some cases (Theorem 2.5.2).<br></div><div><br></div><div>In the unsupervised scenario, the <i>n</i>-TARP method highlights the existence of many different clustering structures in a dataset. However, not all structures present are statistically meaningful. This issue is amplified when the dataset is small, as random events may yield sample sets that exhibit separations that are not present in the distribution of the data. Thus, statistical validation is an important step in data analysis, especially in high-dimensions. However, in order to statistically validate results, often an exponentially increasing number of data samples are required as the dimensions increase. The proposed <i>n</i>-TARP method circumvents this challenge by evaluating statistical significance in the one-dimensional space of data projections. The <i>n</i>-TARP framework also results in several different statistically valid instances of point separation into clusters, as opposed to a unique "best" separation, which leads to a distribution of clusters induced by the random projection process.<br></div><div><br></div><div>The distributions of clusters resulting from <i>n</i>-TARP are studied. This dissertation focuses on small sample high-dimensional problems. A large number of distinct clusters are found, which are statistically validated. The distribution of clusters is studied as the dimensionality of the problem evolves through the extension of the feature space using monomial terms of increasing degree in the original features, which corresponds to investigating non-linear point separations in the projection space.<br></div><div><br></div><div>A statistical framework is introduced to detect patterns of dependence between the clusters formed with the features (predictors) and a chosen outcome (response) in the data that is not used by the clustering method. This framework is designed to detect the existence of a relationship between the predictors and response. This framework can also serve as an alternative cluster validation tool.<br></div><div><br></div><div>The concepts and methods developed in this dissertation are applied to a real world data analysis problem in Engineering Education. Specifically, engineering students' Habits of Mind are analyzed. The data at hand is qualitative, in the form of text, equations and figures. To use the <i>n</i>-TARP based analysis method, the source data must be transformed into quantitative data (vectors). This is done by modeling it as a random process based on the theoretical framework defined by a rubric. Since the number of students is small, this problem falls into the small sample high-dimensions scenario. The <i>n</i>-TARP clustering method is used to find groups within this data in a statistically valid manner. The resulting clusters are analyzed in the context of education to determine what is represented by the identified clusters. The dependence of student performance indicators like the course grade on the clusters formed with <i>n</i>-TARP are studied in the pattern dependence framework, and the observed effect is statistically validated. The data obtained suggests the presence of a large variety of different patterns of Habits of Mind among students, many of which are associated with significant grade differences. In particular, the course grade is found to be dependent on at least two Habits of Mind: "computation and estimation" and "values and attitudes."<br></div>

Recent research has challenged the way in which personality and attitude constructs are measured. Alternatives have been offered as to how non-cognitive responses are modeled, the mode of delivery used when administrating such scales, and the impact of technology in measuring personality. Thus, the major purpose of the studies in this thesis concerns two interrelated issues of personality research, namely the way personality responses are best modeled, and the most optimal mode by which personality items are presented and associated modal issues. Three studies are presented. First, recent developments using an ideal point approach to scale construction are outlined, and an empirical study compares modeling personality items based on an ideal point approach (generalized graded unfolding model; GGUM) and a dominance approach (graded response model: GRM). Second, an extensive review of literature pertaining to the mode effect when transferring paper-and-pencil measures to screen was conducted, in addition to a review of the various types of computerized and innovative items and their associated psychometric information. Finally, nine innovative items were developed using various multimedia features (e.g., video, graphics, and audio) to ascertain the advantages of these methods to present items constructed to elicit response behavior underlying ideal point approaches, namely, typical response behavior. It was found that the dominance IRT model continued to produce superior model-data fit for most items, more attention needs to be placed on developing principles for constructing ideal point type items, the web-based version supplied 20% more construct information than the paper version, and innovative items seem to provide more data-model fit for students with lower personality attributes. While the innovative items may require more initial outlay in terms of time and development costs, they have the capacity to provide more information regarding test-takers’ personality levels, potentially using fewer items.

Recent research has challenged the way in which personality and attitude constructs are measured. Alternatives have been offered as to how non-cognitive responses are modeled, the mode of delivery used when administrating such scales, and the impact of technology in measuring personality. Thus, the major purpose of the studies in this thesis concerns two interrelated issues of personality research, namely the way personality responses are best modeled, and the most optimal mode by which personality items are presented and associated modal issues. Three studies are presented. First, recent developments using an ideal point approach to scale construction are outlined, and an empirical study compares modeling personality items based on an ideal point approach (generalized graded unfolding model; GGUM) and a dominance approach (graded response model: GRM). Second, an extensive review of literature pertaining to the mode effect when transferring paper-and-pencil measures to screen was conducted, in addition to a review of the various types of computerized and innovative items and their associated psychometric information. Finally, nine innovative items were developed using various multimedia features (e.g., video, graphics, and audio) to ascertain the advantages of these methods to present items constructed to elicit response behavior underlying ideal point approaches, namely, typical response behavior. It was found that the dominance IRT model continued to produce superior model-data fit for most items, more attention needs to be placed on developing principles for constructing ideal point type items, the web-based version supplied 20% more construct information than the paper version, and innovative items seem to provide more data-model fit for students with lower personality attributes. While the innovative items may require more initial outlay in terms of time and development costs, they have the capacity to provide more information regarding test-takers’ personality levels, potentially using fewer items.

Recent research has challenged the way in which personality and attitude constructs are measured. Alternatives have been offered as to how non-cognitive responses are modeled, the mode of delivery used when administrating such scales, and the impact of technology in measuring personality. Thus, the major purpose of the studies in this thesis concerns two interrelated issues of personality research, namely the way personality responses are best modeled, and the most optimal mode by which personality items are presented and associated modal issues. Three studies are presented. First, recent developments using an ideal point approach to scale construction are outlined, and an empirical study compares modeling personality items based on an ideal point approach (generalized graded unfolding model; GGUM) and a dominance approach (graded response model: GRM). Second, an extensive review of literature pertaining to the mode effect when transferring paper-and-pencil measures to screen was conducted, in addition to a review of the various types of computerized and innovative items and their associated psychometric information. Finally, nine innovative items were developed using various multimedia features (e.g., video, graphics, and audio) to ascertain the advantages of these methods to present items constructed to elicit response behavior underlying ideal point approaches, namely, typical response behavior. It was found that the dominance IRT model continued to produce superior model-data fit for most items, more attention needs to be placed on developing principles for constructing ideal point type items, the web-based version supplied 20% more construct information than the paper version, and innovative items seem to provide more data-model fit for students with lower personality attributes. While the innovative items may require more initial outlay in terms of time and development costs, they have the capacity to provide more information regarding test-takers’ personality levels, potentially using fewer items.

Recent research has challenged the way in which personality and attitude constructs are measured. Alternatives have been offered as to how non-cognitive responses are modeled, the mode of delivery used when administrating such scales, and the impact of technology in measuring personality. Thus, the major purpose of the studies in this thesis concerns two interrelated issues of personality research, namely the way personality responses are best modeled, and the most optimal mode by which personality items are presented and associated modal issues. Three studies are presented. First, recent developments using an ideal point approach to scale construction are outlined, and an empirical study compares modeling personality items based on an ideal point approach (generalized graded unfolding model; GGUM) and a dominance approach (graded response model: GRM). Second, an extensive review of literature pertaining to the mode effect when transferring paper-and-pencil measures to screen was conducted, in addition to a review of the various types of computerized and innovative items and their associated psychometric information. Finally, nine innovative items were developed using various multimedia features (e.g., video, graphics, and audio) to ascertain the advantages of these methods to present items constructed to elicit response behavior underlying ideal point approaches, namely, typical response behavior. It was found that the dominance IRT model continued to produce superior model-data fit for most items, more attention needs to be placed on developing principles for constructing ideal point type items, the web-based version supplied 20% more construct information than the paper version, and innovative items seem to provide more data-model fit for students with lower personality attributes. While the innovative items may require more initial outlay in terms of time and development costs, they have the capacity to provide more information regarding test-takers’ personality levels, potentially using fewer items.

In this study, the transfer of learning criteria that can be implemented before, during and after a learning programme was investigated. The transfer of learning criteria was identified, after which the Station Management Learning Programme was evaluated to see whether transfer of learning criteria was used during the facilitation of the programme. The study population for the research was comprised of facilitators and station commanders, who facilitated and attended the Station Management Learning Programme in Gauteng as part of their development as Station Commanders. The study methodology involved qualitative and quantitative approaches to data collection, with questionnaires and one-on-one interviews. Descriptive statistics were produced and literature, questionnaires and interviews were examined to establish whether transfer of learning took place. The findings of the study reflect that various learning transfer strategies exist and that these can be used to transfer learning from the classroom to the work environment. The study concludes that a significant number of transfer of learning strategies are already implemented in the South African Police Service, in the presentation of the Station Management Learning Programme. / Adult Basic Education (ABET) / M. Ed. (Adult Education)

<p> </p> <p>Traditional kinship systems involve the organization of individuals who are biologically connected. However, such systems have evolved beyond bloodlines to incorporate individuals that are biologically unassociated but operate in familial-like roles due to shared spaces and/or experiences. Historically, kinship systems or cultural networks have functioned as the cornerstone of survival for those of the Black lived experience. From the days of legalized human chattel slavery to present-day movements seeking justice for the minoritized, the foundation of kinship was typically built through the local church, the assumed maternal positions by Black women, Black secret societies and more. They each served, and continue to serve, as a means for survival and success against a systemically oppressive society. This study explores the notion and existence of kinship systems at historically Black colleges and universities (HBCUs). It specifically examines how fictive kinships through the lens of faculty-student dynamics, religion, and social activities, potentially influences the academic experience of Black students at HBCUs that currently have an above average retention rate. As America’s educational institution has lacked diversity, inclusion, justice, and equity for Black people for countless years, the primary mission of this study was to amplify Black student voices which have traditionally been suppressed. A supplemental goal of this study was to offer Black students tools for introspection that will aid them in navigating possible barriers to (post) educational success. In turn, this study gives insight to predominantly white institutions of higher learning on how to positively enhance the experience and retention of Black students, and the overall structure of diversity and inclusion on campus.</p>

<p>In this dissertation, I investigate the authentic learning experiences of students participating in the Engineering Projects in Community Service (EPICS) program at Purdue University within the framework of authentic education. Utilizing a quantitative approach, the study assesses the performance of new and returning students across five key outcomes that measure authentic learning during a single semester. The analysis of variance (ANOVA) revealed significant main effects for time of assessment and type of student on performance, with an overall improvement in all outcomes observed from mid-term to final evaluations and returning students typically outperforming new students. Interaction effects between time and type were also examined, revealing subtle yet complex dynamics in students’ learning trajectories. The findings hold implications for enhancing authentic learning, especially in engineering design contexts, and offer insights to guide future implementation of and improvements to authentic education initiatives, particularly the EPICS program. Despite certain limitations, the research opens avenues for future investigations into diverse aspects of authentic education in STEM and beyond.  </p>