Root Words 4

Journal of Educational Psychology 2001, Voi. 93, No. 3, 498-520

Copyright 2001 by the American Psychological Association, Inc. 0022-0663/01/S5.00 DOI: I0.1037//0022-0663.93.3.498

Estimating Root Word Vocabulary Growth in Normative and Advantaged Populations: Evidence for a Common Sequence of Vocabulary Acquisition Andrew Biemiller and Naomi Slonim University of Toronto

Root word vocabulary was studied in 2 normative samples (English-speaking, wide socioeconomic range) and in an advantaged sample. The authors estimated that in 2nd grade, the mean normative vocabulary was 5,200 root words, increasing to approximately 8,400 root words by 5th grade. During grades 3-5, the lowest quartile added about 3 root words a day, whereas the highest quartile added about 2.3 words a day. However, by 5th grade, children in the lowest quartile had only reached average 4th-grade level because they had such a small vocabulary in 2nd grade. There is evidence that words are learned in roughly the same order. The implications of these findings suggest (a) that greater efforts should be made to foster vocabulary acquisition in the primary years and (b) that a rough vocabulary curriculum sequence can be identified for the elementary years.

Increasingly, educators are becoming aware of the importance of vocabulary development as an aspect of successful educational achievement. For example, Gough and Tunmer's (1986) "simple view of reading" emphasizes the combined importance of identifying words and understanding language for successful reading. Catts, Fey, Zhang, and Tomblin (1999) reported that language as well as phonological problems must be considered when working with poor readers. Cunningham and Stanovich (1997) reported that vocabulary assessed in first grade predicted over 30% of reading comprehension variance in 11th grade. Conversely, recent studies of early reading intervention such as Pinnell, Lyons, Deford, Bryk, and Seltzer's (1994) and Gregory, Earl, and O'Donoghue's (1993) reports on Reading Recovery, and Madden, Slavin, Karweit, Dolan, and Wasik's (1993) report on the Success for All program

suggest that although these primary programs were effective at promoting word identification skills, they had no significant impact on reading comprehension. Some years ago, Becker (1977) made a similar observation about his own program, DISTAR. He noted that as children progressed beyond second grade, the vocabulary levels of school texts left some competent "readers" (decoders) unable to successfully comprehend grade-level reading material in third or fourth grade. Chall, Jacobs, and Baldwin (1990) reported similar observations in a working-class sample in the later elementary years. Vocabulary development has been of interest since ancient times. In relatively recent years, there has been considerable debate about the size and rate of development of vocabulary, and about how vocabulary is acquired (e.g., see Anglin, 1993; Beck & McKeown, 1990; Curtis, 1987; McKeown & Curtis, 1987; Nagy, Herman, & Anderson, 1985; Nagy & Herman, 1987; and White, Graves, & Slater, 1990). Anglin's recent (1993) monograph provided a particularly careful estimate of vocabulary growth, making clear distinctions between root words (which must be learned), derived words (semantic variations of root words), inflections (syntactic variations), and compounds. Derived, inflected, and compound words may be understood if the root word is known and the relevant semantic or syntactic modification is also known. Thus if plan is understood (as a verb), plan (as a noun), planning, planned, unplanned, and so on may also be understood. Anglin reported growth in root word vocabulary from an average of about 3,100 root words in Grade 1 to about 7,500 root words in Grade 5. Anglin reported evidence that children need to learn about twice as many words as the number of root words, based on observations of overt derivation of word meanings. We suspect that this exaggerates the number of words that must be learned (i.e., what Anglin calls "psychologically basic" words) and that many derived or inflected words may be understood directly without any observable process of derivation. In support of this hypothesis, we note that about two thirds of the derivatives of the root words used in the studies reported herein are reported as "known" by Dale and O'Rourke (1981) at the same or younger age level as the level at

Andrew Biemiller and Naomi Slonim, Institute of Child Study, University of Toronto, Toronto, Ontario, Canada. This work could not have been undertaken without the financial support of the Ontario Ministry of Education and the Social Sciences and Humanities Research Council of Canada, the encouragement of Jeanne Chall (who first urged Andrew Biemiller to study vocabulary issues) and our colleague Robbie Case, the support of Jeremy Anglin and Michael Graves, and information and help from Joseph O'Rourke. We particularly appreciate Janet Astington's comments on a draft of this article. Cheryl Zimmerman spent many, many hours scanning the Living Word Vocabulary into Excel and hand-correcting many scanning problems. Many graduate students contributed to the development of the sentence context test format, notably Hadley Koltun, Rose Nauta, and Amy Sirota. The final version of the test was refined and administered by Saundra Biemiller, Cathy Boote, Lara Chebaro, Lisa Harrower, Kelly Heffernan, Cecilia Kwon, Nancy Pilateris, and Lynn Tan with assistance from numerous other students. Special thanks must be given to the teachers and children from participating schools in the Waterloo Catholic District School Board and the Laboratory School of the Institute of Child Study, University of Toronto. Without their patience and helpful support, this study could not have been completed. Correspondence concerning this article should be addressed to Andrew Biemiller, University of Toronto, Institute of Child Study, 45 Walmer Road, Toronto, Ontario M5R 2X2, Canada. Electronic mail may be sent to [email protected]. 498

VOCABULARY DEVELOPMENT

499

which the root was identified. (This analysis, conducted by Naomi Slonim, will be the subject of a separate paper.) Consequently, we have chosen in this research to focus on the growth of root word vocabulary. Many writers—including Chall (1983/1996), Curtis (1980), Curtis and Longo (1999), and Gough and Tunmer (1986)—have emphasized the parallel importance of basic reading mechanics (decoding skills) and a growing vocabulary for increasing reading or language comprehension. However, although there has been a renewed emphasis in recent years on insuring the provision of decoding skills (e.g., Allington & Woodside-Jiron, 1998), there has been no comparable educational emphasis on the development of language and vocabulary (Adams, 1990; Becker, 1977; Biemiller, 1999b; Graves, Juel, & Graves, 1998). This is surprising considering the evidence provided by Beck, Perfetti, and McKeown (1982); McKeown and Curtis (1987); and reviews by Beck and McKeown (1990); Graves et al. (1998); and Stahl (1999) that various approaches to enhancing vocabulary have strong effects on increasing reading comprehension. Note that once students become fluent readers—Chall's Stage 3 and above—researchers should really be talking about language comprehension rather than reading comprehension. Prior to this stage, comprehension of oral language is usually at a higher level than comprehension of printed language (Curtis, 1980; Sticht & James, 1984).

instruction. On the other hand, if the rate of word acquisition is around seven or eight a day, as Nagy et al. (1985) suggested, it is hard to imagine a curriculum that could systematically introduce that many words. Anglin's (1993) data suggest an average gain of about three root words per day (based on 365 days), or six root words per school day between Grades 1 and 5. The present study includes new information on the rate of root word vocabulary acquisition. The degree to which words are learned in the same order is a matter of practical and theoretical interest. If the order of word learning is highly predictable, it is possible to construct curriculum materials that ensure introduction to words in a developmentally useful order. A well-defined sequence of word acquisition is also of theoretical interest. The data in Dale and O'Rourke's (1981) Living Word Vocabulary (LWV) imply that there is some ordering of word learning. However, the data in LWV are basically organized around age groups rather than around levels of word knowledge in individual children. In previous studies (Biemiller, 1998, 1999a), we found strong Pearson correlations between word means in different grades, suggesting that the order of difficulty remained much the same even as overall levels of word knowledge increased. In the present studies, we examine directly the order in which words are being learned in groups with different levels of vocabulary knowledge.

The role of schooling in vocabulary acquisition has been the subject of much debate. Early (preliteracy) differences in vocabulary growth are associated with social class (Duncan, BrooksGunn, & Klebanov, 1994; Hart & Risley, 1995; McLloyd, 1998). Nagy and Herman (1987) argued that much vocabulary acquisition results from literacy and wide reading. However, obviously much vocabulary acquisition occurs before children become literate and before they are reading books that introduce unfamiliar vocabulary (Becker, 1977). Cantalini (1987) and Morrison, Williams, and Massetti (1998) both reported that vocabulary acquisition in kindergarten and Grade 1 is little influenced by school experience, based on the finding that young first graders have about the same vocabulary (Peabody Picture Vocabulary Test) as old kindergarten children. Cantalini reported the same result for second grade. On the other hand, Huttenlocher, Levine, and Vevea (1998) reported gains of about 10% of words of their test in vocabulary (and other measures including syntax, concepts, and spatial operations) during the 6-month period between October and April (high school instruction) in contrast with an increase of 3% of test words during the 6-month period between April and October (kindergarten and Grade 1, cross-sectional data). The latter study uses a much larger sample than Cantalini and Morrison et al.'s studies, but a much more restricted vocabulary test. However, overall we think it possible that children in kindergarten and Grade 1 are gaining vocabulary increases, but children start and finish at different vocabulary levels rather than acquiring a common grade level vocabulary.

To our knowledge, only one relatively comprehensive assessment has been made of words known by children. This is Dale and O'Rourke's (1981) LWV. In this study, which was carried out over more than two decades, 44,000 individual word meanings were tested to determine the grade level at which they were known by 67%-80% of children. In LWV, "known" meant a correct response on a three-choice multiple choice test. Words were assigned to levels on the basis of the lowest grade at which a sample of 200 or more children passed the words at the criterion 67%80%. Children and adolescents in Grades 4, 6, 8, 10, and 12 were included as well as college students and adults. Words were sent out to be tested with children at a particular grade in a number of schools: "To avoid sectional bias, we sent 50 tests or less of a single form to a single school district. We tested in all sizes of school systems—both rural and urban" (Dale & O'Rourke, 1979, p. 3). If the word fell within the 67%-80% range at that grade level, it was assigned that level. If the word came in above 80%, it was tried at a lower grade level, whereas if it came in below 67%, it was tried at a higher level, until the 67%-80% level was found. However, testing did not go below fourth grade, as the tests involved reading. Thus there are many words reported above 80% correct at Level 4, or fourth-grade words. Similarly, many words are reported below 67% correct at Level 12, meaning that they were not retested with college students. However, approximately 30,000 word meanings were reported known by at least 67% of children or adolescents between Grades 4 and 12.

The absolute rate of vocabulary acquisition is a critical question. As Beck and McKeown (1990) noted in their review of vocabulary development, if the rate of vocabulary acquisition is smaller than is commonly assumed, it is possible to think of instructional approaches to increase vocabulary acquisition. For example, if the average rate of word acquisition is around two to four root words a day with the rest coming from morphological extensions, it is possible to teach or include that many new words in classroom

The LWV includes both root words and derived forms. (Grammatical inflections were usually only included for irregular forms.) Part of the current research included a study of the incidence of root forms in LWV at each level. A Level 2 was created by including words known by more than 80% of children at Level 4. One of the advantages of sampling vocabulary from the LWV levels is that it is possible to estimate the numbers of words known in a relatively clearly defined and educationally relevant population of words. For example, knowing that a child understands 5%

500

BIEMILLER AND SLONIM

of the words in Webster's Third New International Dictionary of the English Language (1981), as Anglin could determine, gives teachers relatively little practical guidance in choosing vocabulary for instruction. Knowing that the same child understands half of some 2500 Level 2 root words provides substantially more useful information. Purposes of This Study The present study had several purposes. First, we wanted to determine the numbers of words at each LWV level and to estimate what proportion were root words. Second, we wanted to examine the validity of LWV's levels in terms of current populations (i.e., can we use the LWV as a reasonable guide to words likely to be known at particular levels—or worth learning?) Associated with this is an extension of LWV to Level 2, or second grade. Third, we wanted to revisit the question of vocabulary growth rates—of how many root words must be learned per day. Our fourth goal was to examine the range of individual differences in vocabulary growth rates as well as group differences between the normative and advantaged samples. Our final goal was to determine the degree to which words are learned in a similar sequence by children with different rates of vocabulary progress. In this article we describe data on the development of root word vocabulary from kindergarten to Grade 6. Data from some or all of these grades are available from three samples: (a) a normative sample of English-first-language (EFL) children with a wide socioeconomic status (SES) range tested in June, 1999; (b) an advantaged sample of mostly EFL upper middle SES children in a university laboratory school in June, 1999; and (c) a second normative sample of EFL children drawn from the same schools but given a different sample of words in November, 1999. The second normative sample was included after some unexpected results appeared in the first two samples.

General Method Organizing and Sampling Root Words: Constructing the Root Word Inventory The first step in the study was to sort the word meanings in LWV by levels and percentage correct for each meaning within the levels. This was done by scanning the entire contents of LWV into an Excel file and sorting word meanings into levels. A Level 2 was created by including all words at Level 4 that were known by more than 80% of children tested. Words known by less than 67% of 12th-grade children were omitted from Level 12. Data on the six levels are given in Table 1. The words were then

ordered by difficulty within the level rather than alphabetically. Samples of 100 words were then drawn from each level by taking every xth word where x = number of words in the level divided by 100. Each 100-word sample was then categorized according to the five categories provided by Anglin (1993): root words (monomorphemic words; e.g., closet, flop), inflected words (grammatical variations; e.g., changed, baits), derived words (one root and one or more derivational affix; e.g., mucky, stillness, talkativeness), literal compounds (words composed of two or more words, one of which is a root word and the other root or derived; e.g., payday, taxpayer), and idioms (compounds whose meaning cannot be determined from knowledge of the component words, e.g., lady's slipper, which is an orchid; eleventh hour, which is a last minute occurrence). We had 90% agreement in coding words as "root" or not. Twenty root words were then randomly drawn from each list for use in a test. About 15% of the words initially randomly selected were discarded as archaic or unsuitable [e.g., churn, ass (donkey), paraffin (wax used in canning), hep (knows answers)]. These were replaced with the next root word from the random list at a similar level of percentage correct as given in LWV. Short sentences were constructed and pretested for each word. Two forms of our Root Word Inventory were constructed from the final list of 20 root words per level. Form A consisted of the odd-numbered words and their sentences (1, 3, etc.); Form B consisted of the even-numbered words and their sentences. Thus each form had 10 words per level for a total of 60 words. One word was deleted from Form A and two from Form B when we discovered coding problems with these words. (Scorers found children's responses to three words hard to score. Responses to these three words have been removed from the calculations.) Both forms were given to one of our test populations. The correlation of children's scores for the two forms was .88 (Pearson). Test words, LWV meanings tested, and test sentences are given in Appendix A. (In Appendix A, words from Forms A and B, definitions, LWV levels, and test sentences are shown in order of mean word difficulty.)

Estimating Root Word Vocabulary Size We had originally thought that we could estimate root word vocabulary size simply by applying the observed percentage correct at each level to the estimated number of root words at that level as shown in Table 1. However, after selecting the root words, we noted that in a number of cases more than one entry was given for a root word. In some cases, the entries had similar meanings. For example, the word fish is given as a Level 2 entry meaning "a water animal" and also as a Level 2 entry meaning "to try to catch a fish." It appears that most children who know one of these meanings will also know the other. (The "water animal" was known by 97% of Level 4 children sampled, and "catch a fish" was known by 94%.) In our view, these items refer to a single root word meaning. We reviewed all of our sampled LWV root word meanings to see if there were other redundant meanings at the same level. We determined that the 120 Form A and B words covered knowledge of 138 separate word meanings—the 120 sampled and 18 other redundant meanings from the same levels. The two

Table 1 Data on Levels in the Living Word Vocabulary Level Variable

2

4

6

8

10

12

Combined levels

Total words Estimated percentage of root words Estimated number of root words Estimated number of independent root words*

4,374 58 2,537

5,257 51 2,681

6,750 38 2,565

5,600 56 3,136

4,359 50 2,180

4,162 45 1,854

30,502 49 14,953

2,207

2,333

2,232

2,728

1,897

1,613

13,010

1

Independent or nonredundant meanings are estimated to be 87% of estimated number of root words.

501

VOCABULARY DEVELOPMENT meanings offish as given earlier is one example. Additional examples are jurisdiction at Level 10, given as both "authority" and "legal power"; envelop at Level 8, referring to "surround" and "wrap up"; and buckle at Level 2, given as "to fasten" and "fastener." There is no clear trend across the levels, so we have used the overall redundancy figure. Thus if 120 sampled words address 138 LWV meanings, then the overall estimate of 14,953 root words between Levels 2 and 12 should be reduced to 13,009 nonredundant root words (87% of the total). We applied the same reduction to estimated root words at each level. This is shown in Table 1. The reduced estimated number of independent root words at each level was used to estimate total root word vocabulary known covered in LWV. The LWV omits some root words tested in Anglin's (1993) study. Of 30 root words in Anglin's vocabulary, 10 were not found in LWV. However, using data provided by Anglin, we ascertained that many of these omitted root words were not known by children in Anglin's study. If we examine the mean number of root words known in the test at each grade, we find 5.44 in Grade 1, 8.11 in Grade 3, and 13.22 in Grade 5. (These figures are slightly higher than those in Anglin, 1993, p. 64. In the Anglin monograph, some scores had been adjusted for multiple choice responses.) If only words that appear in the LWV are used, we find 5.06, 7.20, and 10.78 root words, respectively. Thus using LWV words underestimates total root words known by 8% in Grade 1, 13% in Grade 3, and 22% in Grade 5. Projected underestimates in the other grades have been interpolated to 5% in kindergarten, 11% in Grade 2, 17% in Grade 4, and 26% in Grade 6. Estimated vocabulary at each grade level was first to be determined by multiplying the estimated number of nonredundant words by percentages of test words known, and then further adjusting estimated numbers of words by correcting for LWV underestimates for specific grades, as outlined in the preceding paragraph.

Estimating Vocabulary Growth Rates Once we arrived at estimates of the total number of root words for each child, we could generate estimated root word vocabulary at each grade level sampled. This provided a cross-sectional look at the growth of vocabulary. The estimated rate of vocabulary acquisition per year is simply the difference between observed vocabulary in the higher grade and that in the lower grade. Average gains per day were obtained by dividing annual gains by 365. Similar approaches are taken over longer periods. We can assume that vocabulary growth begins around age 1. The mean age of children tested near the end of Grade 2, minus 1 (age at infancy), yields the number of years of acquisition from infancy to the end of Grade 2 (6.6 acquisition years for children in the two normative samples, 6.9 in the advantaged sample). The mean years for gaining vocabulary from the end of Grade 2 to the end of Grade 5, based on mean ages, was 3.1 years for the normative children (both studies) and 3.0 years for the advantaged children. We now turn to describing methodology and reporting basic results from the three separate studies. We then combine results from the three studies to address the main questions of this research.

Study 1: First Normative Sample

dren were not included because of a major province-wide testing program being conducted at the same time with all third- and sixth-grade children.) The sampling system involved identifying in each school two boys and two girls in each grade whose birthdays were closest to March 30th, and those whose birthdays were closest to Sept. 30. Parents of these children were contacted for permission to include the children in the study. When permission was not given, we moved to the child of the same sex with the next closest birthday to the target. The intent was to identify children randomly, while ensuring a representation of younger and older children in the study. In fact, we were able to obtain between 19 and 25 children per grade. Table 2 gives number of children by sex and mean age by grade in the first normative sample.

Method Testing reported herein was conducted in May and June of 1999. Testing was done individually for children in kindergarten through Grade 2. The investigator introduced herself and said I'm going to ask you what some words mean. Later, I'll ask you to read the same words and some others. Some of these words are common and some aren't. If you want to stop at any time, you can. I'm going to read you a sentence and then ask you what a word means in that sentence. You can use words, pointing, or acting to explain the meaning of that word. We then gave three examples, including, " 'I bought a new car.' In this sentence, what does car mean?" We continued with " 'Johnny fell and broke his arm.' What does arm mean?" If the child had difficulty with these examples, we would go over them and explain how to answer. (In the arm example, pointing to one's arm would do.) A final example involved a hard word: " "The material was translucent.' In this sentence, what does translucent mean?" This example was given so that we could explain that there would be words children did not know, and that it was alright to say, "I don't know." (No children actually knew translucent.) At all levels, nearly all children could do the first two examples (with help) and were successful with some items in the test. In kindergarten to Grade 2, the tester wrote down student explanations for coding later. Because written testing was used in Grades 4 and 5, no probing was done in oral tests with children in kindergarten to Grade 3. We would reread a sentence if the child asked us to, on the assumption that those taking the written test could reread as needed. In oral testing, children were stopped if they made eight errors at a level. (There were 10 words at each level, 60 words total.) In the case of written testing (given to children in Grades 4 and 5), the instructions were as follows:

Table 2 Normative Sample: Mean Age and Numbers of Cases by Grade and Gender

Children Studied This study was intended to provide data on the root word vocabulary of a representative sample of English-speaking children from a range of economic levels. Three schools in a mid-sized Ontario city were included. These drew on children from assisted housing (about one third of the children in one school and one fifth in another) and from working-class and middle-class families. Within each school, we attempted to include eight children in Grades kindergarten, 1,2, 4, and 5. (Third- and sixth-grade chil-

Grade Variable

K

1

2

4

5

Age (years) Gender Boys Girls Combined

6.0

7.0

7.8

10.1

11.0

13 10 23

10 10 20

Note.

11 10 21

K = Kindergarten.

11 8 19

11 14 25

Total

56 52 108

502

BIEMILLER AND SLONIM

Read the whole sentence. What does the word in bold mean? Write a sentence that explains what the word in bold means. It's OK to write "I don't know" if you don't know what the word means. Some of these words are used by students in high school, so we don't expect you to know all of the words you see. The examples were printed with sample answers (including "I don't know" for translucent). Written testing was conducted in groups of 6 to 10 children at a time. If a child experienced difficulty reading the sentences, we would read them to the child. (Very few children required this assistance.) In a prior study (Biemiller, 1998), we had found that there was no difference in levels of word knowledge in Grades 4 and 5 when assessed in oral versus written (with reading assistance) formats. Each response was scored 1, .5, or 0. A score of 1 meant that in the scorer's judgement, the child knew the word well enough to be able to follow a story in which the test sentence could appear. (That is, they would not lose the meaning of the story because of this word.) A score of 0 meant that in the scorer's judgement, the child did not know the word this well. A score of .5 meant that the scorer was unsure of whether the child knew the word well enough to understand a story containing the sentence. Scoring guides for each word were developed and are available from the authors. Two raters scored all children in one class. The total scores for each child by each rater correlated at r = .98. For the final data presented in this article, scoring was done by three trained raters. All questionable cases were resolved together.

The percentage of words was lower in the higher LWV levels (Table 3). Overall, level differences were significant, F(5, 490) = 367.58, p < .001. The Duncan range is 6% (Duncan, .05). There was little difference between Levels 4 and 6 and between Levels 8 and 10. There was a significant interaction between level and grade, F(20, 490) = 6.55, p < .01. This reflects variation in differences between levels at different grades and particularly the fact that in Grades 4 and 5, Level 6 proved slightly easier than Level 4.

Study 2: Advantaged Sample Study 2 was intended to provide data on the root word vocabulary knowledge of advantaged upper-middle-class children attending a university laboratory school, in contrast to the normative population of Study 1. Given the evidence of large differences in vocabulary and vocabulary acquisition opportunities associated with social class before school (e.g., Hart & Risley, 1995), we thought it useful to look at vocabulary growth in a highly advantaged sample. In addition, we were able to administer both forms of our Root Word Inventory to children in Study 2, providing information on reliability.

Results The percentages of words were larger in the higher grades (Table 3). The overall gain was significant, F(4, 98) = 58.29, p < .001, as were differences between grades of 18% or more (Duncan, .05). Note the relatively large gain between Grades 1 and 2. There were no significant sex differences, F(l, 98) = 0.67, ns, and no significant interaction with sex, F(4, 98) = 1.60, ns. The unusually large increase in percentage of words learned between Grades 1 and 2 (from 17% to 35% of all words tested), combined with the relatively low rate of vocabulary acquisition after Grade 2 (from 35% to 57% of all words tested, or about 7% a year) led us to consider a replication. This was carried out in November, 1999.

Children Studied Children in Study 2 attended a private, laboratory school at the University of Toronto. A tuition of $4,000 is charged at this school, limiting the sample to children from relatively affluent families. (There is no test criterion for admission. Children are randomly selected from applicants, subject to diversity and sex distribution constraints.) Between 20 and 22 children per grade were available from junior kindergarten (4-year-old children) to Grade 6. Table 4 shows the mean age and distribution of boys and girls by grade.

Table 3 Normative Sample: Mean Percentage Correct by Grade and Level for Form A (Study 1) Word level Grade

Age lonths)

2

4

6

8

10

12

M

72.2 2.8

46 21

19 14

7 12

3 6

0

0 0

12

84.0 3.0

56 24

23 16

16 16

3

3 9

0

9

0

17 10

M SD

21

M SD

22

M SD

22

93.4 5.5

78 15

44 21

37 19

28 17

22 16

3 5

35 13

M SD

23

121.6 4.1

82 18

50 24

52 22

42 15

35 19

10 13

45 14

M SD Weighted average (K—5)

20

131.8 3.3

94 8 71

65 21 40

70 16 36

48 16 25

44 12 21

21 14

57 10 33

Note.

Grade 3 children were not tested. K = kindergarten.

0

7

7

503

VOCABULARY DEVELOPMENT

Table 4 Ages and Gender of the Advantaged Sample: Study 2 Grade Variable

JK

SK

1

2

3

4

5

6

Age (years) Gender Boys Girls

4.9

6.1

6.9

7.9

9.0

9.9

10.9

11.8

11 10 21

10 10 20

All Note.

10 12 22

10 10 20

11 11 22

11 8 19

12 10 22

12 10 22

Total

87 81 168

JK = junior kindergarten; SK = senior kindergarten.

Method In this study, both forms of the Root Word Inventory were used in Grades 1 to 6. Only Form A was used with junior (4-year-old) and senior (5-year-old) kindergarten children. Procedures were the same as in Study 1. Most testing was done in June, 1999. However, because of schedule problems, Grade 2 children were tested on Form B in September (when they were just beginning Grade 3) rather than June. Similarly, six Grade 3 children were tested in September rather than June. There was no evidence that children tested in September had higher scores.

Results The advantaged sample was tested on two forms of 60 words each in Grades 1 to 6. The two forms yielded similar results: The Pearson correlation between children's percentage-correct scores on Form A and Form B was .88 (n — 126). We again found that percentages of words known were higher in higher grades (Table 5). For combined forms (Grades 1-6), the overall gain was significant, F(5, 114) = 16.73, p < .001. Mean differences of about 19% were different using the Duncan .05 criterion. Note again the very large gain between Grades 1 and 2. There was no significant sex difference. The interaction between grade and sex was significant, F(5, 114) = 2.56, p < .03. This reflects the very low achievement of the Grade 4 boys. Form A was somewhat easier than Form B, at 49% versus 45% correct overall, F(l, 114) = 38.63, p < .001. There was no significant interaction between grade and form, F(5, 114) = 1.79, p = .12. There was a low-order interaction between form and sex, F{\, 114) = 3.51, p = .06. This reflects the fact that girls did slightly better on Form B compared with boys. The mean percentage correct at each Dale and O'Rourke level by each grade are shown in Table 6. Although scores were noticeably higher than the normative sample in the earlier grades, the pattern was similar, with relatively little difference between Levels 4 and 6 and between Levels 8 and 10. (Direct comparisons between the samples are discussed later.) Using a repeatedmeasures analysis of variance (ANOVA), differences between levels were highly significant, F(5, 760) = 734.01, p < .001 for Form A, and F(5, 570) = 517.84, p < .001 for Form B. For Form A, the Duncan .05 range was 3% compared with 4% for Form B. There was the same pattern of significant differences seen with Form A except that Level 12 was not significantly lower than the preceding levels. There were significant Grade X Level interactions, F(35, 760) = 6.85, p < .01 for Form A, and F(25, 570) = 5.04, p < .01 for Form B. Inspection of Table 7 suggests

that in both groups, the size of the difference between Levels 2 and 4 was larger in the lower grades, whereas the size of the difference between Levels 6 and 8 was larger in the later grades for Form A and the earlier grades for Form B. Overall, there were effectively four levels: 2, 4/6, 8/10, and 12. Study 3: Second Normative Sample Study 3 was conducted to see whether the anomalous large gains observed in both the normative and advantaged samples in Grade 2, followed by relatively slower gains in Grades 4 and 5, could be replicated. At the same time, it was possible in Study 3 to include third- and sixth-grade children from a normative sample. There were two other changes. One was the use of oral testing at all grade levels, to see whether the lower rates of gain after Grade 2 might have been due to use of written testing. The other was to modify the conditions for stopping testing. In Studies 1 and 2, oral testing was stopped if a child missed 8 out of 10 items in a level. We thought this might have led to underestimating the performance of kindergarten and Grade 1 children (although Grade 2 children were tested using the same rule). In Study 3, all children were tested until they missed 10 consecutive items.

Children Studied The children were drawn from the same three schools used for Study 1. No Study 1 children were included in the Study 3 sample.

Table 5 Mean Percentage Correct by Gender and Form: Advantaged Sample Grade Gender and form All children Form A Form B Boys, all Form A Form B Girls, all Form A Form B Note.

JK

SK

13

23

13

20

13

Grades 1-6

1 31 31 30 32 32 32 29 29 29

43 46 40 43 47 39 43 45 41

51 53 48 51 54 47 51 52 49

47 50 44 39 42 36 57 59 54

55 57 53 53 55 51 57 59 54

58 59 56 59 62 56 57 57 57

JK = junior kindergarten; SK = senior kindergarten.

47 49 45 47 49 44 49 50 47

504

BIEMILLER AND SLONIM

Table 6

Words Tested

Advantaged Sample: Percentage Correct by Mean and Level for Forms A and B Word level Grade JK (n = 22) Form A M SD SK (n = 20) Form A M SD 1 (n = 22) Form A M SD FormB M SD 2 (n = 22) Form A M SD Form B M SD 3 (« = 22) Form A M SD Form B M SD 4 (n = 22) Form A SD Form B M SD 5 (n = 21) Form A M SD Form B M SD 6 (n = 20) Form A M SD Form B M SD Average (Grades 1-6) Form A Form B

Age (months)

2

58.7 3.6

40 21

6

8

19 12 14 14

4 7

2 7

0 0

72.7 3.5

70 31 24 12 19 14 18 11

3 8

0 23 0 9

82.8 3.0

77 46 34 19 6 20 17 18 13 12

3 31 7 11

80 41 39 11 7 16 20 17 11 14

4 30 9 11

94.8 2.6

4

10 12 M

Method 13 9

94 58 52 34 27 11 46 9 19 23 17 15 12 13 85 57 52 23 15 7 40 11 17 15 14 19 10 19

107.9 4.2

98 66 59 41 37 17 53 4 21 15 12 23 15 11 88 62 56 29 28 22 48 11 16 9 16 23 19 11

118.6 3.7

88 67 59 42 36 11 50 16 25 25 22 21 13 16 78 61 50 26 29 23 44 14 20 17 17 20 22 15

130.2 3.6

Form B of the Root Word Inventory was used. This form was constructed at the same time as Form A, using every other word generated in sampling. Two words were changed from the version used in Study 2 because of ambiguities in scoring the original items.

94 73 71 41 45 17 57 4 17 15 16 17 15 10

Procedures were the same for all Study 3 children as those described for kindergarten to Grade 2 in Study 1, except that testing was not stopped until a child made 10 consecutive nonresponses. In other words, all children were tested individually and orally. The test took approximately 30 min per child.

Results As shown in Table 8, the percentage of correct responses were similar to those in Study 1. Overall, grade differences were significant, F(6, 142) = 43.53, p < .001. Sex differences were not significant nor was the interaction between grade and sex. The largest grade-to-grade gain was again seen between Grades 1 and 2 (from 21% to 37% correct). Overall, in comparing grades, differences of more than 15% were significant (Duncan, .05). Changes in test procedure may have slightly increased observed vocabulary levels in kindergarten (12% in Study 1 versus 17% in Study 3) and Grade 1 (17% vs. 21%). However, these changes did not eliminate the unusually large gains seen between Grades 1 and 2 (35% in Study 1 vs. 37% in Study 3). There was also no evidence that oral testing led to higher estimates of vocabulary. In fact, the level reached by Grade 5 is slightly lower than that seen in Study 1, consistent with the slightly lower levels seen on Form B in Study 2 and with the fact that these children were assessed in November rather than June. Results mostly similar to Study 1 were seen when comparing levels. Overall, differences between levels were highly significant, F(5, 710) = 829, p < .001, as was the interaction between levels and grade, F(30, 710) = 6.52, p < .001. The Duncan range for levels is 3%. Unlike Study 1, Level 6 did differ significantly from Level 4. Otherwise, the pattern was similar, with Level 8 and Level 10 being of similar difficulty.

87 72 52 38 43 24 53 13 13 13 13 14 18 10 141.8 5.0

93 77 75 50 44 17 59 7 18 20 18 21 11 12 85 69 57 48 49 31 56 9 21 19 19 10 17 12 91 65 58 38 33 13 49 84 60 51 29 28 19 45

Note. Grades JK and SK children were not given Form B. JK = junior kindergarten; SK = senior kindergarten.

Sampling was again based on ages, in this case starting with birthdates in January and June, and seeking 4 boys and 4 girls at

each grade level in each school. Table 7 gives the mean age numbers of cases by sex for each grade in Study 3.

Comparisons Involving All Three Studies In this section, we are going to return to the main questions of this research (as discussed in the introduction). Table 7 Second Normative Sample: Mean Age and Numbers of Cases by Grade and Gender Grade Variable

K

1

2

3

4

5

6

Age (years) Gender Boys Girls Grade, all

5.5

6.5

7.4

8.5

9.5

10.5

11.6

11 13 24

11 9 20

Note.

11 11 22

8 10 18

K = kindergarten.

14 11 25

13 16 29

11 7 18

Total

79 77 156

VOCABULARY DEVELOPMENT Table 8

Second Normative Sample: Mean Percentage Correct by Grade and Level (Study 3) Word level Age (months)

2

4

6

8

10

12

M

11

65.6 3.8

62 18

17 13

14 12

3 7

4 6

3 5

17 8

18

78.2 2.5

68 21

28 14

18 12

3 5

8 11

1 21 3 7

M SD

25

89.0 3.6

76 54 17 17

38 15

19 24 12 15

11 37 10 10

M SD

29

102.2 4.3

81 13

53 43 17 15

21 26 12 14

12 39 9 10

M SD

18

114.4 5.0

81 60 11 20

24 11

30 15

17 44 10 10

24

126.0 3.5

88 11

73 54 33 12 12 17

31 14

15 49 8 8

20

138.7 3.0

90 74 57 41 40 10 18 15 18 16

17 53 10 12

81

12 41

Grade

SK M SD 1 M

n

SD

2 3 4 5 M

SD

6

M SD Weighted average

Note.

57

52 16

44

24

27

SK = senior kindergarten.

How Valid Is the LWV? We examined the validity of LWV in two ways. First, we related observed word accuracy to LWV levels. We calculated the mean scores for each word in our test by grade. These are given in Appendixes B and C. We combined these for Grades 1, 2, 4, and 5 (for which we have data from all studies) and correlated these word scores with LWV levels. For the normative samples, the Pearson correlations were .73 for Form A and .68 for Form B. For the advantaged sample, the correlations between combined Grade 1, 2,4, and 5 word means and LWV levels was .79 for Form A and .72 for Form B. Overall, LWV word meaning ratings account for 46% to 62% of observed root word inventory score variance. Second, we looked at the validity of LWV in terms of absolute levels of performance on particular levels in specified grades. LWV words assigned a given level were reported to be "passed" by children in that grade at an average of 73% on three-alternative multiple choice items. We can compare mean root word inventory scores for the normative sample at each level with the percentage expected for children in that grade when the reported mean percentage is adjusted for guessing: Corrected percent correct = observed percent correct — [percent error/(number of alternatives — 1)]. This adjustment brings the expected level score down to 60%. When giving definitions, Grade 4 normative sample children from Study 1 actually were 50% correct on Level 4 words on Form A, whereas those from Study 3 were 60% correct on Level 4 words on Form B. There were no Grade 6 children for Form A, but Grade 5 children were 70% correct on Level 6, whereas Grade 6 children in Study 3 were 57% correct on Level 6.

505

Overall, these data are generally consistent with values expected from LWV. However, inspection of Appendixes B and C makes it clear that many individual words are known at substantially higher or lower values than those predicted by LWV. These data also support Level 2 as based on LWV words at Level 4 that were known by more than 80% of children in the LWV norming. Grade 2 children were 76%-78% correct on the Level 2 words in the two normative studies. (We do not have data on multiple choice meanings at Grade 2.) At Grade 4, these Level 2 words would be expected around 87%. In fact, we observed scores of 81% in both forms. Whereas in general, words from the higher levels proved to be more difficult, it is not reasonable to expect that children in a grade will know all words up to that level. Furthermore, Levels 4 and 6 were very similar in difficulty, as were Levels 8 and 10. Thus although LWV provides a better guide to identifying vocabulary words for curriculum than any other source, we will see that a more precise sequence may exist than that given in LWV.

Comparing Vocabulary Growth in the Three Studies Percentages correct by grade for both forms for all populations are shown in Figure 1. In all groups, there was an unusually large gain between Grades 1 and 2. Between the end of Grade 2 and Grade 6, we found relatively slower growth in root word vocabulary, with the advantaged children starting at somewhat higher levels. The estimated root word vocabulary by grade is shown in Table 9 for both forms and both normative and advantaged children. In addition, combined estimates based on Forms A and B are also shown. Although the advantaged sample had a 20% larger root word vocabulary at the end of Grade 2, the normative sample appeared to have "caught up" by the end of Grade 5, the difference then being only 3%. At this point we consider gains from 1 year of age (when vocabulary begins to be acquired) to the end of Grade 2, and gains

NORM.A NORM.B »OV_A »DV_B

Figure 1. Percentage correct by form for normative and advantaged samples. NORM_A = normative population, Form A vocabulary data (Study 1); NORM_B = normative population, Form B vocabulary data (Study 3); ADV_A = advantaged population, Form A vocabulary data (Study 2); A D V B = advantaged population, Form B vocabulary data (Study 2).

506

BIEMILLER AND SLONIM

Table 9 Estimated Vocabulary, by Sample and Form Normative population Grade

Advantaged population

Study 1, Form A

Study 3, Form B

M

Study 2, Form A

M SD

1,699 916

2,349 1,134

2,924

3,173 1,201

M SD

2,371 1,341

2,967 984

2,669

M SD

5,048 1,829

5,301 1,511

5,175

Study 2, Form B

M

4,314 1,576

4,275 1,477

4,295

6,591 1,815

5,722 1,492

6,157

7,805 1,695

6,988 1,677

7,397

K

M SD

5,759 1,492

M SD

6,889 2,185

6,699 1,506

6,794

7,569 2,441

6,758 2,254

7,164

M SD

9,038 1,653

7,784 1,246

8,411

9,024 1,513

8,345 1,621

8,685

9,736 2,040

9,248 1,934

9,492

M SD

8,737 1,903

Note. Vocabulary is estimated by multiplying the percentage of words known for each form and grade by 13,010 and then further adjusting the total obtained by 1.05 (kindergarten), 1.08 (Grade 1), 1.10 (Grade 2), 1.13 (Grade 3), 1.17 (Grade 4), 1.22 (Grade 5), and 1.26 (Grade 6) to allow for words not included in the Living Word Vocabulary but known by children at different grades, based on Anglin (1993). K = kindergarten.

during Grades 3-5. Gains per day for these two periods are shown in Table 10. The overall conclusion is that children from the normative sample appear to acquire root words at an average rate of 2.2 root words per day from 1 year of age to the end of Grade 2 and a greater rate—2.9 words per day—during Grades 3-5. The advantaged sample appears to have gained root words more rapidly from infancy to Grade 2 (2.4 words per day). After Grade 2, the advantaged sample actually appears to gain words more slowly through Grade 5 (2.3 words per day).

Individual Differences: Vocabulary Growth by Quartiles The quartile sample sizes in the three studies are relatively small when taken grade by grade. However, we have constructed a table summarizing averaged estimated vocabulary scores (combining data from Forms A and B for the normative and advantaged groups) for children in each vocabulary quartile in Grades 1 to 5 (Table 11). Daily gains in root word vocabulary from infancy to the testing age in Grade 2 were estimated by dividing the vocabulary in Grade 2 by (mean age - 1) and then dividing the average

Table 10 Growth in Estimated Root Word Vocabulary: Infancy to Grade 2 and Grade 3 to Grade 5 Normative population

Advantaged population

Variable

Age 1 (infancy) to Grade 2

Grade 3 to Grade 5

Age 1 (infancy) to Grade 2

Grade 3 to Grade 5

Total gain Years of gaina Annual gain Daily gain in root words

5,175 6.6 784 2.2

3,236 3.1 1,044 2.9

6,157 6.9 892 2.4

2,528 3.0 843 2.3

Note- This data is based on data in Table 9. Vocabulary is estimated by multiplying the percentage of words known for each form and grade by 13,010 and then further adjusting the total obtained by 1.05 (kindergarten), 1.08 (Grade 1), 1.10 (Grade 2), 1.13 (Grade 3), 1.17 (Grade 4), 1.22 (Grade 5), and 1.26 (Grade 6) to allow for words not included in the Living Word Vocabulary but known by children at different grades, based on Anglin (1993). * Years of gain equals mean age in Grade 2 minus 1 for infancy to Grade 2, and mean age in Grade 5 minus mean age in Grade 2 for Grade 3 to Grade 5.

507

VOCABULARY DEVELOPMENT

Table 11 Estimated Vocabulary by Quartile Group and Grade for Normative and Advantaged Samples Based on Combined Data From Forms A and B

Grade

1 2 4 5 Infancy-Grade 2 a : gain/day Grade 3-5": gain/day

Average

76
51%-75%

26%-50%

0%-25% Norm

Adv

Norm

Adv

Norm

Adv

Norm

Adv

Norm

Adv

1,122 3,000 4,293 6,614

2,425 4,491 4,045 6,888

2,219 4,838 6,071 8,006

3,700 5,366 6,664 7,924

3,106 5,787 7,476 8,777

4,716 6,304 7,912 9,094

4,030 7,129 9,083 10,065

6,367 8,132 9,980 10,749

2,669 5,175 6,794 8,411

4,295 6,157 7,164 8,685

1.3

1.8

2.0

2.1

2.4

2.4

3.2

3.2

2.2

2.4

3.0

2.1

2.9

2.4

2.5

2.8

2.3

2.4

2.9

2.3

Note. Sample size is about 10 children per normative quartile group and 5 children per advantaged quartile group. Estimates based on data in Table 9. Vocabulary is estimated by multiplying the percentage of words known for each form and grade by 13,010 and then further adjusting the total obtained by 1.05 (Kindergarten), 1.08 (Grade 1), 1.10 (Grade 2), 1.13 (Grade 3), 1.17 (Grade 4), 1.22 (Grade 5), and 1.26 (Grade 6) to allow for words not included in the Living Word Vocabulary but known by children at different grades, based on Anglin (1993). Norm = normative sample; Adv = advantaged sample. a Years of gain equals mean age in Grade 2 minus 1 for infancy to Grade 2, and mean age in Grade 5 minus mean age in Grade 2 for Grade 3 to Grade 5. Ages for specific quartiles were used.

annual gain by 365 days. Daily gains in root word vocabulary from Grade 2 to Grade 5 were estimated by dividing the difference between estimated vocabulary at the end of Grade 2 and vocabulary at the end of Grade 5 by (mean age in Grade 5 — mean age in Grade 2) and then dividing this value by 365 days. Some surprising results can be seen in Table 11. The observed differences between the highest and lowest vocabulary quartiles in Grade 5 were about the same size as in Grade 2. (This conclusion is based on cross-sectional data and is thus somewhat suspect.) Overall, the average magnitude of difference between the highest and lowest quartiles was around 3,900 root words in Grade 2 and was slightly smaller in Grade 5 (3,700 root words). (The difference between the highest and lowest quartiles was greater in Grade 4 and differed more between the normative and advantaged samples: 4,800 for normative and 5,900 for advantaged.)

achievement group. Mean scores for each of these groups of words were calculated for each ability group of children. Evidence can be seen in Figure 2 that words are learned in a roughly fixed order and that at any given level of overall word knowledge there are two or three deciles of root words at the 25%-74% correct range. The groups of children knowing only 3%-10% and ll%-20% of all words in the test mainly knew words from the first two or three sets of words (from the best known decile, the second best known decile, etc.). The group with knowledge of 45% of words overall knew over 75% percent of words in the first four deciles of words. Those knowing 55% or 65% of the words knew over 75% of the first five or six deciles of words, respectively. Overall, these descriptive data strongly sug-

lee

1*=^

Sequence of Word Difficulty The mean percentage correct for each word in Forms A and B by children in each available grade for the normative and advantaged samples are given in Appendixes B and C. In both appendixes, the words were ordered by average difficulty in Grades 1 to 5 (omitting Grade 3) because data were available for these grades in all samples. Pearson correlations between word means in different grades and populations are given in Appendixes D and E. Correlations between average word means for the normative and advantaged population were .96 for Form A and .91 for Form B. In general, these correlations imply that children are learning words in largely the same order. Data in support of this conclusion are presented in the next section. Within each 58- or 59-word form, each group of 6 words represents a decile of 1,300 of the 13,000 estimated Level 2 to Level 12 nonredundant root words in LWV, ranging from the easiest to the hardest. (Only 5 words were used to estimate deciles in the cases of the first decile for Form A and the first and last deciles for Form B.) Achievement groups were based on overall performance on the vocabulary test: 0%-10%, ll%-20%, and so forth. Children from different grades could be included in the same

_ •

\

\

\

-

\

>

\

'

-

-

\

\

x

\

•V

\

\

\

\ \ \

\ \

\

\

\ \

.

\

\

\

\ ^

\

GRP65 6RP55 GRP45

;

\

\

\

\

\

\

\

\

\

\

»--

\

\

N

GRP35 GRP25 GRP15

v

GRP»5 1

Z

3

4

5

Rord D e c i l e

6

7

8

9

10

Group

Figure 2. Percentage of words known at each cumulative level of vocabulary for vocabulary ability groups from 5% of total vocabulary to 65% of total vocabulary. Children knew the following percentages of words tested: GRP05 = 0%-9%; GRP15 = 10%-19%; GRP25 = 20%-29%; GRP35 = 30%-39%; GRP45 = 40%-49%; GRP55 = 50%-59%; GRP65 = 60%-

508

BIEMILLER AND SLONIM

gest that children are acquiring vocabulary in a relatively predictable order.

Discussion There are four main findings of interest from our research. First, the data showed a strong relationship between average knowledge of words in our study and LWV levels, with over 52% of average word score variance associated with LWV levels in the normative sample and 62% of average score variance associated with LWV levels in the advantaged sample. This provides substantial evidence of the validity of Dale and O'Rourke's (1981) work for current populations. Second, in the normative population, we estimate that on average, children had acquired about 5,200 root words by the end of Grade 2, or about 2.2 words per day from 1 year of age. During Grades 3 to 5, children in the normative sample gained an average 3,200 additional root words, or about 2.9 words per day. In the advantaged population, more root words had been acquired by the end of Grade 2—6,200 words, or 2.4 words per day. However, after Grade 2 the advantaged children gained an additional 2,500, or about 2.3 words per day by Grade 5. Third, a comparison of quartile groups in different grades indicates that large differences in root word vocabulary had occurred by Grade 2, with the mean for the lowest quartile being 4,100 less than the mean for the highest quartile in the normative sample and 3,600 less in the advantaged sample. Most differences between children with relatively low vocabulary compared with those with relatively high vocabulary had occurred by the end of Grade 2. After Grade 2, the rate of word acquisition varied from 2.1 to 3.0 root words per day, with lower quartile normative children gaining the most between Grades 3 and 5. Fourth, the study shows that individual words are known in much the same order in all grades from kindergarten to Grade 5 (except that there are a number of words not known at all in kindergarten and Grade 1). This implies that words are being learned in a fairly fixed order. Examination of words known by specific vocabulary achievement groups further supports the view that words are learned in roughly the same order. Thus we can predict roughly which words a child is likely to learn next on the basis of the total number of words known by that child. Several points are discussed in this section, including (a) limitations to using LWV levels for sequencing vocabulary curriculum, (b) the observed rapid increase in vocabulary in Grade 2, (c) the declining difference between the normative and advantaged populations, (d) the significance of individual differences before and after Grade 2, and (e) the significance of a consistent sequence of word acquisition. Educational implications are also discussed, including (a) the implications of differences in the number of words learned by Grade 2, (b) the implications of vocabulary size and gains after Grade 2, and (c) the educational implications of the sequence of word acquisition. Limitations to Using LWV Levels for Sequencing Vocabulary Curriculum As can be seen in Appendix A, words ordered by difficulty as observed in all of our studies do not correspond closely to Dale and O'Rourke's levels, especially in the middle range of word diffi-

culty (ranging from the third to eighth deciles in word difficulty). The numbers of words from each of Dale and O'Rourke's levels in each decile of observed word difficulty in our Studies 1 and 3 are given in Table 12. Words from the middle deciles incorporate words from Level 4 to Level 10. The fifth to sixth deciles of word difficulty included predominantly words from Dale and O'Rourke's Levels 4 and 6, whereas the eighth and ninth deciles of difficulty predominantly included words from Levels 8 and 10. Deciles 4 and 7 included words from many LWV levels. We suggest two reasons why LWV levels have limited value for identifying a curriculum vocabulary sequence. First, children were ordered by grade rather than vocabulary achievement level. As can be seen in Appendixes F and G, children in a given grade were in at least three achievement groups. Thus, grade levels are weak indicators of whether a word will be known. Second, use of three alternative multiple choice items in LWV allowed fairly frequent guessing of correct responses. Finally, in general, we believe that our assessment of word meaning is more accurate than multiple choice data. However, it is possible that in some cases our test sentences may have been unreasonably easy or difficult. For practical purposes, children in the third and fourth grades might focus on root words from Level 4 and 6 words and encounter some from Level 8 and 10 words. In Grades 5 and 6, children might focus on words from Levels 8 and 10, while continuing to encounter some additional Level 4 and 6 words. This is based on the percentages of words known at different levels in Studies 1 and 3. The problem is determining which words from given levels should be used. Unfortunately, we can only suggest common sense as one reviews words from different levels that occur in books used with children in different grades. We suggest keeping track of what words have been introduced (in text and with definitions as needed). By Grade 6, have most of the Level 4 and 6 words been introduced? Have perhaps 50% of the Level 8 and 10 words been introduced? In total, this would involve introducing about 6,900 root words (based on Table 1). However, even children in the low quartile knew 3,000 words by the end of Grade 2, including roughly 1,000 words above Level 2. To bring the lowest quartile to

Table 12 Numbers of Words by Word Difficulty Deciles by Dale and O'Rourke's Levels: Data From Combined Forms A and B for the Normative Population Level Decile

2

4

6

8

10

12

Total

0%-10% ll%-20% 21%-30% 31%^K)% 41%-50% 51%-60% 61%-70% 71%-80% 81%-90% 91%-100%

8 7 4 0 1 0 0 0 0 0

1 3 5 1 2 5 2 1 0 0

1 2 1 3 5 3 2 2 1 0

0 0 2 2 2 2 0 4 4 3

0 0 0 5 2 1 3 4 4 1

0 0 0 1 0 1 5 1 3 7

10 12 12 12 12 12 12 12 12 11

20

20

20

19

20

18

117

All

Note. Three words were omitted because they were changed with different samples.

VOCABULARY DEVELOPMENT average levels, they would need to pick up another 1,500 words a year. As seen in Table 11, cross-sectional data suggest that children in the lowest normative quartile gained about 900 words a year during Grades 3 to 6. What is needed is an increase of 600 root words a year over what we see now. This conclusion is based on small samples and should only be seen as suggestive. However, the present state of the art includes no estimates of actual numbers of root words needed and no guidance on what words might be included.

The Observed Rapid Increase in Vocabulary in Grade 2 In three samples, we have found a rapid increase in vocabulary in Grade 2. Other researchers have not reported unusually large gains in Grade 2. (An exception may be White et al. (1990), who found large gains between Grades 1 and 2 and between 2 and 3.) What might account for this finding? Three possible answers are discussed. Methodology. Our method requires children to give word meanings orally for words presented in a sentence context. (This method or variants of it have been used with adults; e.g., Hazenberg and Hulstijn, 1996). We are not aware of other instances where this "sentence context" approach has been used with young children. Could children simply become much more able to give verbal descriptions of word meanings in Grade 2, as part of the general cognitive changes known at this age? Note that somewhat different words seem to be being mastered in Grade 2 by children in the normative and advantaged samples. (See Appendixes B and C.) We are currently undertaking a study comparing vocabulary levels as assessed by Peabody picture vocabulary methodology and levels as assessed by sentence context methods. We will be contrasting children's ability to identify words using multiple pictured alternatives with a sentence approach with the same Peabody items. Reading. Clearly one difference is that by the end of Grade 2, most children are reading to some degree. We believe that it is possible that when encountering an unfamiliar word while reading, students may be able to pause and consider meanings in a way that they cannot when hearing unfamiliar words in context. However, if reading alone explained the gains in Grade 2, we would expect continued relatively large gains in Grades 3 to 6. This is simply not the case. Cognitive-developmental change. Age 7 has long been recognized as a change point in human development (e.g., Case, 1985; Flavell, 1992; Piaget, 1971). Following Case's (1985, 1992) cognitive-developmental theory, which emphasizes the role of increasing working memory capacity, we believe it is possible that around age 7, children become more capable of asking questions about words they do not know. (We will elaborate on this point in a forthcoming paper.) There is evidence (Beals, 1997) that in conversations, children 5 and younger rarely ask about words they don't understand. (Conversely, any parent knows that young children often ask about objects and actions for which they don't have word; e.g., "What's that?" or "What are you doing?") Using an interview procedure with Grade 5 and 6 children, Biemiller (1999a) found that children report that most words recently acquired were learned either as a result of asking about them or of adult-initiated instruction. We hypothesize that increased cognitive capacity at age 7 may make inquiring about words easier. Briefly,

509

we suggest that increased working memory capacity may allow a child to attend to problems with words without having to "forget about" the story context. Increased capacity may also make learning words from adult instruction easier. This is a hypothesis that can be investigated relatively easily.

The Declining Difference Between the Normative and Advantaged Populations During the Elementary Years By the end of Grade 2, the advantaged children had about 20% more root words (see Table 11). (The difference was larger in Grade 1 and kindergarten, but that is suspect for the reasons just discussed.) The early large difference presumably reflects the effects of an advantaged environment and possibly greater experience with defining or explaining words. As described in Hart and Risley's (1995) book on the experiences of young American children, advantaged children are exposed to many more words than average or disadvantaged children. They participate in more complex verbal discussions. It appears that these differences in experience lead to marked differences in vocabulary development. However, there also appears to be a "catch-up" process underway during the elementary school years, with less advantaged and lower quartile children adding root word vocabulary somewhat more rapidly than their advantaged peers during the Grade 2 to Grade 5 period. By Grade 5 there was no meaningful difference between the normative and advantaged groups in estimated root word vocabulary (3%). It is possible that if more words known largely by collegeeducated people were included (as was the case in Anglin's study from 1993), greater differences between the samples would have appeared. It is also possible that if morphologically complex words were included, differences in favor of the advantaged sample—or the highest quartiles in both populations—would have appeared. However, at present it appears that in root word growth there is a definite "catch-up" phenomenon, with normal children catching up to advantaged children. The critical factor in this catch-up is the apparent relatively lower rate of root word acquisition in the middle elementary years by the advantaged population in our study. It is possible that this finding is unique to the school studied or that the additional words being acquired by these advantaged children are simply not those tested here.

The Significance of Individual Differences Before and After Grade 2 Although there appears to be a catch-up between advantaged and normative populations as a whole, the difference between the highest and lowest quartile groups remains much more constant between Grades 2 and 5. If children of differing vocabulary size can acquire similar numbers of words after Grade 2, it seems likely that they could acquire new words at a similar rate earlier. However before Grade 2, as Becker (1977) observed, there is little effort to introduce hundreds of new words in the primary grades.

The Significance of a Consistent Sequence of Word Acquisition There is a marked similarity in mean scores for words in the different populations and grades when ability and word difficulty

510

BIEMILLER AND SLONIM

are controlled. (See Figure 2 and Appendixes F and G.) What can account for these stabilities? Methodology. It is certainly possible that in some cases, the sentences we used to create "context" for the words may have made the item particularly easy or difficult. Word frequency. Word frequency is usually involved to account for differences in word knowledge (e.g., see references in White et al., 1990). Carroll, Davies, and Richmond (1971) provided substantial data on word frequencies in schoolbook English. However, Carroll et al.'s frequencies account for relatively little variance in word knowledge in this sample of words. The highest Pearson correlation obtained, r = .26, was calculated between SFI and the average scores for words in Grades 1, 2, 4, and 5. (SFI is a logarithmic index of word frequency in Carroll et al.). This accounts for 7% of the observed Root Word Inventory variance. In contrast, the LWV levels account for 52% of root word variance. Adding the frequency index to LWV accounts for no additional Root Word Inventory variance. On the other hand, advantaged population word means account for 83% of normative population word means. Again, neither LWV levels nor the frequency index adds any significant variance to this association. Why does frequency have so little effect? Part of the answer lies in the varying meanings of words. Some of the meanings that we used based on random sampling from LWV were uncommon uses of common words [e.g., beat (wings), bit (computer information), or tree (rack for shoes, hats)]. These words have high frequencies in print, but not in the meanings used. Frequencies of word meanings rather than word forms might lead to better predictions (but would be very hard to produce). It is also possible that some words have different frequencies in oral use than in print. It is clear that factors other than print frequency account for most variation in word knowledge as identified by both LWV and our Root Word Inventory. Sequential pattern made clearer by cross-age ability grouping. The existence of a strong sequence in word acquisition is not surprising. However, it is important to note that when data are grouped by achievement level rather than by grade (as LWV does) or age (as Peabody and Wechsler do), sequential patterns become clearer. Note in Appendixes F and G that in the present data, specific "achievement levels" typically include children from three or four grades. Are there underlying developmental factors explaining the order of word acquisition? Certainly some of the words rarely known even by Grade 6 may involve complex meanings requiring further cognitive advances. Possible examples might be Reformation, locomotion, and oligarchy. However, there are other words virtually unknown by sixth graders that may simply not have been experienced. Examples might be valor, parch, and destitute. These and others appear to be concepts that could be understood at younger ages. Frankly, at this point, we believe that there is a stronger observed sequence of word learning than can be explained solely by cognitive constraints, as can be seen in Appendix A. We wish we could provide a better explanation for the clear sequence we see. We can be clearer about what does not explain this sequence (e.g., print word frequency, cognitive constraints) than about what does explain this observed sequence of word acquisition.

The Educational Implications of These Findings Each of these findings has considerable significance for educational practice. Implications of word acquisition by Grade 2. At the end of Grade 2, children in the lowest quartile had 2000 fewer root words than the average. Feitelson and her colleagues (Feitelson, Goldstein, Iraqi, & Share, 1991; Feitelson, Kita, & Goldstein, 1986) have demonstrated that it is possible to increase rates of vocabulary and language acquisition significantly prior to Grade 2. By far the simplest way to reduce vocabulary differences apparent in Grade 2 or 3 would be to make some effort to foster vocabulary growth in the preschool and early primary years. Implications of word acquisition after Grade 2. It is important to note that children in the normative sample appear to be catching up to the advantaged sample between second and fifth grade. However, the absolute difference in root word vocabulary between the least and most advanced quartiles remains quite large in all grades in both the normative and advantaged samples. The lowest Grade 5 quartiles in both populations had a root word vocabulary around the level of Grade 4 children near the median (see Table 11). Thus schooling appears to allow some catching up across social classes but may not reduce differences between those who started school with large vocabularies and those who did not. Various studies have shown that it is possible to increase rates of vocabulary acquisition in the middle-grade years (see Stahl, 1999, for a summary). It remains to be seen whether low-quartile children can come closer to average fifth- or sixth-grade achievement. It may also be necessary to recognize, as Chall and Conard (1991) suggested, that children who are lagging in vocabulary development may need reading materials that are somewhat more restricted in their vocabulary demands, especially for "independent" (unassisted) reading. Either there must be variation in the vocabulary demands of books read by different children, or the range of ages in classes must vary somewhat more so that children have more similar vocabulary achievement. To what extent a somewhat looser definition of ages appropriate to grades would help remains a subject of much debate. However, these data would be consistent with allowing a wider variation of ages for particular grades, as suggested in Biemiller (1993). Implications of a sequence of word acquisition. Root words appear to be learned in roughly the same order by most children. (We also have preliminary evidence suggesting that the same is true even for children whose first language is not English; Biemiller, 1999a). To some degree, this word order can be predicted from LWV levels. This suggests that a plausible vocabulary curriculum sequence can be established empirically. It becomes reasonable to define bodies of vocabulary that should be largely known by certain grades, or better, by points in a curriculum (through which students may not progress at the same rate). This may be especially important before children are reading—in kindergarten and Grade 1. For example, it appears that by the end of Grade 2, average children understand most Level 2 words, about half of Level 4 words, and about a third of Level 6 words from the LWV. It would be worth trying to bring most children to this level of vocabulary achievement. On a vocabulary test, we may infer that they are familiar with most of the words up to an identifiable range of words, are likely to be learning another identifiable range of words, and probably

VOCABULARY DEVELOPMENT need relatively little emphasis on words more difficult than that second range of words. As they progress, we can identify ranges of words that should then be introduced.

A Final Note We began this article by observing that although vocabulary development is crucial for school success, it has not received the attention and interest that work on identifying printed words and spelling have received. The findings presented herein support the need to create a more systematic approach to facilitating vocabulary development in the schools. If we are roughly correct, an average rate of about 3 root words per day (6 if all words were to be learned at school) is a potentially manageable task. Evidence that all children are adding words at similar rates after Grade 2 suggests that with greater educational focus on vocabulary, children could start adding words at a similar level from kindergarten or earlier. Evidence that there is a well-defined sequence to what words are learned makes the problem of planning a vocabulary curriculum more realistic. A move to a more teacher-centered approach to vocabulary development is a needed complement to the move to increased focus on phonics skills in reading that has occurred in the past decade.

References Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press. Allington, R. L., & Woodside-Jiron, H. (1998). The politics of literacy teaching: How "research" shaped educational policy. Educational Researcher, 25(8), 4-12. Anglin, J. M. (1993). Vocabulary development: A morphological analysis. Monographs of the Society for Research in Child Development, 55(10, Serial No. 238). Beals, D. (1997). Sources of support for learning words in conversation: Evidence from mealtimes. Child Language, 24, 673-694. Beck, I., & McKeown, M. (1990). Conditions of vocabulary acquisition. In R. Barr, M. L. Kamil, P. B. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research, (Vol. 2, pp. 789-814). New York: Longman. Beck, I. L., Perfetti, C , & McKeown, M. G. (1982). Effects of long-term vocabulary instruction on lexical access and reading comprehension. Journal of Educational Psychology, 74, 506-521. Becker, W. C. (1977). Teaching reading and language to the disadvantaged—What we have learned from field research. Harvard Educational Review, 47, 518-543. Biemiller, A. (1993, December). Lake Wobegon revisited: On diversity and education. Educational Researcher, 22, 7-12. Biemiller, A. (1998, April). Oral vocabulary, word identification, and reading comprehension in English second language and English first language elementary school children. Paper presented at the annual meeting of the Society for the Scientific Study of Reading, San Diego, CA. Biemiller, A. (1999a, April). Estimating vocabulary growth for ESL children with and without listening comprehension instruction. Paper presented at the annual conference of the American Educational Research Association, Montreal, Quebec, Canada. Biemiller, A. (1999b). Language and reading success. Cambridge, MA: Brookline Books. Cantalini, M. (1987). The effects of age and gender on school readiness and school success. Unpublished doctoral dissertation, Ontario Institute for Studies in Education, Toronto, Ontario, Canada.

511

Carroll, J. B., Davies, P., & Richmond, B. (1971). The American Heritage word frequency book. Boston: Houghton Mifflin. Case, R. (1985). Intellectual development: Birth to adulthood. New York: Academic Press. Case, R. (1992). The mind's staircase: Exploring the conceptual underpinnings of children's thought and knowledge. Hillsdale, NJ: Erlbaum. Catts, H. W., Fey, M. E., Zhang, X., & Tomblin, J. B. (1999). Language basis of reading and reading disabilities: Evidence from a longitudinal investigation. Scientific Studies of Reading, 3, 331-361. Chall, J. S. (1996). Stages of reading development. (2nd ed.). New York: Harcourt Brace. (Original work published 1983) Chall, J. S., & Conard, S. S. (1991). Should textbooks challenge students? New York: Teachers College Press. Chall, J. S., Jacobs, V. A., & Baldwin, L. E. (1990). The reading crisis: Why poor children fall behind. Cambridge, MA: Harvard University Press. Cunningham, A. E., & Stanovich, K. E. (1997). Early reading acquisition and its relation to reading experience and ability 10 years later. Developmental Psychology, 33, 934-945. Curtis, M. E. (1980). Development of components of reading skill. Journal of Educational Psychology, 72, 656-669. Curtis, M. E. (1987). Vocabulary testing and vocabulary instruction. In M. G. McKeown & M. E. Curtis (Eds.), The nature of vocabulary acquisition (pp. 37-52). Hillsdale, NJ: Erlbaum. Curtis, M. E., & Longo, A. M. (1999). When adolescents can't read: Methods and materials that work. Cambridge, MA: Brookline Books. Dale, E., & O'Rourke, J. (1979). The living word vocabulary: The words we know. Boston: Houghton Mifflin. Dale, E., & O'Rourke, J. (1981). The living word vocabulary. Chicago: World Book/Childcraft International. Duncan, G., Brooks-Gunn, J., & Klebanov, P. (1994). Economic deprivation and early childhood development. Child Development, 65, 2 9 6318. Feitelson, D., Goldstein, Z., Iraqi, J., & Share, D. I. (1991). Effects of listening to story reading on aspects of literacy acquisition in a diglossic situation. Reading Research Quarterly, 28, 70-79. Feitelson, D., Kita, B., and Goldstein, Z. (1986). Effects of listening to series stories on first graders' comprehension and use of language. Research in the Teaching of English, 20, 339-356. Flavell, J. H. (1992). Cognitive development: Past, present, and future. Developmental Psychology, 28, 998-1005. Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading and reading disability. Remedial and Special Education, 7, 6-10. Graves, M. F., Juel, C , & Graves, B. B. (1998). Teaching reading in the 21st century. Boston: Allyn & Bacon. Gregory, D., Earl, L., & O'Donoghue, B. (1993). A study of reading recovery in Scarborough: 1990-1992 (Publication No. 92/93-15). Scarborough, Ontario, Canada: Scarborough Board of Education. Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore: Brookes. Hazenberg, S., & Hulstijn, J. H. (1996). Defining a minimal receptive second-language vocabulary for non-native university students: An empirical investigation. Applied Linguistics, 17, 145-163. Huttenlocher, J., Levine, S., & Vevea, J. (1998). Environmental input and cognitive growth: A study using time-period comparisons. Child Development, 69, 1012-1029. Madden, N. A., Slavin, R. E., Karweit, J. L., Dolan, L. J., & Wasik, B. A. (1993). Success for all: Longitudinal effects of a restructuring program for inner-city schools. American Educational Research Journal, 30, 123-148. McKeown, M. G., & Curtis, M. E. (Eds.). (1987). The nature of vocabulary acquisition. Hillsdale, NJ: Erlbaum. McLloyd, V. C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53, 185-204.

512

BIEMILLER AND SLONIM

Morrison, F. J-, Williams, M. A., & Massetti, G. M. (1998, April). The contributions of 1Q and schooling to academic achievement Paper presented at the Annual Meeting of the Society for the Scientific Study of Reading, San Diego, CA. Nagy, W. E., & Herman, P. A. (1987). Breadth and depth of vocabulary knowledge: Implications for acquisition and instruction. In M. G. McKeown & M. E. Curtis (Eds.), The nature of vocabulary acquisition (pp. 19-36). Hillsdale, NJ: Erlbaum. Nagy, W. E., Herman, P. A., & Anderson, R. C. (1985). Learning words from context. Reading Research Quarterly, 20, 233-253. Piaget, J. (1971). Piaget's theory. In P. H. Mussen (Ed.), Carmichael's manual of child psychology (3rd ed., pp. 703-732). New York: Wiley.

Pinnell, G. S., Lyons, C. A., Deford, D. E., Bryk, A. S., & Seltzer, M. (1994). Comparing instructional models for the literacy education of high-risk first graders. Reading Research Quarterly, 29, 9-38. Stahl, S. A. (1999). Vocabulary development. Cambridge, MA: Brookline Books. Sticht, T. G., & James, J. H. (1984). Listening and reading. In D. Pearson (Ed.), Handbook of research on reading (Vol. 1, pp. 293-317). New York: Longman. Webster's third new international dictionary of the English language. (1981). Springfield, MA: Merriam. White, T. G., Graves, M. F., & Slater, W. H. (1990). Growth of reading vocabulary in diverse elementary schools: Decoding and word meaning. Journal of Educational Psychology, 82, 281-290.

513

VOCABULARY DEVELOPMENT

Appendix A Root Words From Forms A and B Showing Words Ordered by Mean Difficulty, Meanings Selected, Living Word Vocabulary (LWV) Level, and Test Sentences LWV level

Test sentence

Word

Meaning tested

First decile (0-1,300 words) Form A: Grades 1, 2, 4, 5: 91%-79% fish flood throat match cafe Form B: Grades 1, 2, 4, 5: 93%-87% spread

A water animal Unusual flow of water Passage from stomach to mouth Thing to light fire Eating place

2 2 2 2 2

Johnny caught a fish. The flood caused a lot of damage to the town. He felt a lump in his throat. Where is the box of matches? She met him at the cafe.

To distribute (over a surface, as in buttering or painting) Injection Close Sound from mouth Stick knife into

2

Use a knife to spread the jam.

6 2 2 4

Go to the doctor to get your shot. The drugstore was near the hospital. Stephanie's voice could be heard from far away. He pretended to stab his friend.

Worthless things Take number from another Finished doing To expect and wish for Moved through the air" Cry and scream

6 2 2 2 4 4

Take that stuff outside. Subtract the smaller number from the bigger number. I've done all my work. I hope that it rains today. The ball had flown a great distance. The baby made a fuss.

A circled string To try to hear Fall Strike at a ball Keeps, keep School subject

2 2 2 4 2 6

He made two loops with his shoe laces. You should listen to your mother. The ball dropped from his hand. Jamie took his second swing at the ball. He kept his old hockey trophies. John got his math work done quickly.

Tiny piece of wood Big iron hook To fasten Large rock Free from fear Exactly

8 4 2 2 6 4

He got a sliver in his foot. He threw the anchor. Buckle your seatbelt. The boulder was too heavy to move. When there is war people do not feel secure. I want just the right dress.

Room Dark spot cast by light New, not spoiled Look shyly Rolled up bit Fair dealing

4 2 2 4 8 4

There is space for only three passengers. He was afraid of his own shadow. A fridge keeps food fresh. She peeped around the corner. She gave me a wad of tissue paper. The criminal was brought to justice.

shot near voice stab Second decile (1,301-2,600 words) Form A: Grades 1, 2, 4, 5: 77%-67% stuff subtract done hope flown fuss Form B: Grades 1, 2, 4, 5: 85%-74% loop listen drop swing kept math Third decile (2,601-3,900 words) Form A: Grades 1, 2, 4, 5: 66%-59% sliver anchor buckle boulder secure right Form B: Grades 1, 2, 4, 5: 72%-54% space shadow fresh peep wad justice Fourth decile (3,901-5,200 words) Form A: Grades 1, 2, 4, 5: 57%-55% cobra tally react thud drama blab Form B: Grades 1, 2, 4, 5: 5Wc-Al% dodo through matting haul gull litter

Snake Count Act backb Dull sound Plays Tell secret

6 8 10 8 6 10

The cobra lived in the house. The teacher kept a tally of days missed. When the cat saw the mouse, she didn't react. There was a thud in the next room. She enjoys watching drama productions. He made a promise not to blab.

Extinct bird From start to end A straw fabric Distance travelled Bird Disorder

10 4 12 10 6 10

The dodo is extinct. The test continued through the day. The floor matting was frayed. It was a long haul from Mexico to Canada. The gulls were eating the food. The room was littered with empty cups.

(Appendix continues)

514

BIEMILLER AND SLONIM

Appendix A (continued) Word Fifth decile (5,201-6,500 words) Form A: Grades 1, 2, 4, 5: 51%-42% stock peeve shimmer straight mammoth root Form B: Grades 1, 2, 4, 5: 41%-34% Vaseline parcel possum distant transit man

Sixth decile (6,501-7,800 words) Form A: Grades 1, 2, 4, 5: 42%-24% know thigh because beat astronomy victim Form B: Grades 1, 2, 4, 5: 31%-23% knoll period envelop garble duplex lash Seventh decile (7,801-9,100 words) Form A: Grades 1, 2, 4, 5: 23%-16% tree polo guard former dibs induct Form B: Grades 1, 2, 4, 5: 21%-11% text nation ominous narrow writhe curious Eighth decile (9,101-10,400 words) Form A: Grades 1, 2, 4, 5: 13%-8% alias vice robust rotary swoon junction Form B: Grades 1, 2, 4, 5: 10%-5% lust character cartilage matron delinquent whittle

Meaning tested

Supply, provide Annoy Faint gleam Direct Huge Dig around Petroleum jelly Package Animal Long ago Public transportation Humankind

LWV level

6 8 10 4 6 10

6 6 4

6 8 2

Recognize Upper part of leg For the reason that Flap wings About stars Injured person

4 6

Small hill A time in history Surround To mix up Two homes in one Fasten with rope

10 4 8 12

Rack for shoes, hats Game played on horseback A defense First of two Next claim to To bring in

12

Schoolbook A country Threatening Lacking a broad view

6 4 12 10

Twist about Odd, strange

12 10

False name Evil habit Strong and healthy Wheel-like motion Faint Joining

10 10 8 8 12

Strong desire Nature of

10

Tough tissue Married woman Breaks laws Cut wood with knife

4 4

4 6

8

6

4

6 10 12 12

6

8 8

10 6

4

Test sentence

The boy stocked the shelves in the grocery store. My brother peeves me. The moonlight shimmered on the lake. The dog went straight home. It was a mammoth building. She rooted around in her basement for the sleeping bag. The jar of Vaseline is on the shelf. The parcel was delivered to the office. He saw a possum. The year my mother was born seems distant to me. The children took transit to school. Man has always had trouble with the weather. He knows that person. The thigh is Marty's favorite part of the chicken. She went home because she was sick. The robin beat its wings. You use telescopes in astronomy. She was the victim. The boy ran up and over the knoll. The period of the dinosaurs was very interesting. Fog enveloped the city. He garbled his words. He lives on one side of a duplex. Lash the suitcase to the roof of the car. The guests hung their hats on the coat tree. They were watching polo. Keep your guard up. The former Prime Minister spoke to the people. I have dibs on the car when she gets home. I wonder if the Baseball Hall of Fame will induct John Doe. She had a text about history. He had lived in two nations. There were ominous shadows in the alley. Many people have a narrow understanding of the pollution issue. The caterpillar writhed about. What a curious thing to say! When asked for his name, he gave an alias. Gambling is a vice. Construction workers must be very robust. The phone had a rotary dial. The sight of blood caused her to swoon. The junction links the two highways. Their lust for battle was strong. Difficult times in life may show the true character of a person. She suffered from torn cartilage. The matron came to the social tea. The delinquent arrived at the courthouse. He whittled a stick.

515

VOCABULARY DEVELOPMENT

Appendix A (continued) Word

Ninth decile (10,401-11,700 words) Form A: Grades 1, 2, 4, 5: 7%-3% bit franchise sequence inquisitive vain popular Form B: Grades 1, 2, 4, 5: 5%-2% jurisdiction perpendicular empty republic discord cow Tenth decile (11,701-13,000 words) Form A: Grades 1, 2, 4, 5: 3%-l% etch question valor parch destitute cognac Form B: Grades 1, 2, 4, 5: \%-O% locomotion lance abrasive Reformation oligarchy

Meaning tested

LWV level

Test sentence

Item of computer data Chain of businesses Connect in series Curious Conceited Representing the people

12 10 8 8 10 12

There are 8 bits in 1 byte. He bought a restaurant franchise. The sequence of events was surprising. The inquisitive scientist made many observations. People do not like her because she is vain. He was elected by popular vote.

Authority

10

Upright position Without meaning Representative government Clash of sounds Female of species

8 10 8 12 6

The police officer had no jurisdiction outside the city limits. Climb the perpendicular pole. The book had an empty plot. The Republic of Mexico chose not to join. The band's concert was full of discord. The whale cow met the others.

Engrave with acid Problem Courage Dry up Terribly poor French brandy

12 12 10 8 12 12

I am going to etch a metal plate in art class today. She presented a difficult question. The valor of the knight was praised far and near. I was parched after my walk. He was destitute. Leave the cognac on the table.

Ability to move Cut open Scratch material Protestant era Ruled by a few

8 8 12 12 12

His locomotion was poor. He lanced the wound. That material is abrasive. The Reformation was a long time ago. A few countries in the world are oligarchies.

Note. The meanings given were selected randomly. Many words also have other given meanings. " Not a root word. Included in test by error. b A derived word but probably learned as a basic word.

(Appendixes continue)

516

BIEMILLER AND SLONIM

Appendix B Correct Percentage of Form A Words Ordered by Normative Average for Grades 1,2, 4, and 5 Grade SK

T \X/W

Word cognac destitute parch valor question etch popular vain inquisitive sequence franchise bit junction swoon rotary robust vice alias induct dibs former guard polo tree victim astronomy beat because thigh know root mammoth straight shimmer peeve stock blab drama thud react tally cobra right secure boulder buckle anchor sliver fuss flown hope done subtract stuff cafe match throat flood fish Note.

LWV

l

2

3

4

5

6

1, 2, 4, 5(M)

level

Norm

Adv

Norm

Adv

Norm

Adv

Adv

Norm

Adv

Norm

Adv

Adv

Norm

Adv

12 12 8 10 12 12 12 10 8 8 10 12 6 12 8 8 10 10 12 12 10 6 4 12 6 4 4 4 6 4 10 6 4 10 8 6 10 6 8 10 8 6 4 6 2 2 4 8 4 4 2 2 2 6 2 2 2 2 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 10 5 10 0 5 10 0 5 0 0 3 5 0 0 3 0 0 10 24 12 17 33 29 12 62 53 48 43 8 22 48 52 72 67 69

0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 1 3 45 5 35 20 5 3 20 10 6 33 0 25 40 15 1 60 20 28 50 78 70 53 65 60 55 60 53 50 83 80 70 92 75

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 5 0 0 0 5 6 0 0 11 5 5 11 19 0 21 5 11 0 26 3 6 0 11 11 21 42 24 32 42 45 13 48 45 34 37 42 42 58 79 71 76 87

14 1 4 0 0 0 0 10 14 11 16 0 14 7 16 8 2 9 16 14 7 16 18 14 30 32 45 25 54 42 31 28 45 28 31 40 24 41 59 26 45 83 76 59 73 69 86 76 69 91 62 90 64 78 90 86 91 93 95

0 0 0 0 0 0 0 0 2 1 0 0 8 0 6 12 8 10 6 0 12 8 9 24 8 15 30 24 24 44 40 26 40 48 48 46 40 54 54 62 52 52 60 60 62 62 70 76 70 78 68 64 72 78 82 86 100 96 90

18 0 5 5 5 0 0 5 16 12 1 5 19 5 5 7 7 3 27 9 10 18 19 23 23 32 59 25 55 21 41 32 64 59 43 43 55 64 75 77 73 84 75 73 77 89 84 77 75 86 91 100 86 80 100 98 95 100 98

32 7 7 7 3 0 3 26 10 10 10 13 22 1 19 31 16 19 25 44 13 19 63 44 50 44 69 63 84 41 69 44 37 94 53 44 87 69 94 72 97 100 69 75 81 100 84 84 84 97 97 100 100 84 97 100 100 100 100

1 1 0 5 0 0 0 9 9 1 14 13 11 18 14 9 24 9 20 20 22 24 29 22 26 27 35 48 57 44 53 57 57 63 72 52 83 67 78 72 83 70 70 70 61 67 57 89 78 61 85 83 87 89 76 89 87 93 89

14 1 18 19 0 5 2 16 10 37 10 9 5 10 28 19 14 12 23 28 39 34 59 18 46 75 41 62 75 28 50 64 86 57 59 57 77 68 91 64 64 95 68 68 80 62 84 61 86 80 89 95 91 80 91 93 93 91 93

1 1 5 1 8 10 10 5 10 23 13 15 11 25 25 20 18 28 38 55 50 50 45 45 60 55 43 50 65 60 75 70 78 80 83 80 92 93 90 80 82 85 65 83 90 78 88 85 73 90 92 95 100 100 100 92 100 95 97

26 5 23 21 0 10 14 14 19 39 23 9 37 14 21 19 12 21 25 37 48 48 73 25 48 82 46 73 73 39 57 64 66 68 46 55 91 82 82 71 62 91 73 89 77 84 80 84 73 91 82 95 95 91 89 95 93 95 93

31 1 16 5 4 8 15 31 24 58 50 16 58 8 46 39 5 24 24 43 39 58 70 23 58 100 77 81 77 73 73 73 46 62 85 62 88 96 92 58 66 92 77 77 69 65 81 77 85 88 88 100 100 92 100 100 96 92 96

1 1 1 2 2 3 3 4 5 6 7 7 7 11 11 12 12 13 16 19 21 22 22 23 23 27 28 32 39 42 42 43 45 51 51 51 55 55 56 56 57 57 59 59 61 62 65 66 67 68 70 70 75 77 79 86 90 90 91

18 2 13 11 1 4 4 11 15 25 12 6 19 9 17 13 9 11 23 22 26 29 42 20 37 55 48 46 64 32 45 47 65 53 45 49 62 64 77 59 61 88 73 72 77 76 84 75 76 87 81 95 84 82 92 93 93 95 95

LWV = Living Word Vocabulary; SK = senior kindergarten; Norm = normative sample; Adv = advantaged sample.

517

VOCABULARY DEVELOPMENT

Appendix C Correct Percentage of Form B Words Ordered by Normative Average for Grades 1,2, A, and 5 Grade

SK Word oligarchy reformation abrasive lance locomotion cow discord republic empty perpendicular jurisdiction whittle delinquent matron cartilage character lust curious writhe narrow ominous nation text lash duplex garble envelop period knoll man transit distant possum parcel Vaseline litter gull haul matting through dodo justice wad peep fresh shadow space math kept swing drop listen loop stab near voice shot spread Note.

LWV

level

12 12 12 g 8

6 12 8 10 8 10 4 6 10 8 8 10 10 12 10 12 4

6 6 8 12 8 4 10 2 8 6 4 6 6 10 6 10 12 4 10 4 8 4 2 2 4

6 2

4 2 2 2 4 2 2

6 2

Norm

0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0

0 5 0

0 5 2 0 9 0 2 0 0 7 14 9 3 7 9 25 9 9 0 14 14 18 14 18 16 28 66 41 16 52 23 75 59 89 52 71 75 68 93

1

6

5

4

3

2

Norm

Adv

Norm

Adv

Norm

Adv

Norm

Adv

Norm

Adv

Norm

0 0 0 0 3 0 0 0 1 0 0 0 0 3 0 6 0 3 0 0 0 0 0 3 0 6 6 1 17 23 6 3 23 11 17 6 22 34

0 0 0 5 5 0 2 0 9 1 5 7 1 2 7 7 0 9 14 5 9 12 1 19

0 0 0 0 1 1 3 2 1 0 8 16 0 4 10 13 4 7 13 11 10 14 8 40 26 26 38 28 36 21 34 24 44 48 50 50 38 48 56 50 66 78 66 60 66 72 76 80 76 78 78 80 86 92 90 94 94 98

0

0 0 0 0 1 4 1 0 4 4 4

1 0 22 0

25 12 17 23 36 34 50 28 42 42 31 59 39 45 45 69 50 53 42 50 67 64 72 64 56 72 81 53 67 70 83 95 89 95 86 89 86 94

4 0 0 5 21 1 9 3 1 13 17 27 4 6 23 46 11 15 17 46 2 50 30 44 25 46 63 59 21 57 54 81 75 67 48 55 81 59 71 63 79 83 75 88 90 71 92 86 96 92 100 96 96 98

0

32

1 3 28 1 14 7 25 0 14 1 14 21 3 7 21 43 25 30 32 43 5 50 14 30 21 50 41 71 5 36 62 39 82 71 73 52 95 26 50 25 71 39 55 77 77 64 82 86 91 66 95 95 89 93

13 8 39 1 27 5 20

4 9 9 21 18 21 9 21 24 21 38 35 35 28 31 49 54 54 38 43 55 59 41 37 54 64 64 43 69 79 71 62 78 86 76 86 97 84 90 86 90 93 100 95

0 0 0 1 1 1 3 6 6 12 6 1 17 14 17

95 89

86 84

100 83

95 100 93 91

94 95

84 61

100 96

76 83

6 17 17 22 11 39 50 64 50 36

64 50 70 81 78 78 81 89 83 83

3

0 14 11 2 55 32 32 50 55 59 14 77 9 18 12 16 48 34 73 73 73 82 68 84 41 86 84 68 77 100 100 75 77

0 0 5 3 0 0 11 5 0

0 11 0 0 26 53 5 5 11 37 5 24 8 45 11 16 24 68 0 61 42 71 42 55 79 39 68 21 37 45 39 45 58 90 92 55 84 82 87 84 100 95 95 76 92

87 100 89

1 2 5

0 5 24 9

39 10 27 18 5 8 36 6 21 50 50 16 27 18 55 18 53 55 84 6 52 59 48 50 75 80 47 95 42 42 46 82 52 68 86 86 75 95 86 95 64 100 95 98 93 100 95 95 86

6

15 95 12 1 43 13 8 24

64 34 29 29 88 20 50 24 29 10 53 57 76 10 38 95 55 86 76 88 91 79 20 43 50 50 65 79 95 93 86 91 72 100

62 95 93

3 6 5 15 1 5 16 13 13 35 25 20 5 40 80 26 16 43 50 10 40 45 63 35 55 58 90 40 70 60 60 80 55 65 95 80 45 40 70 78 73 90 90 100 68 88 85 83 97 100 100 100 90 100 85 95 95

LWV = Living Word Vocabulary; SK = senior kindergarten; Norm = normative sample; Adv = advantaged sample. (Appendixes continue)

5 (M) 1, 2, 4,

Adv 16 0 55 16 23 5 20 28 100 21 1 60 21 1 43 70 40 20 46 90 23 73 25 43 35 55 65 75 23 45 85 48 95 80 85 90 93 40 80 36 80 58 92 80 100 95 80 90 97 45 95 90 93 87 100 73 83 63

Norm

0 0 0 0

1 2 2 2

3 4 5 5 5 7 8 8 10 11 12 14 14 17 21 23 27 27 27 29 31 34 35 38 39 39 41 41 42 46 47 49 51 54 57 58 64

67 72 74 76 77 83 83 85 87 89 91 93 93

Adv 4 3 17 3 12 3 12 7 31 3 5 21 4 4 20

42 16 18 22 43 10 34 12 31 11 30

34 56 4 47 58 49 65 64 75 49 80 19 37 33 44 49 56 84 84 69 85 77 91 63 94 92 87 86 93

91 87 78

518

BIEMILLER AND SLONIM

Appendix D Form A: Correlations Between Word Means in Different Grade Groups and With Dale and O'Rourke's Levels Grade K Grade

LWV

1

level

Norm

Adv

Norm Adv

-.69 - . 76

— .86

—

Norm Adv

-.75 -.80

.92 .81

.88 .93

_ .86

Norm Adv

-.71 -.73

.77 .73

.87 .88

Norm Adv

-.66 -.77

.62 .64

Norm Adv Average Norm Adv

-.66 -.77 -.76 -.80

Norm

2 Adv

Norm

.85 .79

.93 .94

.95

.75 .77

.71 .71

.86 .87

.57 .61

.72 .75

.67 .70

.82 .76

.88 .91

.88 .83

4 Adv

Norm

.93 .86

.93 .91

.91

.84 .86

.90 .86

.90 .90

.93 .97

.97 .94

.94 .98

5 Adv

Norm

.95 .90

.92 .96

.92

.94 .92

.91 .95

.92 .91

Average Adv

Norm

.90 .94

.96

K i l

2 4 5

Note.

Number of observations = 59. LWV = Living Word Vocabulary; Norm = normative sample; Adv = advantaged sample; K = Kindergarten.

Appendix E Form B: Correlations Between Word Means in Different Grade Groups and With Dale and O'Rourke's Levels Grade 1

2

LWV level

Norm

Adv

Norm Adv

-.70 -.75

.88

—

Norm Adv

-.63 -.74

.87 .82

Norm Adv

-.64 -.62

Norm Adv

Grade

3

Norm

Adv

.85 .90

.86

—

.85 .77

.85 .87

.95 .86

-.63 -.61

.80 .73

.80 .86

Norm Adv

-.69 -.57

.78 .63

Norm Adv Average Norm Adv

-.67 -.48 -.70 -.68

4

Norm

Adv

.90 .92

.87

—

.93 .82

.85 .89

.94 .84

.84 .77

.92 .72

.92 .81

.72 .54

.77 .69

.86 .67

.88 .78

.88 .91

.97 .85

5

Norm

Adv

.83 .94

_ .80

—

.92 .74

.90 .85

.92 .68

.91 .73

.88 .67

.89 .82

.93 .94

.98 .87

.90 .97

6

Norm

Adv

.89 .87

.79

—

.85 .64

.87 .85

.93 .75

.96 .82

.88 .96

.96 .91

Average

Norm

Adv

.80 .93

— .76

—

.77 .93

.93 .89

.71 .89

Norm

Adv

— .91

—

1 2

A

4

3

6

Note, n — 58. LWV = Living Word Vocabulary; Norm = normative sample; Adv = advantaged sample.

519

VOCABULARY DEVELOPMENT

Appendix F Normative Population: Performance of Vocabulary Achievement Groups on Words of Varying Difficulty Average of grades 1,2,' i, 5 Word decile and grade

Norm

Adv

Vocabulary ability group 0%-10%

ll%-20%

21%-30%

31%-40%

41%-50%

51%-60%

61%-70%

71%-80%

1 2 13 21 38 63 84 70 85 91

2 10 18 30 55 72 87 91 94 97

5 21 36 75 77 83 94 89 94 100

(none)

0 0 8 6 4 18

0 0 3 5 8 16

0 0 0 3 6 9

/ 5 14 20 42 56 64 77 95 95

1 8 20 42 54 79 80 86 97 96

4 16 28 58 81 83 87 93 97 100

0 0 5 10 6 11 3 35

0 0 2 3 4 7 8 24

0 0 1 0 1 3 5 10

Total

Form A: Percentages Decile 10 9 8 7 6 5 4 3 2 1

2 5 11 20 32 47 56 62 71 87

8 14 11 27 49 49 69 76 84 94

0 0 0 0 0 2 0 11 17 49

0 1 0 1 5 8 6 24 47 74

0 0 4 1 13 23 22 53 65 89

0 0 5 9 23 45 58 68 76 91 Form A: iVs

Grade K 1 2 4 5 Total

11 6 1 0 0 18

8 6 3 1 0 18

2 6 4 3 0 15

0 1 7 5 1 14

0 0 0 0 0 0

21 19 26 23 19 108

Form B: Percentages Decile 10 9 8 7 6 5 4 3 2 1

0 3 7 15 27 38 46 62 80 91

8 10 18 23 28 60 44 71 84 87

0 0 0 0 0 0 1 14 32 36

0 1 2 1 4 6 9 25 49 73

1 1 1 4 12 21 25 51 74 92

1 3 8 17 25 36 46 68 83 93

(none)

Form B: Ns Grade K 1 2 3 4 5 6 Total

4 2 0 0 0 0 0 6

10 5 1 2 0 0 0 18

1 10 5 6 1 1 0 30

1 1 11 8 6 2 3 32

0 0 0 0 0 0 0 0

22 18 25 29 18 24 19 155

Note. Grade 3 children were not given Form A: Norm = normative sample; Adv = advantaged sample; K = kindergarten. Percentages in italics are between 0 and 24. Percentages in bold are between 75 and 99. All other percentages are between 25 and 74.

(Appendixes continue)

520

BIEMILLER AND SLONIM

Appendix G Advantaged Population: Performance of Vocabulary Achievement Groups on Words of Varying Difficulty Average of Grades 1,2,- 4,5 \17__.J

*1 ^

Vocabulary ability group

"1

Word decile and grade

Norm

Adv

0%-10%

ll%-20%

21%-30%

31%-4O%

41%-50%

51%-60%

61%-70%

71%-80%

7 9 10 28 53 50 76 82 91 97

8 25 19 45 75 65 90 83 95 97

21 38 25 61 79 74 96 91 95 97

31 61 41 72 94 72 93 94 97 98

0 0 5 6 5 3 4 0 23

0 0 0 6 8 7 10 7 38

0 0 0 2 4 6 4 6 22

0 0 0 0 1 0 3 4 8

4 16 12 29 36 62 56 77 87 84

12 19 24 46 51 77 73 85 92 92

17 41 37 59 70 79 79 88 94 99

45 50 58 63 88 79 88 100 100 100

4 1 9 7 9 6 36

0 3 6 2 4 5 20

0 1 2 5 5 6 19

0 0 0 0 1 2 3

Total

Form A: Percentages Decile 10 9 8 7 6 5 4 3 2 1

2 5 11 20 32 47 56 62 71 87

8 14 11 27 49 49 69 76 84 94

0 0 1 0 4 2 / 6 21 37

0 0 0 1 10 9 5 23 42 68

; 0 1 0 17 10 26 55 61 89

; 3 2 3 29 32 51 75 82 95 Form A: Ns

Grade JK SK 1 2 3 4 5 6 Total

10 2 1 0 0 1 0 0 14

9 5 2 1 0 1 0 0 18

1 8 9 1 0 0 0 1 20

2 5 5 3 4 4 0 2 25

22 20 22 19 22 22 21 20 168

Form B: Percentages Decile 10 9 8 7 6 5 4 3 2 1

0 3 7 15 27 38 46 62 80 91

8 10 18 23 28 60 44 71 84 87

(none)

0 0 0 2 0 6 4 21 61 48

0 1 1 3 5 29 23 53 68 86

4 4 3 15 22 52 32 66 87 83 Form B: Ns

Grade 1 2 3 4 6 Total

0 0 0 0 0 0 0

3 0 0 1 0 0 4

9 2 0 2 0 0 13

6 12 2 5 2 1 28

22 19 19 22 21 20 123

Note. Junior and senior kindergarten (JK and SK, respectively) children were not given Form B. Norm = normative sample; Adv = advantaged sample. Percentages in italics are between 0 and 24. Percentages in bold are between 75 and 99. Ail other percentages are between 25 and 74. Received August 10, 2000 Revision received October 26, 2000 Accepted October 27, 2000