Computerizing reading training: Evaluation of a latent semantic analysis space for science text

The effectiveness of a domain-specific latent semantic analysis (LSA) in assessing reading strategies was examined. Students were given self-explanation reading training (SERT) and asked to think aloud after each sentence in a science text. Novice and expert human raters and two LSA spaces (general reading, science) rated the similarity of each think-aloud protocol to benchmarks representing three different reading strategies (minimal, local, and global). The science LSA space correlated highly with human judgments, and more highly than did the general reading space. Also, cosines from the science LSA spaces can distinguish between different levels of semantic similarity, but may have trouble in distinguishing local processing protocols. Thus, a domain-specific LSA space is advantageous regardless of the size of the space. The results are discussedin the context of applying the science LSA to a computer-based version of SERT that gives online feedback based on LSA cosines.     

Typing versus thinking aloud when reading: Implications for computer-based assessment and training tools

The goal of this study was to assess the impact of modality of production of think-aloud protocols on reading strategies. Readers in two studies spoke or typed protocols for narrative or science texts and completed comprehension tests for each text. Human judges identified the presence of paraphrasing, bridging inferences, and elaborating within the protocols. Reading comprehension skill was assessed with the Nelson-Denny test. With respect to narrative texts, paraphrasing and bridging were less frequent when readers were typing than when they were thinking aloud. With respect to science texts, less-skilled readers made bridging inferences more frequently when typing than when speaking. Conversely, skilled readers generated more paraphrases than bridges when typing thoughts but not when speaking. These results have implications for computer-based tools for reading assessment and intervention.     

Assessing the format of the presentation of text in developing a Reading Strategy Assessment Tool (R-SAT)

We are constructing a new computerized test of reading comprehension called the Reading Strategy Assessment Tool (R-SAT). R-SAT elicits and analyzes verbal protocols that readers generate in response to questions as they read texts. We examined whether the amount of information available to the reader when reading and answering questions influenced the extent to which R-SAT accounts for comprehension. We found that R-SAT was most predictive of comprehension when the readers did not have access to the text as they answered questions.     

Collaborative dialogue patterns in naturalistic one-to-one tutoring

Naturalistic one-to-one tutoring is more effective than traditional classroom teaching methods, but there have been few attempts to examine the features of normal tutoring that might explain its advantage. This project explored dialogue patterns in two samples of naturalistic tutoring with normal unskilled tutors (as opposed to expert tutors): graduate students tutoring undergraduates in research methods and high school students tutoring 7th graders in algebra. We analysed the extent to which those tutoring protocols manifested components that have been emphasized in contemporary pedagogical theories and intelligent tutoring systems: active student learning, sophisticated pedagogical strategies, specific examples and cases, collaborative problem solving and question answering, deep explanatory reasoning, convergance toward shared meanings, feedback, error diagnosis and remediation, and affect. The most prominent components consisted of collaborative problem solving, question answering, and explanation in the context of specific examples. We identify frequent dialogue patterns that characterize these collaborative processes.     

How to Assess a Model's Testability and Identifiability

Formal definitions are given of the following intuitive concepts: (a) A model is quantitatively testable if its predictions are highly precise and narrow. (b) A model is identifiable if the values of its parameters can be ascertained from empirical observations. (c) A model is redundant if the values of some parameters can be deduced from others or if the values of some observables can be deduced from others. Various rules of thumb for nonredundant models are examined. The Counting Rule states that a model is quantitatively testable if and only if it has fewer parameters than observables. This rule can be safely applied only to identifiable models. If a model is unidentifiable, one must apply a generalization of the Counting Rule known as the Jacobian Rule. This rule states that a model is quantitatively testable if and only if the maximum rank (i.e., the number of linearly independent columns) of its Jacobian matrix (i.e., the matrix of partial derivatives of the function that maps parameter values to the predicted values of observables) is smaller than the number of observables. The Identifiability Rule states that a model is identifiable if and only if the maximum rank of its Jacobian matrix equals the number of parameters. The conclusions provided by these rules are only presumptive. To reach definitive conclusions, additional analyses must be performed. To illustrate the foregoing, the quantitative testability and identifiability of linear models and of discrete-state models are analyzed. Copyright 2000 Academic Press.     

Components of events and processes

Rips and Estin (1998) provided evidence that mental events such as dreaming are more homogeneous than physical events such as checking out a book; that is, their parts are more difficult to distinguish. In their experiment, participants listed more distinctive properties for the parts of physical events than for the parts of mental events. However, the physicality of stimuli was confounded with temporal aspects. Mental stimuli tended to be processes, and physical stimuli, events. This study tested homogeneity with new stimuli separating out the factors of physicality and aspect. Consistently, both physicality and aspect had significant effects on the perceived homogeneity of activities, as measured by the number of listed parts, the number of distinctive properties of each part, and homogeneity ratings. The study shows that homogeneity is strongly influenced by aspect but that physicality remains a robust factor for homogeneity, even after taking aspect into account.     

The sounds of dominance.

A study was conducted to determine the effects of vocal cues on judgements of dominance in an interpersonal influence context. Physical measures of human vocal cues and participant ratings of dominance were obtained from videotapes of actors delivering short influence messages. After controlling for linguistic and visual content of messages, results indicated that mean amplitude and amplitude standard deviation were positively associated with dominance judgments, whereas speech rate was negatively associated with dominance judgments. An unexpected interaction revealed that mean fundamental frequency (F₀) was positively associated with dominance judgments for male speakers but not significantly associated with dominance judgments for female speakers. F₀ standard deviation was not significantly associated with dominance judgments. Results support the conclusion that dominance judgments are inferred from multiple sources of information and that some vocal markers of dominance are more influential than others.     

Time flies faster if a person has a high working-memory capacity

Attention affects the perception of time, and the ability to control attention is reflected in measures of working-memory capacity. Individuals with low working memory capacity have more difficulty maintaining focus on a task than high-capacity individuals, particularly when faced with contextual distracters. This experiment examined the effect of working-memory capacity on the perception of temporal duration while performing a cognitive task. We predicted that low-capacity participants would be more likely to direct attention away from the cognitive task and towards the contextual distraction of time, and consequently perceive temporal duration more accurately, and perform the cognitive task less accurately, than high-capacity participants. The results showed that when performing both tasks simultaneously, low-capacity participants were less accurate than high-capacity participants on the cognitive task, but were more accurate on the timing task. High-capacity participants, conversely, were more accurate in the non-temporal cognitive task at the cost of monitoring duration.