Abstract
Goldstein, H., Ziolkowski, R. A., Bojczyk, K. E., Marty, A., Schneider, N., Harpring, J., & Haring, C. D. (2017). Academic vocabulary learning in first through third grade in low-income schools: Effects of automated supplemental instruction. Journal of Speech, Language, and Hearing Research, 60, 3237–3258.
Study Purpose
Vocabulary serves as a pivotal link between oral and written language (National Reading Panel [NRP], 2000). Students with limited vocabularies are at risk for difficulties in developing fluent reading skills and in comprehending reading passages (Nash & Donaldson, 2005; Scarborough, 2005). For children who are at risk for reading disabilities due to limited vocabulary, explicit teaching may be particularly important (Coyne, McCoach, Loftus, Zipoli, & Kapp, 2009; Foorman & Torgesen, 2001). To determine the potential for overcoming limited vocabularies, there is a need to examine the effects of explicit vocabulary instruction for children identified with developmental delays as well as children at risk for language and reading disabilities. The purpose of this study was to determine to what extent children in low-income schools learn novel academic vocabulary words as a result of explicit vocabulary instruction.
Participants
All first graders in two elementary schools participated in the study. Over 90% of the participants were eligible for reduced or free lunch. Students with disabilities and at risk for disabilities, who were enrolled at lease part-time in general education classrooms, were included in the study. The students were randomly assigned to a vocabulary instruction group and a phonics instruction group. Students were tracked from first through third grade. First-grade students were randomly assigned to two treatment conditions that were administered simultaneously within classrooms: a vocabulary instruction group and a phonics instruction comparison group. The students were randomly assigned in first grade and then continued to receive the intervention in the same experimental condition in second and third grade.
Materials/Procedures
The vocabulary words chosen for teaching were what have been termed Tier 2 words (Beck, McKeown, & Kucan, 2002). Tier 1 words are basic words that rarely require direct instruction (e.g., boy); Tier 3 words are situation specific and are seldom used outside of a specific context (e.g., decibel). In contrast, Tier 2 words are found across a variety of linguistic and academic environments (e.g., layer, consequence, adult). Academic vocabulary incorporates both Tier 2 and Tier 3 and is of considerable importance for developing the academic language needed for children to be successful in content areas. Beck et al. (2002) recommend prioritizing Tier 2 words to maximize impact on vocabulary knowledge for students at risk for reading disabilities.
The vocabulary instruction and the comparison condition were both administered simultaneously in listening centers run by research assistants. Students independently listened to prerecorded interactive activities and interventions through headphones for 20 min, the same activity was repeated 4 days per week. Each student was provided with a daily packet that contained a copy of a decodable book from the Open Court Curriculum, a worksheet, and a pencil. All students listened to the same decodable book and engaged in choral reading at the beginning of each session (approximately 5 min). Students listened to the CD and completed the vocabulary or phonics worksheets independently following the instructions provided. Scripts for both conditions were developed. All scripts required active participation and responding from students by asking questions and providing time for spoken and written responses. Eighteen weeks of instruction were presented in four 4- to 5-week units presented during the school year.
In Grades 1 and 2, the curriculum included one “anchor” (Tier 1) word from Open Court stories and six novel Tier 2 vocabulary words (two nouns, two verbs, and two adjectives). In Grade 3, the curriculum included only six Tier 2 words. Over 3 years, students were taught 324 challenging Tier 2 words. The words were chosen from the Coxhead Academic Word List (Coxhead, 2000). (The words are listed in the article, but are also available on several websites such as https://www.victoria.ac.nz/lals/resources/academicwordlist). The Coxhead (2000) wordlist scales words based on their frequencies in textbooks across disciplines. For each lesson an Open Court story was always read first. The anchor and Tier 2 words were linked to the Open Court story. Each script contained a sequence of script elements for each word, all incorporating interactive activities. The scripts presented user friendly definitions for the novel words. Students were given opportunities to identify the words, say the words, relate words to pictures, and define words.
Assessment Measures
Four curriculum-based assessments of targeted vocabulary words were used to monitor the acquisition of levels of word knowledge. These researcher-made measures mapped onto the levels 2 to 4 of word learning outlined by Dale (1965): 2 = word recognition, 3 = receptive identification, 4 = expressive labeling and decontextualized definitions.
Recognition Probes
Recognition probes were administered to determine the student’s ability to recognize the target word. The task required the student to identify the target word from two choices: the target word and a phonetically matched nonsense or rare word. For example, children were asked, “Which of these words have you heard before: demolish or puh-NEW-kuv?” This measure maps onto Level 2 of Dale’s (1965) taxonomy: Has the child ever heard the word before, and if so, can he or she recognize it when presented with a distractor word?
Receptive Identification Probes
The receptive identification probes represented a generalization task to evaluate whether children had the depth of word knowledge to identify novel pictures of the words they learned. The receptive probes were not the same target pictures that children were exposed to during the intervention, so it was not a test of whether the child simply remembered the training pictures. The receptive probes mapped onto Level 3 of Dale’s (1965) taxonomy. Receptive identification probes were modeled after the Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4).
Expressive Labeling Probes
Expressive labeling probes assessed students’ ability to provide a label for the novel vocabulary word when given the trained picture stimulus with or without the definition. This measure maps onto Level 4 of Dale’s (1965) taxonomy. The student viewed the picture of the target item that they were trained with during the intervention on a single plate and was asked to label the picture in response to the question, “What is this?” If the child did not provide a correct response, the child was asked, “What word means ______?” (e.g., What word means ten years?—for decade). A correct response required the student to provide the trained vocabulary word; all other responses were scored as incorrect.
Decontextualized Definition Probes
Decontextualized definition probes were administered to determine the student’s level of word knowledge without contextual support from text or illustrations. These probes map onto Level 4 of Dale’s (1965) taxonomy. The task required the student to explain the meaning of the novel vocabulary words (i.e., synonym or definitional knowledge) in response to the examiner’s request to “Tell me everything you know about ___.” If the child only provided one attribute of a word, the child was prompted further by saying, “Tell me something else about ____.” A correct response required the student to provide at least one synonym or brief description (i.e., function, characteristic, etc.) of the target word; all other responses were scored as incorrect.
Unit Assessments
The primary dependent variables included in this report were unit assessments. Unit assessments included all four subtests and were administered every 5 weeks in the fall semester and every 4 weeks in the spring semester.
Norm-Referenced Measures
Norm-referenced measures of vocabulary, language, and cognitive development were administered at the beginning of the study. The standardized measures of vocabulary include the PPVT-4 (Dunn & Dunn, 2007) and the Expressive Vocabulary Test, Second Edition (EVT-2; Williams, 2007). The Clinical Evaluation of Language Fundamentals–Fourth Edition (CELF-4; Semel, Wiig, & Secord, 2003) was used to assess oral language. The Kaufman Brief Intelligence Test, Second Edition (KBIT-2; Kaufman & Kaufman, 2004) was administered to obtain an estimate of verbal and nonverbal intelligence.
Results
The vocabulary and phonics groups differed significantly on vocabulary assessments at each unit test for each measure reflecting deeper levels of knowledge, from word recognition to receptive identification to expressive labeling to decontextualized definitions.
The effect sizes were large, with mean differences between groups averaging 19% for the word recognition tests and 21% for the receptive identification tests.
The mean differences averaged 38% and 18% for the expressive labeling and decontextualized definition tests, respectively.
Of the 324 academic vocabulary words taught in 3 years, children in the experimental group recognized a mean of 298 words (92%). The large effect size reflects recognition of words 33% above chance compared with 13% above chance for the comparison group.
The experimental group receptively identified a mean of 234 words (72%). The large effect size reflects receptive identification of words 38% above chance compared with 16% above chance for the comparison group.
Chance performance is not a factor for the expressive labeling and decontextualized definition tasks. The experimental group expressively labeled a mean of 135 pictures (42%) and defined a mean of 70 words (22%). This compares to knowledge of a mean of 10 to 12 (3%–4%) words for the comparison group. Because of these multiple measures, these results provide an indication of deeper knowledge of learning of targeted vocabulary words.
Performance on the various vocabulary assessments was consistent with the researchers’ predictions, with the best performance on the Level 2 recognition task and poorest performance on the Level 4 decontextualized definition task. Children who received vocabulary instruction demonstrated Level 2 and Level 3 knowledge of most words, even though Level 4 knowledge was not demonstrated for the majority of words.
Children in the vocabulary instruction group who also had an individualized education program (IEP) benefited from the intervention but not to the same degree as those without an IEP.
Children with initial higher PPVT-4 scores benefited more from vocabulary instruction than those with lower PPVT-4 scores.
Rate of vocabulary learning increased from first through second grade but decreased in third grade.
The reason for the decrement in word learning in third grade is not clear. It is possible that the target vocabulary words were relatively more difficult in third grade. As the researchers began to exhaust words from the Academic Word List (Coxhead, 2000), they were left with more abstract words that were difficult to incorporate into stories. Of interest is that the comparison group, who exhibited much poorer vocabulary learning, exhibited a similar pattern—they also performed better in second grade than third grade. The academic vocabulary words selected for this intervention were challenging. Despite being characterized as Tier 2 words, these words rarely seemed to be encountered in kindergarten to third grade enough to be learned by children in the comparison group; they learned to label 10 of the targeted words and to define 12 words in 3 years, whereas the intervention group learned to label 135 words and to define 70 words on average.
This investigation demonstrated that it is possible to significantly increase students’ academic vocabularies. The study did not report on the relationships between better academic vocabulary knowledge and reading comprehension. The authors do acknowledge that replication of the results may be problematic because they employed an automated intervention managed by researchers and graduate students. However, the investigation does show that is possible to increase students’ academic vocabulary by explicitly teaching the vocabulary. Speech-language pathologists may find the Word Feast books for elementary, middle school, and high school which are available from Pro-Ed, particularly useful. These books, designed by and for speech-language pathologists, use research-based methodologies for teaching academic vocabulary. For upper elementary and middle school students, there is the Word Generation program developed by noted literacy researchers. It has an extensive collection of lessons for student from fourth through eighth grade intended for use by teachers. It is available for free download from http://wordgen.serpmedia.org/