No abstract

We have developed the TextEvaluator system for providing text complexity and Common Core-aligned readability information. Detailed text complexity information is provided by eight component scores, presented in such a way as to aid in the user's understanding of the overall readability metric, which is provided as a holistic score on a scale of 100 to 2000. The user may select a targeted US grade level and receive additional analysis relative to it. This and other capabilities are accessible via a feature-rich front-end, located at http://texteval-pilot.ets.org/TextEvaluator/.

We present an automated method for estimating the difficulty of spoken texts for use in generating items that assess non-native learners' listening proficiency. We collected information on the perceived difficulty of listening to various English monologue speech samples using a Likert-scale questionnaire distributed to 15 non-native English learners. We averaged the overall rating provided by three nonnative learners at different proficiency levels into an overall score of listenability. We then trained a multiple linear regression model with the listenability score as the dependent variable and features from both natural language and speech processing as the independent variables. Our method demonstrated a correlation of 0.76 with the listenability score, comparable to the agreement between the nonnative learners' ratings and the listenability score. * We would like to thank to Yuan Wang for data collection, Kathy Sheehan for sharing text difficulty prediction system and insights, and Klaus Zechner, Larry Davis, Keelan Evanini, and anonymous reviewers for comments.

This paper considers whether the sources of linguistic complexity presented within texts targeted at 1st‐grade readers have increased, decreased, or held steady over the 52‐year period from 1962 to 2013. A collection of more than 450 texts is examined. All texts were selected from Grade 1 textbooks published by Scott Foresman during the targeted time period. Analyses are implemented using the TextEvaluator® tool, a comprehensive text complexity evaluation tool designed to help teachers, textbook publishers, and test developers identify and quantify text‐based sources of comprehension difficulty within informational, literary, and mixed texts. Results suggest that 1st‐grade textbooks published over the past half century have included an increasing proportion of informational passages, and this shift has been accompanied by the following specific changes: (a) an increase in the proportion of words that tend to appear less frequently in printed text, (b) an increase in the proportion of words that are more characteristic of academic text as opposed to fiction or conversation, (c) lower levels of referential cohesion, (d) lower levels of narrativity, and (e) fewer instances of an interactive/conversational style. These findings suggest that, in contrast to the claim of a “general, steady decline” in textbook complexity, text‐based sources of comprehension difficulty within Grade 1 texts have either risen or held steady throughout the past half century.

Writing assistance systems, from simple spelling checkers to more complex grammar and readability analyzers, can be helpful aids to nonnative writers of English. However, many writing assistance systems have two disadvantages. First, they are not designed to encourage skills learning and independence in their users; instead, users may begin to use the system as a crutch. Second, they use a "one-size-fits-all" approach, treating all writers' problems as equivalent. In this paper we describe TechWriter, a personalizable writing assistance program for advanced learners of English that encourages skills learning. We describe TechWriter's basic writing assistance functionalities, how it can be used by writers alone and working together with writing tutors, who it can be personalized for, and how it can help writers acquire better writing skills over time.