What research says about vocabulary instruction for students with learning disabilities.

The importance of vocabulary knowledge to school success, in general, and reading comprehension, in particular, is widely documented (Anderson & Nagy, 1991; Baker, Simmons, & Kame'enui, 1998; Becker, 1977; Cunningham & Stanovich, 1998). Further, the recent RAND report, Reading for Understanding, suggests that vocabulary and word knowledge can contribute to improved comprehension, and it provides a sound rationale for increased emphasis on vocabulary instruction (Snow, 2002). While vocabulary knowledge is fundamental to reading comprehension (Baumann & Kame'enui, 1991; Stanovich, 1986), existing research does not support a specific vocabulary development method or program to address the discrepancies in word knowledge between students with poor and rich vocabularies (Baker, Simmons, & Kame'enui, 1998; National Research Council, 1998; Snow, 2002). Rather, the development of proficient reading skills is documented as the most effective independent word learning strategy.

To illustrate the notion that a significant amount of students' vocabulary growth may develop through independent reading, Anderson and Nagy (1991) point out that an average fifth-grade student who spends 25 minutes a day reading will encounter approximately 1 million words of text per year. If only 2% of the words are unfamiliar to students, then they would encounter 20,000 new words each year. Overall, if 1 word out of every 20 words is learned, that would account for 1,000 new words learned each year. The importance of independently learning words from context is also summarized by Adams:

 
   While affirming the value of classroom instruction 
   in vocabulary, we must also recognize its 
   limitations. By our best estimates, the growth in 
   recognition vocabulary of the school age child 
   typically exceeds 3,000 words per year, or more 
   than eight per day. This order of growth cannot 
   be ascribed to their classroom instruction, nor 
   could it be attained through any feasible program 
   of classroom instruction. (1990, p. 172) 

Based on estimates of independent reading word counts, students who read independently for at least 10 minutes each day appear to experience substantially higher rates of vocabulary growth than students who do very little independent reading (Adams, 1990; Anderson & Nagy, 1991; Cunningham & Stanovich, 1998).

Unfortunately, students who struggle with reading often fail to engage in the volume of independent reading necessary to significantly improve vocabulary development (Baker, Simmons, & Kame'enui, 1998; Stanovich, 1986). For students with learning disabilities, strategies for contextual word learning are also less proficient (Pany, Jenkins, & Schreck, 1982). As a result of ineffective word learning strategies, students with disabilities have a fragmented and less complete knowledge of words, as well as a narrow understanding of particular word features (Swanson, 1986). Overall, differences in the amount of independent reading, lack of strategies to learn words from context, and diffuse word knowledge appear to be the most critical obstacles to vocabulary development for students with disabilities (Stahl & Shiel, 1999). Because learning vocabulary during independent reading is very inefficient for students with reading difficulties, vocabulary and word learning skills must be taught. Yet, current practice deemphasizes vocabulary instruction despite the established robust and reciprocal relationship between vocabulary and reading comprehension (Snow, 2002).

Although vocabulary instruction for students with disabilities is an emerging area of research, general guidelines for instruction have been presented in the literature (e.g., Baumann & Kame'enui, 1991; Stahl & Shiel, 1999). First, poor readers should be encouraged to read. Because the use of text is important for vocabulary development and word learning, texts should be selected carefully so that a book's overall context can help students learn words. Research on the richness of vocabulary used in sources of spoken and written language has revealed that speech is "lexically impoverished" when compared to written language. Although this seems counterintuitive given the often simplistic nature of children's literature, "the relative rarity of the words in children's books is, in fact, greater than that in all of the adult conversation, except for courtroom testimony" (Cunningham & Stanovich, 1998, p. 9). In other words, children's stories are valuable sources of rare and unique words as well as rich and diverse vocabulary.

Second, vocabulary must be taught directly and sequentially (Biemiller, 2001). According to Stahl and Shiel (1999), 300 to 400 new word meanings can be taught a year through explicit instruction. When considering the importance of vocabulary to comprehension, 300 to 400 words are significant and suggestive of the need for explicit vocabulary instruction. According to research on explicit vocabulary instruction, selected vocabulary should include words that are important for understanding text, as well as functionally important words or words that students will encounter often (Stahl, 1986). Explicit instruction should also include the use of a word's context and definition, opportunities for "deep processing" (e.g., finding a synonym or antonym, making up a novel sentence with the word, classifying the word with other words, and relating the definition to one's own experience), and multiple exposures to the new word (Stahl). Finally, vocabulary should be taught through productive approaches that optimize word learning (Snow, 2002). For example, rather than focusing on a set of targeted words, instruction might focus on one word with multiple semantic connections to other words (Stahl & Shiel). Productive approaches might include teaching students strategies to learn words from context, word parts (e.g., decomposing words to examine affixes and roots), or semantic mapping (Baumann & Kame'enui, 1991).

Given that students with learning disabilities often have inadequate vocabulary knowledge and difficulties with learning as a language-based activity, it is critical to examine what research says about vocabulary instruction, specifically for students with learning disabilities. Although previous reviews of vocabulary instruction are available, they have focused exclusively on students without disabilities (e.g., Baumann & Kame'enui, 1991; National Reading Panel, 2000; Stahl & Fairbanks, 1986). A review of vocabulary instruction for students with learning disabilities is lacking even though a reasonable number of vocabulary intervention studies with samples of students with learning disabilities is available. Therefore, the purpose of this review is to summarize the findings of vocabulary intervention research studies designed for students with learning disabilities to document effective and productive approaches to word learning and to provide directions for classroom practice and future research in vocabulary instruction.

METHODS

LITERATURE SEARCH PROCEDURES AND SELECTION CRITERIA

Computer searches of PsycINFO and ERIC databases from 1978 to 2002 were conducted to locate appropriate studies. Descriptors for the computer search included "reading," and "vocabulary or concepts," "concept mapping or semantic mapping," or "mnemonics," and "instruction, interventions, or strategies," and "disabilities." In addition, we carefully examined the references from the identified studies themselves. Finally, we hand-searched the following journals to locate the most recent literature: Education and Treatment of Children, Exceptional Children, Journal of Learning Disabilities, The Journal of Special Education, Learning Disability Quarterly, Learning Disabilities Research and Practice, Reading and Writing Quarterly, Overcoming Learning Difficulties, and Remedial and Special Education. It must be noted that we did not include other sources of the literature (e.g., Dissertation Abstracts International) of contact active researchers in this area for unpublished studies. As such, conclusions based on this review should be viewed as tentative.

To judge the appropriateness of each article, we evaluated studies using six criteria. First, only published studies were included in the review. Unpublished doctoral dissertations and presentations were excluded. Thus, the review may represent a potential bias toward more sound research designs and stronger effect sizes than perhaps are found in unpublished articles (Lipsey & Wilson, 1993). Second, experimental, quasi-experimental, and single-subject design studies were included in the review. Third, the recipients of the intervention were identified as children with learning disabilities. Fourth, the studies had to include elementary, junior high, or high school students with learning disabilities. Fifth, studies that described vocabulary instruction related to sight word reading (e.g., Keel, Slaton, & Blackhurst, 2001) rather than word knowledge (meaning) and/or did not present evidence of the effects of systematic use of vocabulary strategies on vocabulary learning were excluded. For example, although the Reyes, Duran, and Bos (1989) study addressed vocabulary strategies, it was not included because the primary purpose of this study was to determine the effects of code switching of language alternation on the vocabulary performance of bilingual students with learning disabilities. Finally, studies included in this review had to measure vocabulary outcomes. Interrater agreement for article inclusion and exclusion was 96% (range = 90% to 100%).

Using these criteria, a total of 20 studies were located. However, the final review included 19 relevant vocabulary studies, because 1 study (Woodward, Carnine, & Collins, 1988) reported data from a previous study (Johnson, Gersten, & Carnine, 1987). Interrater agreement on coding of pertinent sample and study characteristics for each study was conducted. The mean interrater agreement was 93% (range = 74% to 100%). We calculated and reported effect sizes (ES) or percentage of nonoverlapping data (PND) when sufficient data (e.g., means and standard deviations, clear graphs and charts) were provided for group and single-subject design studies. Group effects were calculated using Cohen's d, in which ES = M posttest score of the experimental group - M posttest score of the control group/pooled standard deviation. In some situations in which means and standard deviations were not provided, effect sizes were calculated using F or t-test scores provided (see Murawski & Swanson, 2001). When a study did not include a control group and used a pretest-posttest research design on the treatment group only, we used the standardized mean change to calculate the intervention effect (Becker, 1988). The effect size in such instances was determined by computing the difference between the posttest and pretest means divided by the pretest standard deviation.

RESULTS

STUDY CHARACTERISTICS

This review of 19 studies included 17 (89.5%) group design studies and 2 (10.5%) single-subject design study. Of these studies, 12 were published from 1978 to 1989 and 7 were published from 1990 to 1996. Three studies included two experiments (Boettcher, 1983; Horton, Lovitt, & Givens, 1988; Mastropieri, Scruggs, Levin, Gaffney, & McLoone, 1985), 1 study included three experiments (Pany, Jenkins, & Schrek, 1982), and 1 study included six experiments (Bos & Anders, 1992). However, Experiment 2 in the study by Horton et al. and Experiment 1 in the Pany et al. study included only students without disabilities. As such, these were not included in the review. The 19 articles yielded a final sample of 27 investigations. A total of 901 students with learning disabilities participated in the studies reviewed.

Sixteen investigations reported specific grade levels with ranges of Grades 4 through 12, and 19 investigations reported participants' ages with ranges of 9 to 16.2 years. Twelve investigations reported the gender of participants (229 males and 96 females). The length of interventions varied from 1 session to 15 sessions (M = 5.25; SD = 3.71, n = 26) and session duration ranged from 2 min to 50 min (M = 36.20; SD = 16.54, n = 20). Regarding intervention setting, 30% were conducted in special education classrooms; 11% in content area classrooms; and 18% in a library, a computer lab, of quiet rooms adjacent to the classroom. Eleven investigations (41%) did not report the settings in which the research was conducted. With regard to delivery of instruction, researchers (52%) most often delivered the instruction, followed by teachers (26%), computers (19%), and a librarian (4%). In general, instruction occurred most often in groups (63%), followed by individual instruction (33%) and instruction in pairs (4%). While the majority of investigations (78%) did not provide any treatment fidelity information, 1 investigation (4%) reported high treatment fidelity of 98%, 4 investigations (15%) indicated that a checklist or teaching script was developed to ensure treatment consistency, and 1 investigation (4%) reported using videotaping to ensure treatment fidelity. Only 2 investigations (7%) reported reliability data for the dependent measures. Transfer data were collected in 8 investigations (30%). Maintenance data were collected in 15 investigations (56%). Twenty-five investigations (93%) employed tests developed by researchers and 2 investigations (7%) used a standardized test.

FINDINGS BY TYPE OF INTERVENTIONS

Interventions were categorized as keyword of mnemonic strategies, cognitive strategies, direct instruction (DI), activity-based method, constant time delay (CTD), and computer-assisted instruction (CM). Table 1 provides descriptive information for each of the intervention studies.

Keyword or Mnemonic Strategies. Keyword or mnemonic strategies involve two components, keyword and imagery links (Mastroprieri, Scruggs, & Fulk, 1990). The keyword link provides a similar sounding familiar word to associate with the unknown vocabulary word. The imagery link provides a picture of the definition of the unknown word interacting with a picture of the keyword. The phonetic and visual imagery components of the target word and its definition serve to facilitate memorization and recall of target vocabulary. Six studies were found that implemented a keyword mnemonic strategy to help students learn key vocabulary.

In a two-experiment study, Mastropieri, Scruggs, Levin, Gaffney, and McLoone (1985) compared the keyword strategy with a DI method on the recall of vocabulary definitions. Students in both experiments were stratified by grade level and randomly assigned to the two treatment conditions. Students in the mnemonic instruction condition were taught to use the keyword strategy. In the DI condition, students were taught using drill and practice procedures. While an interactive picture of the vocabulary word and word clue were presented in the mnemonic condition, students in the DI condition were provided with a picture of the vocabulary word. Students in both conditions were taught 16 low-frequency, unfamiliar English vocabulary words. In Experiment 1, the instructor provided both the keywords and mnemonic pictures to 32 students with learning disabilities in Grades 7, 8, and 9, whereas in Experiment 2, 37 students with learning disabilities in Grades 6, 7, and 8 had to self-generate the mnemonic imagery. Results indicated that students in the mnemonic instruction…