Embracing Complexity: Assessing Morphological Knowledge for Middle School Students

• In Event: Dealing With Hard Words: Measurement, Relations, and Instruction of Academic Vocabulary That Considers Word Properties

Abstract

In measuring academic language, the morphological nature of academic vocabulary must be taken into account. Most academic vocabulary is morphologically complex where words are built from morphemes (i.e., roots and affixes) that are combined to express different meanings and serve varied syntactic roles (e.g., tasteful, distaste). Morphological knowledge plays a major role in students’ acquisition of academic language (e.g., Nagy & Townsend, 2012; Pacheco & Goodwin, 2013) with students who are morphologically aware often able to ascertain the meanings of unfamiliar words like nationalistic through analysis of the morphemic units and their composite meanings (Anglin, 1993; Nagy & Anderson, 1984). As such, assessing morphological knowledge can provide important understandings about what students know about word parts and how they are used to form academic words.

Assessing morphological knowledge is complex (Bowers et al., 2010; Carlisle, 2010; Carlisle & Goodwin, 2013; Nagy, Carlisle, & Goodwin, 2014). Many researcher-developed measures were created when less emphasis was placed on reporting validity, reliability, and other technical characteristics of measures. They provide limited information on design decisions, such as criteria for words chosen to be part of the assessment. Also, most were developed to answer a particular question and focus on one aspect of morphological knowledge (e.g., suffix use) rather than examining the multiple dimensions that capture the relation of morphological knowledge and literacy.

The goal of our study is to deepen understanding of morphological knowledge by presenting data on a large number of morphological items that vary in content and task type. We consider performance of 250 fifth through eighth graders on 130 morphological items which represent 12 morphological tasks of different formats (e.g., sentence completion, judgment, analogy, etc.) and include varied content (e.g., different morphological relationships, lexicality, word parts, frequency, transparency, and length) that could impact recognition and application of morphological knowledge to literacy tasks. Results regarding the dimensionality of morphological knowledge and item psychometric properties will be presented. Also, discussion of good and poor items will be included.

Methods will involve dimensionality explorations using both exploratory and confirmatory factor analysis. Item and person parameters will be estimated using IRT with BILOG-MG software (Zimkowski et al., 2003). One- and two-parameter logistic (1-PL, 2-PL) multiple group analyses (i.e., to account for vertical equating across grades) will be conducted to identify which items are most and least informative and which fall into an appropriate range of difficulty. Data is still being collected and analyzed, but initial results suggest that morphological knowledge is best considered a multidimensional construct best fit by a bifactor model where there is a core set of knowledge that highlights the presence of meaningful units within words and there are also different aspects of morphological knowledge that can be applied in varied ways to support performance in different areas of literacy. Overall, results from this study can guide researchers in how to best assess morphological knowledge, which is critical when considering the morphological nature of academic vocabulary.

Authors

• Amanda P. Goodwin, Vanderbilt University

• Yaacov Petscher, The Florida Center for Reading Research