Gallagher, Taboada Barker, Beck, and Buehl (2019) evaluated yet another academic vocabulary intervention for middle school students. They concluded that language minority (LM) students (called “English Bilinguals” in the study) “need explicit instruction to improve vocabulary knowledge” (p. 15). They claimed to show that LM students were unable to acquire any new academic words incidentally from the texts their teachers read to them.
Their data do not support this conclusion.
The Study
The study examined the performance of 149 sixth-graders in a seven-week vocabulary intervention called “U.S. History for Engaged Readers.” There were both English-only (EOs) (called “English Monolinguals” in the paper) and LM students (called “English Bilinguals”). The LM group appeared to include both native bilinguals and English language learners.
Gallagher and colleagues selected a group of seven general academic words and 14 discipline-specific social studies terms to be taught explicitly to students. They also counted other academic and social studies words that appeared only in the assigned texts that teachers were to read to students in class (with the students following along in the book). Teachers were told that for those words, they should “draw attention” to them when reading aloud, but not give explicit instruction on them.
Observations were made in each of the seven classrooms in the study to determine which words teachers actually taught and which words were only found in the related texts. They then measured how many of these “explicit” words were learned compared to the number of text-only, “incidental” words that were acquired.
Table 1 summarizes the results from the LM and EO students, with pretest and post-test scores on explicit and incidental words (from their Table 2, p. 9).
Table 1: Test Results on Reading Comprehension (Pretest) and Vocabulary for EO and LM Students
[table "33" not found /]The researchers found that while EO students made significant progress on both explicit and incidental words, LM students only made statistically significant progress on the explicitly taught words. This finding leads Gallagher et al. to conclude that LMs cannot acquire academic vocabulary from reading alone.
The Problems
There are several flaws in the study:
(1) EOs and LMs were clearly at different grade levels at the start of the school year. The researchers said they were both groups were “approximately at [a] fifth-grade level” (p. 6). But a look at the scale scores and norming sample used by Gates-MacGinitie shows otherwise: EOs scored at the spring, fifth grade level, while LMs scored at the spring, third grade level – a two grade level difference. The books used for the incidental words were 3rd to 5th grade level, meaning that many may have been too difficult for the LM students.
Further evidence of the differences between the groups can be found in the pretest scores. The LM group was lower in both incidental (47% v. 54%) and explicitly-taught (44% v. 58%) words compared to the EO group.
These differences alone may account for the LM group’s poor showing on the incidental word test. The texts were much more comprehensible to the EOs (and at or below their grade level) than to LMs.
(2) The number of times the incidental words occurred in the texts was very low, far below what “read-and-test”* studies on L2 adults have found sufficient for acquisition. Nation (2014), for example, estimated that a word would need to occur at least 12 times to give adults a reasonable chance of acquiring it incidentally. Some estimates are as high as 20 to 25 times.
Of the 12 words Gallagher et al. tested that appeared to most or all students incidentally, more than half occurred four times or fewer, and only one was repeated 12 times or more.** It is a surprise that any of the words were acquired in these conditions. Nevertheless, the EO group, for whom the texts were probably more comprehensible, still improved their incidental score by 14 percentage points.
(3) The authors fail to consider the efficiency of vocabulary instruction. They state that improving the vocabulary of LM students is an urgent priority in order to prevent them from falling further behind their EO classmates. They should therefore be interested in finding the most time efficient way of improving that vocabulary.
Yet they report no details on how much time was spent on the explicit words versus the (minimal) time spent reading incidental target words in the texts. It is impossible to determine precisely the relative efficiency of explicit learning vs. incidental acquisition.
When we consider, however, that the EO group improved equally on the explicit and incidental words, incidental acquisition was clearly the faster route. Had LM students encountered texts closer to their grade level, they may have achieved similar results. Other research (e.g. Mason, 2007; McQuillan, 2019a; 2019b; 2019c) has found that incidental acquisition is usually far superior to explicit instruction in time efficiency.
(4) There were other issues that may have influenced the outcome of the study.
- Students didn’t choose their own texts, which means a lack of interest and/or comprehensibility (as noted above) could have prevented them from making progress.
- Test reliability on the vocabulary measure was below what is normally considered acceptable (Cronbach’s alpha: .60-.71). This means that the measure may not have been able to detect changes in vocabulary knowledge.
- Students had to complete a very challenging vocabulary test, on which they had to distinguish between “immutable related and circumstantially related words” (p. 8). While there’s a place for such “deep knowledge” probes, they may fail to pick up partial acquisition, especially in cases such as this where target words were repeated so few times.
The researchers made no mention of the extensive reading studies which provide ample evidence that L2 students, like all students, can and do acquire vocabulary through reading. Gallagher et al. provide no serious counter-evidence to that finding.
*Read-and-test studies give subjects a text with unknown words in it, then have them complete a surprise vocabulary test on those words after their reading. The studies give us estimates of how often a word typically needs to occur to be acquired by most readers. Gallagher and colleagues cite only one source in their literature review on L2 word acquisition (Huckin & Coady, 1999), an article which itself spends less than a paragraph discussing read-and-test studies.
**I excluded one outlier on the test that appeared 118 times (colony), since only three of the 92 LM students encountered it in the incidental condition only.
Follow Us!