Temporal data are increasingly prevalent and important in analytics. Time series (TS) data are chronological sequences of observations and an important class of temporal data. Fields such as medicine, finance, learning science and multimedia naturally generate TS data. Each series provide a high-dimensional data vector that challenges the learning of the relevant patterns This dissertation proposes TS representations and methods for supervised TS analysis. The approaches combine new representations that handle translations and dilations of patterns with bag-of-features strategies and tree-based ensemble learning. This provides flexibility in handling time-warped patterns in a computationally efficient way. The ensemble learners provide a classification framework that can handle high-dimensional feature spaces, multiple classes and interaction between features. The proposed representations are useful for classification and interpretation of the TS data of varying complexity. The first contribution handles the problem of time warping with a feature-based approach. An interval selection and local feature extraction strategy is proposed to learn a bag-of-features representation. This is distinctly different from common similarity-based time warping. This allows for additional features (such as pattern location) to be easily integrated into the models. The learners have the capability to account for the temporal information through the recursive partitioning method. The second contribution focuses on the comprehensibility of the models. A new representation is integrated with local feature importance measures from tree-based ensembles, to diagnose and interpret time intervals that are important to the model. Multivariate time series (MTS) are especially challenging because the input consists of a collection of TS and both features within TS and interactions between TS can be important to models. Another contribution uses a different representation to produce computationally efficient strategies that learn a symbolic representation for MTS. Relationships between the multiple TS, nominal and missing values are handled with tree-based learners. Applications such as speech recognition, medical diagnosis and gesture recognition are used to illustrate the methods. Experimental results show that the TS representations and methods provide better results than competitive methods on a comprehensive collection of benchmark datasets. Moreover, the proposed approaches naturally provide solutions to similarity analysis, predictive pattern discovery and feature selection.

The current study is a follow up to a previous evaluation of Believe It!, an internet-based career development program for adolescent girls. This study attempted to extend the program's effectiveness by manipulating animated agent appearance based on literature suggesting that agent appearance has implications for human-computer program interface. Participants included 52 Latinas (ages 11 to 14) randomly assigned to view one of two versions of the revised career program. Each version contained identical content but included animated agents designed to represent different ethnicities. Pre and post-treatment scores for three career belief measures and an occupational stereotype measure were analyzed using a MANCOVA. The results were not significant and further analyses revealed that the results were confounded by complications with the perceived ethnicity of the animated agents. Despite a lack of significance the results provide enriching information about Latina adolescent perception of ethnicity.

Lots of previous studies have analyzed human tutoring at great depths and have shown expert human tutors to produce effect sizes, which is twice of that produced by an intelligent tutoring system (ITS). However, there has been no consensus on which factor makes them so effective. It is important to know this, so that same phenomena can be replicated in an ITS in order to achieve the same level of proficiency as expert human tutors. Also, to the best of my knowledge no one has looked at student reactions when they are working with a computer based tutor. The answers to both these questions are needed in order to build a highly effective computer-based tutor. My research focuses on the second question. In the first phase of my thesis, I analyzed the behavior of students when they were working with a step-based tutor Andes, using verbal-protocol analysis. The accomplishment of doing this was that I got to know of some ways in which students use a step-based tutor which can pave way for the creation of more effective computer-based tutors. I found from the first phase of the research that students often keep trying to fix errors by guessing repeatedly instead of asking for help by clicking the hint button. This phenomenon is known as hint refusal. Surprisingly, a large portion of the student's foundering was due to hint refusal. The hypothesis tested in the second phase of the research is that hint refusal can be significantly reduced and learning can be significantly increased if Andes uses more unsolicited hints and meta hints. An unsolicited hint is a hint that is given without the student asking for one. A meta-hint is like an unsolicited hint in that it is given without the student asking for it, but it just prompts the student to click on the hint button. Two versions of Andes were compared: the original version and a new version that gave more unsolicited and meta-hints. During a two-hour experiment, there were large, statistically reliable differences in several performance measures suggesting that the new policy was more effective.

The current study analyzed existing data, collected under a previous U.S. Department of Education Reading First grant, to investigate the strength of the relationship between scores on the first- through third-grade Dynamic Indicators of Basic Early Literacy Skills - Oral Reading Fluency (DIBELS-ORF) test and scores on a reading comprehension test (TerraNova-Reading) administered at the conclusion of second- and third-grade. Participants were sixty-five English Language Learners (ELLs) learning to read in a school district adjacent to the U.S.-Mexico border. DIBELS-ORF and TerraNova-Reading scores were provided by the school district, which administers the assessments in accordance with state and federal mandates to monitor early literacy skill development. Bivariate correlation results indicate moderate-to-strong positive correlations between DIBELS-ORF scores and TerraNova-Reading performance that strengthened between grades one and three. Results suggest that the concurrent relationship between oral reading fluency scores and performance on standardized and high-stakes measures of reading comprehension may be different among ELLs as compared to non-ELLs during first- and second-grade. However, by third-grade the correlations approximate those reported in previous non-ELL studies. This study also examined whether the Peabody Picture Vocabulary Test (PPVT), a receptive vocabulary measure, could explain any additional variance on second- and third-grade TerraNova-Reading performance beyond that explained by the DIBELS-ORF. The PPVT was individually administered by researchers collecting data under a Reading First research grant prior to the current study. Receptive vocabulary was found to be a strong predictor of reading comprehension among ELLs, and largely overshadowed the predictive ability of the DIBELS-ORF during first-grade. Results suggest that receptive vocabulary scores, used in conjunction with the DIBELS-ORF, may be useful for identifying beginning ELL readers who are at risk for third-grade reading failure as early as first-grade.

The purpose of this study was to investigate the impacts of visual cues and different types of self-explanation prompts on learning, cognitive load and intrinsic motivation, as well as the potential interaction between the two factors in a multimedia environment that was designed to deliver a computer-based lesson about the human cardiovascular system. A total of 126 college students were randomly assigned in equal numbers (N = 21) to one of the six experimental conditions in a 2 X 3 factorial design with visual cueing (visual cues vs. no cues) and type of self-explanation prompts (prediction prompts vs. reflection prompts vs. no prompts) as the between-subjects factors. They completed a pretest, subjective cognitive load questions, intrinsic motivation questions, and a posttest during the course of the experience. A subsample (49 out of 126) of the participants' eye movements were tracked by an eye tracker. The results revealed that (a) participants presented with visually cued animations had significantly higher learning outcome scores than their peers who viewed uncued animations; and (b) cognitive load and intrinsic motivation had different impacts on learning in multimedia due to the moderation effect of visual cueing. There were no other significant findings in terms of learning outcomes, cognitive load, intrinsic motivation, and eye movements. Limitations, implications and future directions are discussed within the framework of cognitive load theory, cognitive theory of multimedia learning and cognitive-affective theory of learning with media.

This study examines the effect of the translation of traditional scientific vocabulary into plain English, a process referred to as Anglicization, on student learning in the context of introductory microbiology instruction. Data from Anglicized and Classical-vocabulary lab sections were collected. Data included exam scores as well as pre and post-course surveys on reasoning skills, impressions of biology, science and the course, and microbiology knowledge. Students subjected to Anglicized instruction performed significantly better on exams that assessed their abilities to apply and analyze knowledge from the course, and gained similar amounts of knowledge during the course when compared to peers instructed with standard vocabulary. Their performance in upper-level courses was also better than that of their traditionally educated peers. Hypotheses related to the effect are presented and evaluated; implications for instruction are discussed.

Although women of color have increased their presence in the workplace, many obstacles restricting career opportunities still exist. It is important that mental health professionals contribute in providing interventions to increase career opportunities for women of color. The purpose of this research is to add to the repertoire of interventions by studying the irrational career beliefs of Black women. This research utilizes the Believe It! program, an online career development program that focuses on altering irrational/maladaptive career beliefs that can prevent young females from pursuing career opportunities. An early study of Believe It! found it to be effective for Caucasian females, however the effects for minority females were less clear. The current study re-examined the effectiveness of Believe It! for minorities by altering the appearance of the animated character within the program. It was hypothesized that young African American women interacting with African American animated agents would display greater rationality in terms of career beliefs compared to young African American women interacting with Caucasian animated agents. Forty-four African American girls between the ages of eleven to fifteen were pre-tested with a battery of assessment devices addressing the irrationality of the girls' career beliefs. The measures included the Career Myths Scale, the Career Beliefs Inventory, the Occupational Sex-role Questionnaire, and the Believe It! measure. Four to eight days later, participants engaged in the online Believe It! Program; they were randomly assigned to either a matched condition (viewing the program with an African American animated agent) or a mismatched condition (viewing the program with a Caucasian animated agent). After completion of the intervention, participants were post-tested with the same assessment battery. MANCOVA and ANCOVA analyses showed that participants in the matched condition consistently benefitted from the matched intervention. Implications for this research are discussed.

The purpose of this experimental study was to investigate the effects of textual and visual annotations on Spanish listening comprehension and vocabulary acquisition in the context of an online multimedia listening activity. 95 students who were enrolled in different sections of first year Spanish classes at a community college and a large southwestern university were randomly assigned to one of four versions of an online multimedia listening activity that contained textual and visual annotations of several key words. Students then took a comprehension and vocabulary posttest and a survey to measure cognitive load and general attitudes towards the program. Results indicated that textual annotations had a significant positive effect on listening comprehension and that visual annotations had a significant positive effect on how successful students felt. No statistically significant differences were found for other variables. Participants also reported positive attitudes towards vocabulary annotations and expressed a desire to see more annotations during multimedia listening activities of this type. These findings provide further evidence of the impact that multimedia may have on language acquisition. These findings have implications for multimedia design and for future research. Language listening activities should include a variety of vocabulary annotations that may help students to understand what they hear and to help them learn new vocabulary. Further research is needed outside of the laboratory, in the online and increasingly-mobile language learning environment in order to align the research with the environment in which many students currently study. The incorporation of motivation into multimedia learning theory and cognitive load should be explored, as well as new measures of cognitive load.