View source for REAP Study on Vocabulary Stretch (Spring 2006)

== REAP Study on Vocabulary Stretch ==
 

=== Logistical Information ===

{| border="1"
|+ 
|-
| '''Contributors''' || Maxine Eskenazi, Alan Juffs, Michael Heilman, James Sanders, Lois Wilson   
|-
| '''Study Start Date''' || January, 2006  
|-
| '''Study End Date''' || April, 2006  
|-
| '''Learnlab Courses''' || English Language Institute Reading 4 (ESL LearnLab) 
|-
| '''Number of Students''' ||  33
|-
| '''Total Participant Hours (est.)''' ||  220
|-
| '''Data in Datashop''' || no 
|}

=== Abstract ===

There are competing goals in providing a text for practice of target vocabulary items.  First, there is a goal to maximize the density of practice opportunities in a given text.  That is, texts should be preferred that contain many target vocabulary words so that the learner can encounter as many words as possible in the time available.  Second, there is a goal to provide texts that can be comprehended by the learner in order to prevent frustration and allow deep processing of the meaning of the context around the target words.  If too many unfamiliar words appear in a text, target words or not, then the reader will not be able to comprehend or recall the meaning of the text.  Third, there is a goal not to overload the student with too many new words to remember, even if a text can be recalled.

Therefore, the tutor should seek a proper balance between the goals of providing sufficient quantity and difficulty of practice.  This study was aimed at locating this balance.  

=== Glossary ===

''Vocabulary Stretch:'' of a text, the percentage of unfamiliar or unknown vocabulary words that appear.

=== Research question ===

What percentage of potentially unknown target vocabulary words is optimal in a text given for practicing target vocabulary words?

=== Dependent variables ===

Post- and retention test cloze question performance for words answered incorrectly on the pre-test.

Post- and retention test sentence production performance for words answered incorrectly on the pre-test.

Number of words practiced.

=== Independent variables ===

Preferred number of target vocabulary words in a given text (2 or 4 words per text).

=== Hypotheses ===

=== Findings ===

The study found no significant differences between the two conditions (2 or 4 words/text).  This null result may be due to statistical power and mistakes in the study design.  The tutor had difficulty providing high quality readings with four vocabulary words.  Thus, the tutor often had to give texts with 3 or even 2 words to students in the 4-word condition.  Additionally, the optimal number of words per text (the density of practice) may be higher than 4 words.  There may be a threshold at which the density of practice becomes too high.  This threshold may actually be at 5 or even 10 words given the average length of the texts used by the REAP tutor.  Some research suggests that up to 5% of the words in a text can be unknown before comprehension is impeded (Laufer, 1992).  Although the task for that study was different (reading comprehension rather than vocabulary acquisition), it suggests that the optimal level of stretch may be such that as many as 20 words may be unknown in a text with 400 word types (which is approximately the mean number of word types in REAP texts).  Other research, however, suggests higher thresholds up to 98% (Nation, 2001).

Although the results for stretch from this study were inconclusive, a great deal of interesting data were gathered.  Some interesting results have been found, and further analysis is possible. 

Many potential variables exist that could lead to successful learning of new vocabulary: general ESL ability as measured by the MTELP, time on task, number of words looked up, pre-test scores on the academic word list. A step-wise linear regression showed that the most significant predictor of acquisition of new vocabulary was the size of the student's vocabulary scores on the academic word list prior to instruction (r= 0.77, p < .0001). However, the effect of this prior knowledge was diminished by the number of texts the students read r= -0.17). Together, in the stepwise regression, these two variable accounted for 67% of the variance (r2= .67).

This finding corroborates findings of Stanovich (1986) and James (1996) who refer to the phenomenon as the "Matthew effect" (after a Bible passage), whereby the rich get richer and the poor get poorer. In this case, it is not 'general richness' expressed in terms of higher general ESL proficiency, but specifically knowledge of words on the academic word list. In other words, the more knowledge components in the domain of acquisition a learner brings to the task, the greater the potential for future learning. 

It was also found that students frequently access dictionary definitions for non-target words that are well below their expected reading level (e.g., "hitting").  This may lead to revisions of curricula for ESL vocabulary learning.

=== Descendents ===

=== Notes ===

Target vocabulary lists were determined by a lengthy pre-test using multiple choice cloze questions.  Each question takes 30-60 seconds on average, depending on the student.  In later studies, self-assessment pre-tests were preferred due to time constraints and initial impressions of the students.  Lengthy pre-tests for determining target word lists leave a poor initial impression of the tutor.

=== Annotated bibliography ===

Laufer, B. (1992). How much lexis is necessary for reading comprehension? In Vocabulary and Applied Linguistics, Pierre J. L. Arnaud and Henri Béjoint (eds.), 126–132. London: Macmillan.