Running Head: RACE, MATERNAL EDUCATION, AND BUS … · Race, maternal education and Bus Story...
-
Upload
truongnhan -
Category
Documents
-
view
214 -
download
0
Transcript of Running Head: RACE, MATERNAL EDUCATION, AND BUS … · Race, maternal education and Bus Story...
Race, maternal education and Bus Story retells 1
Running Head: RACE, MATERNAL EDUCATION, AND BUS STORY RETELLS
IN PRESS (2011): Journal of Speech, Language, and Hearing Research
The effects of race and maternal education level on children‘s retells of the Renfrew Bus Story –
North American Edition
Anne van Kleeck
University of Texas at Dallas
Alissa Lange
State University of New York at Buffalo
Amy Louise Schwarz
University of Texas at Dallas
Race, maternal education and Bus Story retells 2
Abstract
Purpose: The Renfrew Bus Story North American edition (RBS-NA) is widely used clinically and
in research for determining children‘s language abilities, although possible influences of race and
maternal education on RBS-NA performance are unknown. The current study compared RBS-NA
retells of four groups of children: African American (AA) and European American (EA) whose
mothers had an education level of high school or less (<HS), or had an education level higher
than high school (>HS).
Method: Statistical tests were conducted on 163 kindergartners‘ story retells using raw scores for
all four RBS-NA measures: Information, Sentence Length, Complexity, and Independence.
Results: A 2 x 2 ANOVA indicated main effects for maternal education and race for the
Information score, with <HS and AA children scoring lower. For measures not meeting
ANOVA assumptions, 2 x 2 ANOVAs using ranked data indicated significant differences across
groups for maternal education on Sentence Length, Complexity, and Independence scores, with
<HS scoring lower. There were no additional effects of race.
Conclusion: There are systematic effects of maternal education and race on children‘s RBS-NA
performance, which is important for both researchers and clinicians to take into account when
using this instrument.
Race, maternal education and Bus Story retells 3
The effects of race and maternal education level on children‘s retells of the Renfrew Bus Story –
North American edition
Producing narratives requires the integration of a variety of linguistic and social-
cognitive skills. In the linguistic realm, producing a narrative requires coordinating semantic,
syntactic, morphologic, and phonological elements both within and across sentences. Social-
cognitive skills impact narrative production via the cultural conventions that shape the way
narratives are organized, shared, and adapted to a listener‘s level of background and shared
information on the topic. Due to this complexity, assessing young children‘s narrative skills is
considered to be both ―an ecologically valid and educationally relevant means of language
assessment‖ (e.g., Pankratz, Plante, & Insalaco, 2007, p. 390).
Narrative is important in its own right as a ―ubiquitous and universal construct used to
inform, socialize, entertain, and teach‖ (McGregor, 2000, p. 55). A number of researchers have
hypothesized that skill with narratives in preschool and the early elementary years would also be
a helpful component in later ability to comprehend connected text when reading (Anderson,
Anderson, Lynch, & Shapiro, 2003; Scarborough, 2001; Sénéchal & LeFevre, 2002; Storch &
Whitehurst, 2002; Tabors, Snow, & Dickinson, 2001; van Kleeck, 2007, 2008). Indeed, several
empirical studies have found that narrative abilities do predict later reading comprehension,
reading fluency, or written narrative skills in both children who have learning disabilities and
those who are typically developing (e.g., Feagans & Appelbaum, 1986; Griffin, Hemphill, Camp,
& Wolf, 2004; Reese, Suggtate, Long, & Schaughency, 2009; Snyder & Downey, 1991; Tabors,
et al., 2001). In a large study of over 1,500 Spanish/English bilingual children from kindergarten
to grade three, oral narratives (retells of a wordless picture book) were in fact the best predictor
of third grade reading comprehension (Miller, et al., 2006). Furthermore, narrative skills
Race, maternal education and Bus Story retells 4
uniquely contribute to reading fluency even after controlling for receptive vocabulary and
decoding skills (Reese, et al., 2009).
In the studies in which narrative ability has not predicted later reading ability (e.g.,
Menyuk, et al., 1991; O'Neill, Pearce, & Pick, 2004; Roth, Speece, & Cooper, 2002; Snow,
Tabors, Nicholson, & Kurland, 1995; Stein & Glenn, 1982), researchers have measured
children‘s reading outcomes in the second grade or earlier when word recognition still dominates
both the curriculum and tests of reading ability. Narrative ability would not be expected to
predict performance until children are required to begin engaging in independent comprehension
of longer and more complex connected text, which typically occurs in the third or fourth grade.
And indeed, using structural equation modeling, Storch and Whitehurst (2002) found that the
strong and direct link between oral language and literacy found during preschool disappeared
during the first two years of formal reading instruction, but reappeared in the third and fourth
grade. Others have also found that the links between oral language skills and reading
comprehension become stronger as children get older (e.g., Vellutino, Tunmer, Jaccard, & Chen,
2007). So, the importance of narrative ability during the preschool years appears to accrue when
the curriculum shifts from what is frequently referred to as ―learning to read‖ in kindergarten
through second grade to independently ―reading to learn‖ beginning in third grade.
When evaluating children‘s narrative productions, the complexity of the task is
acknowledged in that two levels of measurement are typically conducted, a micro- and a more
macro-level (Petersen, Gillam, & Gillam, 2008). The micro-level measures focus on such things
as word diversity and frequency and sentence level complexity. These measures are the same as
those typically used when analyzing spontaneous language samples of conversational dialogue.
So, for example, syntactic complexity is often assessed via measures such as mean length of
Race, maternal education and Bus Story retells 5
utterance in words or morphemes (MLU) and percent of complex sentences produced. The
lexical skills of a child are captured by determining the diversity of words used in the sample,
often using a measure such as the number of different words (NDW). When using these
measures, narratives are considered to be one type of language sample, one that can either be
spontaneously generated by the child or consist of a retell of a story told to the child. The
Renfrew Bus Story North American edition (RBS-NA) story retell used in the current study
captures the micro-level of narrative analysis at the sentence level with two measures. One is
called Sentence Length, and it indicates the length of the child‘s five longest sentences, the other
is called Complexity, and it indicates the amount of complex syntax produced. The RBS-NA does
not include micro-level measures at the word level.
To consider how children are able to organize their narratives at a more global level that
goes beyond the word and sentence level, macro-level measures are used. This level is typically
tapped by assessing children‘s ability to organize their narratives into the elements of story
grammar (e.g., Serpell, Baker, & Sonnenschein, 2005), or by assessing the amount of
information from the story that is included in story retell tasks. The RBS-NA contains one macro-
level measure called Information.
To date, The Bus Story Test (Renfrew, 1969) is the only norm-referenced screening
measure of young children‘s narrative abilities that spans the preschool and kindergarten age
range, a period of development during which it is important to be able to measure the language
skills foundational to later reading, and work to strengthen those that are weak. This assessment
tool was originally developed in England as the Bus Story Test, and now has a North American
version, the Renfrew Bus Story – North American edition (RBS-NA: Glasgow & Cowley, 1994)
that was used in the current study. This assessment includes a story and twelve accompanying
Race, maternal education and Bus Story retells 6
pictures, presented on four pages, about a bus that runs away from its driver. The story is read to
the child who then retells the story to the examiner. In addition to the micro- and macro- level
measures mentioned above, the child‘s retell is also analyzed for the amount of prompting
required for the child to discuss the story, a measure called Independence. Both versions of the
Bus Story have been shown to be predictors of persistent language impairment (Bishop &
Edmundson, 1987; Pankratz, et al., 2007; Stothard, Snowling, Bishop, Chipchase, & Kaplan,
1998), and the British version has also shown a strong relationship to later literacy (Stothard, et
al., 1998).
The Bus Story is widely used in research in the United States and the United Kingdom
that has focused on children who are developing their language in a typical fashion as well as
those with language impairments (e.g., Anderson, Morse, Catroppa, Haritou, & Rosenfeld, 2004;
Anderson, et al., 1997; Bishop & Edmundson, 1987; Botting, Conti-Ramsden, & Crutchley,
1997; Briscoe, Gathercole, & Marlow, 1998; Conti-Ramsden & Botting, 1999; Conti-Ramsden,
Crutchley, & Botting, 1997; Dockrell & Lindsay, 2001; Gallagher, Frith, & Snowling, 2000;
Girolmetto, Wiigs, Smyth, Weitzman, & Pearce, 2001; Gray, 2004; Hohen & Stevenson, 1999;
Hughs, Dunn, & White, 1998; Kennedy, et al., 2006; Kovas, et al., 2005; Mackie & Dockrell,
2004; Nathan, Stackhouse, Goulandris, & Snowling, 2004; Paul, 1996; Paul, Hernandez, Taylor,
& Johnson, 1996; Paul & Smith, 1993; Storch & Whitehurst, 2002; Stothard, et al., 1998). In the
United Kingdom, it is believed to be widely used clinically (Norbury & Bishop, 2003), being
―regularly employed in primary language units‖ (Botting, 2002, p. 2) that serve children with
specific language impairments (Conti-Ramsden & Botting, 1999). According to the distributor of
the North American version of the test, test sales are about 60% to Canada and 40% to the
Race, maternal education and Bus Story retells 7
U.S.A. Buyers include both speech-language pathologists and researchers in speech-language
pathology and education (Lange, 2010).
Concerns have been raised that the North American version of this assessment tool may
over-identify minority children as having poor narrative skills. In a review of this test, Haccoun
(2001) noted that African American (AA) children in the normative data appeared to obtain
lower scores, but unfortunately, information on SES is not provided in the norms. Data from
2008 shows that AA children are over three times more likely to be poor than their European
American (EA) counterparts (Federal Interagency Forum on Child and Family Statistics, 2010).
Furthermore, children living in households headed by single mothers were five times more likely
to be living in poverty than those living in households headed by two parents. For African
Americans, the poverty rate for households headed by single mothers was 50.2%, and the
number of female headed households in which there are children under 18 is approximately 2 ½
times higher for African Americans than for the total US population. Given the overall confound
of race and income, it is unclear if AA children‘s lower scores are related to race or SES.
Two key variables often used in determining SES are income and parental education. As
Dollaghan and her colleagues discuss (Dollaghan, Campbell, Paradise, Feldman, Janosky, Pitcair
& Kurs-Lasky, 1999), education level is considered to be the more stable (Huston, McLoyd, &
Garcia-Coll, 1994) and less controversial (Hauser, 1994) of the two. Furthermore, maternal
education level is the best predictor of intellectual attainment when family income, maternal
education level and paternal education level are compared (Mercy & Steelman, 1982). This may
be because, except for cases of extreme poverty, maternal education level is the component of
SES most strongly related to parenting measures (Bornstein, Hahn, Suwalsky, & Haynes, 2003;
Hoff, Laursen, & Tardif, 2002).
Race, maternal education and Bus Story retells 8
More directly tied to language ability, in a study of 701 African American preschoolers,
performance on the Preschool Language Scale-3 (Zimmerman, Steiner, & Pond, 1992) was
significantly different for different maternal education levels, but not for monthly income for
children of the 590 mothers who provided this information (Qi, Kaiser, Milan, Yzquierdo, &
Hancock, 2003). Furthermore, 81 percent of these mothers reported being single female heads of
household. Research from these various perspectives supports our use of maternal education
level as a marker for SES in the current study.
Previous research does not illuminate the potential impact of race on children‘s
performance on the RBS-NA. Although a study by Price and her colleagues (Price, Roberts, &
Jackson, 2006) used the Bus Story Test to study the structural development of AA children‘s
narratives, their data are not helpful for determining whether EA and AA children perform the
same or differently on the Renfrew Bus Story – North American edition because (a) no EA
children were studied, (b) the British version of the test was used (the North American version
was not yet available when their data were collected), and (c) these researchers developed their
own method of coding the narratives using a story grammar paradigm rather than following the
standardized analyses in the test manual.
Pankratz et al. (2007) obtained Renfrew Bus Story – North American edition retells from
racially and ethnically heterogeneous groups of preschoolers with specific language impairment
and those who were developing language in a typical fashion. When they grouped together
minority children in their sample (Hispanic, African American, and other), they found that,
compared to the EA preschoolers, the minority children scored significantly lower on the
Sentence Length score (number of words in the five longest utterances produced in the retells),
but not on the Information score (information provided by the child that corresponded to
Race, maternal education and Bus Story retells 9
information in the original story). This pattern also held when comparing just the Hispanic
children to the EA children, but there were not enough children in the other minority groups to
conduct similar analyses.
These studies just reviewed lead to potential concern regarding impact of race on
children‘s performance on the RBS-NA. We know of no research related to the impact of
maternal education level on the RBS-NA. However, there are two relevant studies that have used
other narrative measures.
Reese had her colleagues (Reese et al., 2009) explored the relationship between maternal
education level and three aspects of story book retell narratives produced by 61 six-year-old
(Study 1) and 39 seven-year-old (Study 2) children. In both studies, maternal education did not
significantly correlate with any of the oral narrative measures. It should be noted, however, that
there were no children of university graduates in this sample. This narrow range of maternal
education level may have obscured the impact of maternal education level on the children‘s
retells.
Kulkofsky and Klemfuss (2008) examined narratives of 112 preschoolers that focused on
a previously witnessed staged event. In Study 1 (n = 46), 89% of the mothers had at least a
college degree; in Study 2 (n = 66), 72% of the mothers had a college degree. There were no
main effects or interactions for maternal education, nor did maternal education correlate with
narrative quality. However, here again the range of maternal education in this study may have
been too narrow, although in this case containing too many mothers with higher education levels.
Given that previous research sheds little light on the topic, the goal of the current study
was to determine the effects, if any, of race and maternal education on kindergarten children‘s
Race, maternal education and Bus Story retells 10
retells of the Renfrew Bus Story – North American edition (RBS-NA) (Glasgow & Cowley,
1994).
Method
Participants
Data were gathered from a sample of the children participating in the large-scale research
project, Scaling-up TRIAD: Teaching Early Mathematics for Understanding with Trajectories
and Technologies, for which approval from the primary university‘s Institutional Review Board
was obtained following all local and federal guidelines for the protection of human subjects. This
study was a U.S. Department of Education-funded randomized control trial involving two urban
public school districts. District agreement to participate was obtained, and all eligible schools
within each district were then randomly assigned to Control or Treatment groups, including all
preschool classrooms in each school (the treatment took place the school year before the retell
data for the current study was collected when the children were in kindergarten). Thus, all
classrooms within each school were assigned to one condition. Children in the initial sample (n =
1305) were from 106 classrooms, in 42 schools in Buffalo, New York and Boston,
Massachusetts. On average, 81.6% of children in these schools were receiving free or reduced
school lunches, and the racial/ethnic distributions across the two school districts were the
following: 43% African American, 35% European American, 19% Hispanic or Latino, and 4%
Asian American).
Maternal education and race. The maternal education status was taken from a
question in the Parent Questionnaire of the TRIAD study. The questionnaire included
questions about parent education and employment, as well as questions about the child
relevant to the math study, such as parent-perceived child counting ability. The question
Race, maternal education and Bus Story retells 11
about maternal education from this measure asked the respondent to report the highest
level of education attained by the child’s mother by indicating one of seven categories (see
Table 1). The self-report Parent Questionnaire data were obtained by mail or by phone by a
trained interviewer during the child’s pre-kindergarten year.
Table 1 shows the number of African American (AA) and European American (EA)
children according to the seven maternal education categories used in the TRIAD study. The race
information was provided by the school districts involved. In the current study, we divided these
seven education categories to create two levels of maternal education, high school and less
(<HS), and higher than high school (>HS). However, to assure equivalence in maternal
education for the AA and EA children in the current study, we made sure that our sample had
equal numbers of children in each of the original seven maternal education categories. To do
this, we set the number of children in each of the seven maternal education categories as the
smallest number of either AA or EA children in that group in the original sample.
For example, the top section of Table 1 indicates that our search of the TRIAD database
yielded 26 AA and 13 EA children whose mothers had an eighth grade education. As shown in
the middle section of Table 1, we included all 13 of the EA children in our sample for this study.
We then randomly selected 13 out of the 26 AA children to participate in this study. We
followed this same procedure for each maternal education category until we had an equal number
of children by race in each maternal education category, as shown in the middle section of Table
1. We then collapsed the seven maternal education categories into two categories as shown in the
bottom section of Table 1, which assigned children to either the <HS or the >HS maternal
education level. Splitting the data into these two maternal education levels made sense
conceptually and, given the sample we were using, this particular split for maternal education
Race, maternal education and Bus Story retells 12
level also allowed us to have roughly equivalent sample sizes at each maternal education level.
Children excluded. We excluded children who were enrolled in special education
programs because they were not distributed proportionally in the original sample across race and
the seven maternal education categories. We also excluded control group children because there
were far fewer of them, and we did not want to have groups in which treatment and control
children were mixed, since the narrative-retell data was collected at the end of the math
intervention. This was the conservative approach to take given that in other research on a math
intervention with preschoolers (reported in Preschool Curriculum Evaluation Research
Consortium, 2008), very small effects (effect sizes ranged from d = 0.08 to 0.17) on language
abilities were found (immediately post treatment or one year later) as measured by the Peabody
Picture Vocabulary Test, Third Edition (Dunn & Dunn, 1997) or the Test of Language
Development Primary, Third Edition (Newcomer & Hammill, 1997). Our final sample for this
study included 172 children, with 86 AA and 86 EA children (see Table 2). A total of eighty-
two children fell within the <HS maternal education level, and 90 fell within the >HS maternal
education level.
Measures: Renfrew Bus Story – North American edition
The RBS-NA is a standardized, norm-referenced screening measure of oral language
using a story retell narrative language sample. The assessment involves telling a child a
story, and then asking the child to retell the story using the pictures in a wordless
storybook as prompts. The stories that children retold while taking the RBS-NA were audio
and video recorded, transcribed and scored on the four measures described in the test
manual. For transcription, utterances were segments of language that consisted of main
clauses and their attached subordinate clauses or sentence fragments. As noted earlier,
Race, maternal education and Bus Story retells 13
there are three primary measures rating the content (macro-level) and form (micro- or
sentence-level) of the story retells, including Information, Complexity, and Sentence
Length. One additional subscale, Independence, taps the child’s self-initiative in generating
the story retell.
The Information subscale is a macro-level measure of how many of the 32 key pieces of
information (including both words and phrases) from the original story the children used in their
story retell. The emphasis is on how much of the information the child uses in the retell, so
children can get credit for a response that matches the ‗key‘ in meaning, even if not in identical
word use. Still, many of the ‗key‘ words or phrases are only given full credit if they do match
exactly. The total possible raw score, 52, includes some items worth 2 points. This subscale
represents the proficiency level of a set of integrated skills. In order to score well on this
measure, children must remember the names of the key words or concepts (memory), know the
meaning of the words well enough to use them appropriately in their retell (vocabulary), and
have a sufficient understanding of story structure to use the words in the right sequence
(book/story knowledge).
Sentence Length is a straightforward micro-level measure of morpho-syntactic
complexity. For this purpose, the RBS-NA uses the average number of words in the five longest
sentences a child generates in the retell. Complexity, also a micro-level narrative measure, looks
at the number of utterances containing a subordinate clause that children use in their story retells.
It is also an indicator of syntactic development.
The Independence score relates to neither macro- nor micro-level structure of the
narrative itself, but is focused instead on the amount of prompting the child needs from the
assessor in order to retell the story. The total prompt score is the summed prompt scores with a
Race, maternal education and Bus Story retells 14
maximum of 4 possible prompts available per picture, and prompts are reverse scored. For
example, ―no prompt‖ is given 4 points, while the 4th
prompt, ―The bus…,‖ is scored 0 points.
Children who retell the story without any need for prompting from the assessor will get the
highest Independence score, while children who need prompts will score lower on Independence.
The Independence measure was scored as directed in the manual.
Procedure
Assessors were trained prior to administering the RBS-NA. Assessors included research
assistants and retired school teachers. Training consisted of a lecture on the importance of strict
adherence to protocol, viewing practice assessments, and real-time test administration practice
with feedback.
Children were individually administered the RBS-NA in the fall semester of kindergarten;
the math treatment had occurred during the prior school year in pre-kindergarten. They were
assessed in an open space (e.g., library, hallway) in their school by a trained assessor.
Assessments were videotaped and audio taped to facilitate transcription. Audiotapes served as a
backup in case of videotape malfunctioned, and also aided in the transcription process when the
videotapes were difficult to hear.
Reliability. The reliability of the transcription and scoring procedures was calculated for
the full sample (n = 1027). Reliability of the transcription procedure was determined by
evaluating a one-way, single-measure interclass correlation (ICC) using a random 7% sample of
stories that were transcribed by two trained research assistants. The two sets of transcriptions
were then expert scored by the same researcher, and the ICCs between the pairs of scores on
three of the measures, which represented the reliability of the transcription process from the tape,
were 0.99 for Information, 0.97 for Sentence Length, and 0.94 for Complexity. Transcription
Race, maternal education and Bus Story retells 15
reliability for the Independence score was not determined, because this was an objective
recording of the type of prompt that was provided and recorded by the assessor during test
administration. Scorers/transcribers checked 100% of the transcriptions for the accuracy of the
Independence scores given by the assessors. Errors were made in less 1% of the time, and these
were corrected.
Inter-rater reliability for the scoring procedure was calculated using ICCs to determine
scoring agreement between two different scorers using the same transcriptions (a random 10% of
samples for Information, Sentence Length, and Complexity, and a random 8% for
Independence). A one-way ICC with single measures were used to account for error due to
having more than two raters, but not always the same two raters rating each target. Agreement
was 0.97 for Information, 0.94 for Sentence Length, 0.81 for Complexity, and 0.94 for
Independence.
Results
Box plots were used to identify outliers for each dependent measure. Specifically, data
points were considered outliers if they were beyond the first and third quartiles by 1.5 times the
interquartile range (Hinkle Wiersma, & Jurs, 2003). The box plots indicated that each dependent
measure included two to three outlier scores. A total of nine children produced the outlier scores
across all four dependent measures. These nine children were distributed relatively equally
across the four groups. For ease of reporting, as shown in Table 2, we removed these same nine
children from the analysis of each dependent measure; so our analysis is based on an n = 163
with 82 African American children and 81 European American children. Seventy-nine children
fell within the <HS category while 84 fell within the >HS category.
We first wanted to know whether our results were due to differences in gender and site of
Race, maternal education and Bus Story retells 16
testing (Buffalo vs. Boston). Only the Information score was normally distributed so we
conducted a two-tail independent t-test on the Information score variable and two-tail Mann
Whitney Tests on the remaining three measures. Table 3 contains the results of the gender
analysis and Table 4 contains the results of the site analysis. The sample contained 80 males and
83 females (eliminating any outliers as discussed above). There were no significant differences
for gender for any of the variables. The sample contained 133 children from Buffalo and 30
children from Boston. There were no significant differences for site for any of the variables. As
such, we proceeded in our planned analyses collapsing over both gender and site.
Two-factor, fixed-effect, analyses of variance (ANOVA) with two levels for the
independent variable race (AA or EA) and two levels for the independent variable maternal
education (<HS and >HS) were planned for four measures in the RBS-NA test manual
(Information, Sentence Length, Complexity, and Independence). A power analysis indicated that
an experimental effect could be detected only if the independent variables explained at least
3.86% of the variance of a given dependent measure (Abdi, Edelman, Valentine, & Dowling,
2009). For main effects, pooled variances to determine Cohen‘s d effect sizes were calculated
according to Cortina and Nouri (2000) and confidence intervals were calculated according to
Altman (2000). We followed Cortina and Nouri‘s (2000) recommendation to only report
Cohen‘s d effect sizes for main effects because there is disagreement on what procedure to use
when calculating Cohen‘s d values for interactions. We also followed their procedure to calculate
the pooled variances needed for the Cohen‘s d calculations. Since only two of the test manual
measures had standard scores, we chose to report findings for all measures using raw scores for
consistency and ease of reporting. Each dependent measure was tested for meeting the
assumptions of ANOVA using Chi Square for independence, the Lilliefors test (Lilliefors, 1969)
Race, maternal education and Bus Story retells 17
for normality, and the O‘Brien test of equal variance (O'Brien, 1979, 1981; Trujillo-Ortiz &
Hernandez-Walls, 2003). The means and standard deviations for the measures from the RBS-NA
are shown in Table 5. Only the Information score met the assumptions of ANOVA. The
Sentence Length data and the Complexity data had a right skewed distribution and equal
variance, while the Independence data had a left skewed distribution and unequal variance.
Attempts to transform these data to meet the assumptions of ANOVA were unsuccessful. When
no nonparametric test exists that fits the experimental design, Conover (1999) recommends
conducting an ANOVA on the raw data and on the rank transformed data. ―If the two procedures
give nearly identical results, the assumptions underlying the usual analysis of variance are likely
to be reasonable and the regular parametric analysis valid‖ (p. 419-420). Thompson (1991) states
that ANOVAs on ranked data are valid only when there is one significant main effect. In this
situation, the ranked data conform to a non-central X2 distribution. When the ANOVA on ranked
data indicates more than one significant main effect and a significant interaction, the approach of
using ranked data in ANOVA is invalid (Leys & Schumann, 2010; Thompson, 1991).
When we conducted ANOVAs on both the raw and ranked data for the three variables
not meeting the ANOVA assumptions (i.e., sentence length, complexity, independence) as
Conover (1999) recommends, we found that there was only one main effect in each case. Given
that our data conform to the situation in which ANOVA using ranked data is valid according to
Thompson (1991), we reported the results for sentence length, complexity, and independence as
ANOVAs based on ranked data. We set the alpha level at .05 and applied the Bonferonni
correction, which resulted in a correct alpha of .0125 (.05/4, because we had one inferential test
for each of the four dependent measures). The results of the inferential tests are shown in Table
6.
Race, maternal education and Bus Story retells 18
For the Information score, there was a main effect for both maternal education F(1,159) =
10.75, MS = 450.96, p = 0.001, Cohen‘s d = 0.49, 95% CI [0.47, 0.52] and race F(1,159) =
17.94, MS = 752.36, p < 0.001, Cohen‘s d = 0.65, 95% CI [0.63, 0.68], but no interaction
F(1,159) = 0.48, MS = 20, p = 0.491. Maternal education explained 6.33% of the variance, with
>HS children (M = 22.45) producing an Information raw score that was significantly higher than
the score achieved by the <HS children (M = 19.10). Race explained 9.51% of the variance, with
EA children (M = 23.01) producing an Information raw score that was significantly higher than
the score achieved by the AA children (M = 18.67). Recall that our power analysis indicated that
only 3.86% of the variance needed to be explained in order to detect an experimental effect. As
such, the power of our sample was more than adequate.
For the remaining three dependent measures, which did not meet the assumptions of
ANOVA, there were main effects for maternal education, no main effects for race, and no
significant interactions. The following results are reported as X2 values instead of as F values
because ANOVAs conducted on ranked data conform to a non-central X2 distribution
(Thompson, 1991). For Sentence Length, the main effect for maternal education X2(1,159) =
7.72, MS = 16,348.99, p = 0.006, Cohen‘s d = 0.47, 95% CI [0.45, 0.50] explained 4.63% of the
variance, with the five longest sentences of >HS children (M = 9.28) having greater average
length than the <HS children (M = 8.28). For Complexity, the main effect for maternal education
X2 (1,159) = 9.17, MS = 17,951.20, p = 0.029, Cohen‘s d = 0.48, 95% CI [0.46, 0.51] explained
5.45% of the variance, with >HS children (M = 2.73) producing more complex sentences than
the <HS children (M = 2.10). For the Independence score, the main effect for maternal X2 (1,159)
= 13.02, MS = 25,194.50, p = 0.004, Cohen‘s d = 0.66, 95% CI [0.64, 0.69] explained 7.31% of
the variance, with >HS children (M = 41.32) requiring less prompting than the <HS children (M
Race, maternal education and Bus Story retells 19
= 37.91).
Discussion
Maternal Education Effects
All four scores of the RBS-NA showed significant differences between children whose
mothers had a lower education level (<HS) as compared to a higher education level (>HS), with
the children in the higher maternal education group scoring higher in all instances. These
findings are supported by other research using spontaneous conversational language samples and
standardized, norm-referenced tests. Dollaghan and her colleagues (1999), for example, analyzed
the language samples and Peabody Picture Vocabulary Test –Revised (PPVT-R) (Dunn, 1980)
scores of 240 three-year old children. Comparing three levels of maternal education, they found
statistically significant linear trends across educational levels for their three spontaneous
language sample measures (mean length of utterance in morphemes, the number of different
words, and the total number of words) and on the standardized test of receptive vocabulary
(PPVT-R).
The mean length of utterance measure used by Dollaghan et al. has been found to be very
highly correlated with the mean length of a child‘s five longest utterances (e.g., Malakoff,
Mayes, Schottenfeld, & Howell, 1999), which was the Sentence Length measure used in the
RBS-NA. As such, these two findings are logically connected and support each other (although
the children were different ages and the language samples were collected differently). On a more
general level, the RBS-NA and the PPVT-R used by Dollaghan and her colleagues are both
standardized, norm-referenced language tests. At this level, the two sets of results are similar in
finding maternal education effects. Qi and her colleagues (Qi, Kaiser, Milan, & Hancock, 2006)
also found that among the 482 low-income AA preschoolers in their sample, maternal education
Race, maternal education and Bus Story retells 20
level had a significant impact on the children‘s PPVT-III performance, with children‘s whose
mothers had a high school degree or less scoring 11 points lower than children whose mothers
had a Bachelor‘s degree. However, the PPVT and the RBS-NA each focus on very different
aspects of language competence (oral narrative skill versus single word receptive vocabulary),
which weakens support that results on one test provide for the other.
A standardized, norm-referenced language test more closely aligned with the RBS-NA
(which requires the integration of many language skills in producing a narrative) is the Preschool
Language Scale-3, or PLS-3 (an omnibus test of language abilities). Qi et al., 2003, found
significant effects of maternal education level on preschoolers‘ performance on the PLS-3 within
low-income groups of both AA and EA preschoolers. The findings with the PLS-3 support those
obtained for the RBS-NA in the current study.
For research focused specifically on narratives, the findings of the current study are at
odds with previous research. Recall that Reese and her colleagues (Reese et al., 2009) found no
relationship between maternal education and three aspects of the storybook retell narratives in
six and seven year olds. However, also recall that this study contained no mothers who were
university graduates. This compressed range may have obscured any impact of education level
on the children‘s narrative language abilities. In the current study, 48 mothers (28% of our
sample) had Bachelor‘s or Master‘s degrees. The greater spread in maternal education levels in
our study may account for the differences in findings between our study and that of Reese et al.
(2009). Furthermore, different story retell tasks were used in each study, so it is also possible that
the different stories that were retold in these two studies might explain differences in the results.
In their study of preschoolers, Kulkofsky and Klemfuss (2008) found no main effects or
interactions for maternal education, nor did maternal education correlate with narrative quality.
Race, maternal education and Bus Story retells 21
However, this study may have had too large a proportion of college graduates in the sample.
Recall that in Study 1 (n = 46), 89% of the mothers had a college degree, and in Study 2 (n = 66),
72% did. Furthermore, both the type of narrative elicited and measures used in this study were
quite different than those used in the RBS-NA, so it is difficult to compare the results obtained
from them. The preschoolers‘ retells in the Lulkofsky and Kelmfyss study were of a previously
witnessed staged event, which was measured for overall quality and for the following four
components of narrative quality: volume (number of propositions), complexity (number of words
per proposition), descriptive texture (number of descriptive words), and cohesion (marking
temporal, causal, conditional, & optional states). These do not correspond well with any of the
RBS-NA measures.
In the Price et al. (2006) study of 65 African American children retelling the Bus Story at
age 4 and again at kindergarten entry, it was likewise found that their methods of scoring the Bus
Story (using a story grammar analysis yielding a maximum of 39 points) did not correlate with
maternal education, which ranged from 9 to 18 years (M = 13.0; SD = 2.02). This does not align
with the current findings, because we found a main effect for race, and maternal education level
on the RBS-NA, but there was no significant interaction. If there had been an impact for maternal
education level only for the EA children in the current study, we would have found a significant
interaction. However, the lack of similar findings in these two studies could possibly be due to
differences between the Information score as determined in the RBS-NA manual and the story
grammar analysis these authors conducted on the Bus Story narratives. Although both are macro-
level measures, they may be dissimilar enough to yield different findings. Furthermore, the
pictures are quite different for the British and North American editions, with the British edition
Race, maternal education and Bus Story retells 22
featuring a double-decker bus that most children in the U.S. would be unfamiliar with. It is
possible that this had an influence on the different findings of these two studies.
In summary, our findings that maternal education level impacted all measures of the RBS-
NA correspond well with research on spontaneous language samples produced by three-year olds
that were not narratives (Dollaghan et al., 1999) and performance on other standardized language
tests (Dollaghan et al., 1999; Qi et al., 2003; Qi et al., 2006), but they stand in contrast to
research that has specifically analyzed narratives produced by children ages 4 to 7 (Kulkofsky &
Klemfuss, 2008; Price et al, 2006; Reese et al., 2009).
Race Effects
Although there was a main effect for race for the Information score only, with AA
children performing more poorly than their EA peers, the Information score is a particularly
important one. When researchers report only one score from either version (British or American)
of the Bus Story, the Information score is almost always chosen (e.g., Bishop & Edmundson,
1987; Botting, et al., 1997; Conti-Ramsden & Botting, 1999; Conti-Ramsden, et al., 1997;
Kovas, et al., 2005; Mackie & Dockrell, 2004; Paul, 1996; Paul, et al., 1996; Paul & Smith,
1993; Storch & Whitehurst, 2002; Stothard, et al., 1998). When two scores are reported, they are
the Information score and the Sentence Length score, so here again, the Information score is of
considerable importance in how either version of the Renfrew Bus Story tends to be used in
research (e.g., Anderson, et al., 2004; Anderson, et al., 1997; Briscoe, et al., 1998; Dockrell &
Lindsay, 2001; Gallagher, et al., 2000; Girolametto, et al., 2001; Gray, 2004; Hohen &
Stevenson, 1999; Hughs, et al., 1998; Kennedy, et al., 2006; Nathan, et al., 2004). Indeed, few
studies report more than one or two of these measures. The current findings regarding the
relatively strong effect of children‘s race on their RBS-NA Information score is as such a very
Race, maternal education and Bus Story retells 23
important one, at least for use of this language assessment tool in the United States, from which
the AA sample in the current study was drawn.
These findings are not consistent with other data on the RBS-NA reported by Pankratz et
al. (2007), who found that the small subset of minority children (mostly Hispanic) in their
sample scored significantly lower on Sentence Length, but not on the Information score.
However, given that these children were mostly Hispanic, it is not possible to compare their
findings to those of the current study.
The current findings are also not consistent with a small study conducted by Serpell and
his colleagues (2005). These researchers read a published children‘s story, The Monkey and the
Crocodile (Galdone, 1969) to four groups of AA and EA children second graders, two from low-
income (n = 27 AA and 14 EA) and two from middle-income (n = 26 AA and 14 EA)
backgrounds. To assess their comprehension of the story, the children were asked a number of
literal and inferential questions tapping seven key features of the story. Finally, they were asked
to retell the story. None of the groups differed on two of the three measures used — answers to
the questions about the story, and the length of their retells in words. However, on story telling
quality, as measured by the number of story structure elements included in the retell, the low-
income AA children scored significantly higher than the middle-income AA children.
Furthermore, the low-income children‘s story telling quality scores were on average higher than,
but not significantly different from, both groups of EA children. Serpell et al. noted ―that low-
income African American children exhibited oral narrative skills on a par with the middle-
income European American children is a striking finding‖ (p. 163) and that these findings were
―strikingly different from those obtained for most of the other competency measures‖ (p. 163).
The findings on the effect of race on children‘s Information score in the current study do
Race, maternal education and Bus Story retells 24
dovetail with some, but not all, more general level with findings regarding other standardized,
norm-referenced language tests. The effect size in the current study was d = 0.65, which means
that the average information score performance of the AA children was 0.65 standard deviations
below that of the EA children. This number is similar to the discrepancy between the
performance of AA and EA children on other standardized, norm-referenced language tests
focusing on other aspects of language development besides narrative ability.
For example, Brooks-Gunn and her colleagues (2003) report findings from two large
scale studies on children‘s performance on the Peabody Picture Vocabulary Test –Revised,
PPVT-R (Dunn, 1980), a test of single-word receptive vocabulary. Data from 315 five year-olds
from the Infant Health and Development Program (IHDP) study showed a white-black gap of
1.63 standard deviations. When adjusted for family income, maternal education, and home
environment, the gap was reduced to 0.86 standard deviations. They reported additional data
from 1,354 five and six-year olds from the National Longitudinal Survey of Youth (NLSY),
which showed a white-black gap of 1.15 standard deviations. When adjusted for family income,
maternal education, and home environment, this gap was reduced to 0.73 standard deviations,
which is very close to the .65 standard deviations found in the current study.
Numerous subsequent studies of a new edition of this test, the PPVT –III (Dunn & Dunn,
1997), which some scholars believe may be less biased than the PPVT-R (Washington & Craig,
1999), also show that low-income AA children score on average substantially below the
expected mean based on national norms, ranging from -0.60 to -1.48 SD from the mean (e.g.,
Campbell, Bell, & Keith, 2001; Champion, Hyter, McCabe, & Bland-Stewart, 2003; Qi et al.,
2006; Restrepo, et al., 2006; Washington & Craig, 1999). And, there is evidence here again that
maternal education explains part, but not all, of this discrepancy in the scores of AA children
Race, maternal education and Bus Story retells 25
(Washington & Craig, 1999).
Although race differences have been found on various versions of the PPVT in some
studies, the study by Qi et al., 2006, found no significant race differences when low- income AA
children (n = 482) were compared to low-income EA preschoolers (n = 42) on the PPVT-III, the
Expressive Vocabulary Test (Williams, 1997), or the PLS-3. The authors note that caution
should be used in interpreting their results, since the sample sizes of AA and EA children were
substantially different. With this caution in mind, we do not find support for the race differences
found on the RBS-NA in the current study on these tests of receptive and receptive vocabulary
and an omnibus test of language abilities.
In sum, the race difference on the Information score of the RBS-NA dovetails with some,
but not all, previous research regarding the performance of AA children on standardized, norm-
referenced language tests.
Future Research
There is clearly a need to sort out possible reasons for the discrepancy in scores between
children of different races on the RBS-NA (and on other standardized language tests). One reason
often suggested for standardized test score discrepancies and/or achievement discrepancies (often
measured by standardized test scores) is that these children are so much more likely to live in
poverty (e.g., Duncan, Yeung, Brooks-Gunn, & Smith, 1998). We attempted to control for this
possibility in the current study by carefully equalizing the number of children from the two
different races that were in each of the seven specific categories of maternal education that
comprised our two levels of maternal education. Our results did indicate that children‘s
performance on all measures used in the RBS-NA, at least when using maternal education as a
measure of socio-economic status, was clearly influenced by socio-economic status. However,
Race, maternal education and Bus Story retells 26
there was also a performance gap on the Information score on the RBS-NA based on race alone,
and the effect size for it (d = 0.65) was larger than the effect size for maternal education (d =
0.49).
This general phenomenon of a test and/or achievement discrepancy between EA and AA
children has been referred to as the ―black-white achievement gap‖ (e.g., Craig & Washington,
2006), and other possible explanations for it in addition to poverty remain to be explored.
Scholars have suggested a variety of additional factors such as the potential cultural bias of tests
(e.g., Chamberlain & Madeiros-Landurand, 1991; Flowers, 2007; Thomas-Tate, Washington,
Craig, & Packard, 2006; Wyatt, 2002), discrepancies in teacher quality (e.g., Connor, Son,
Hindman, & Morrison, 2005; Talbert-Johnson, 2004; Wayne & Youngs, 2003), differences in
home literacy practices for EA and AA preschoolers (e.g., Anderson-Yockel & Haynes, 1994;
Vernon-Feagans, Hammer, Miccio, & Manlove, 2001), and mismatches for AA children between
the language of the classroom (and of standardized language tests) and the language of home
(see van Kleeck & Schwarz, under review).
Limitations of the Current Study
Before we are assured of maternal and racial difference in children‘s performance on the
Information score of the RBS-NA, it is essential that the current study be replicated. First of all,
the children in the current study all took place in a mathematics intervention during the previous
school year. It is possible, although we consider it unlikely, that the mathematics treatment
differentially impacted the narrative retell skills of these kindergartners as a function of either
race or maternal education.
An additional weakness stems from the fact that all of the children in this study were
from urban classrooms in two cities in the Northeastern United States, and most of them were
Race, maternal education and Bus Story retells 27
from just one of those cities. Greater geographic diversity would be preferable. Another
limitation was that there could be other unmeasured child, teacher, family, school, or even
neighborhood characteristics that impacted the results. Also, we had information only on
mothers‘ education levels as a measure of socio-economic status in the current study. Other
unmeasured aspects of a child‘s SES background were factors such as family income and
parental occupation. Further work that could tease out the independent contribution of these
factors would provide a more detailed explanation for our findings. And finally, the current study
compared only EA and AA children. It is essential to additionally determine how children from
other non-mainstream backgrounds perform on the RBS-NA.
Clinical Implications
The findings of the current study suggest that clinicians should use caution in using the
RBS-NA to determine the presence or absence of language impairment in AA children (at least
for the Information score) and children from low-SES backgrounds (for all RBS-NA scores).
Similar discrepancies in scores for African American children on other standardized language
tests have lead to suggestions that spontaneous language sampling, supplemented with other
informal language assessments, provide more valid methods for assessing these children‘s
language skills in an unbiased fashion (e.g., Craig & Washington, 2006).
Spontaneous language samples can be conversational, or they can be a narrative
generated by the child. As such, we might have expected that children‘s story retells, such as
those generated in the RBS-NA, would produce results similar to spontaneous language samples
and not show differences based on race alone. For the RBS-NA measures focused at the sentence
level, Sentence Length and Complexity, this was true, as there were no significant differences on
these measures based on a child‘s race.
Race, maternal education and Bus Story retells 28
The Information score, however, showed a medium large (d = 0.65) effect size for race,
and the importance of this score is highlighted by the fact that it is often the only RBS-NA score
used in many research studies, as noted earlier. The Information score requires more than being
able to generate language at the sentence level. As also noted earlier, it requires the integration of
other skills, such as memory (of key words and concepts), adequate knowledge of the specific
vocabulary in the story to be able to use it in a retell, and sufficient knowledge of story structure
to convey the information in the proper sequence. So, the language differences found on the
RBS-NA that were impacted by a child‘s race appear to reflect differences in language ability
beyond the sentence level. But this may not mean that AA children have a deficit in producing
language beyond the sentence level. It may instead reflect that the narrative a child is required to
produce in the RBS-NA is one that was not chosen by the child, and hence has no social
importance to him or her, unlike narratives the child may be quite adept at producing.
More specifically, it may be that the RBS-NA reflects what Bloome and his colleagues
(Bloome, Katz, & Champion, 2003) have referred to as ―narrative as text.‖ Narrative as text is
emphasized in school, and it tends to separate the existence of the storyteller from the
story. Narrative as performance, on the other hand, explicitly builds social relationships
and identities, but is of diminished importance in school. AA children are often skilled at
narrative as performance, but may lack experience with narrative as text that is important
in school.
The instructions for The Renfrew Bus Story reflect a more “school-like” task. The
examiner says, “I’m going to tell you a story about this bus. Listen very carefully because
when I’m finished, you’re going to tell me the story on this tape recorder. Are you ready?”
To the extent that The Renfrew Bus Story emphasizes narrative as text as opposed to
Race, maternal education and Bus Story retells 29
narrative as performance, AA children might be expected to perform more poorly on it than
their EA peers. Nonetheless, narrative as text is an important academic skill, so the RBS-NA
may be predicting future school difficulties for AA children, even if it does not reflect a
general problem with narrative generation. Future research will need to distinguish these
different types of narratives before the source of AA children’s difficulty with the
Information score on the RBS-NA can be discerned.
The clinical implications of the findings in the current study for children whose mothers
had lower education levels – they scored significantly lower on all RBS-NA measures – was
underscored by Dollaghan et al. (1999). These scholars concluded that there is a ―critical need to
examine the influences of maternal education and other sociodemographic variables on all norm-
referenced tests and measures of spontaneous language production and comprehension, as a
prerequisite to using such measures to compare children or to identify children with language
disorders‖ (p. 1441). Our results indicate that the RBS-NA will likely over-identify children from
low SES backgrounds as having language impairments.
Acknowledgments
The research reported here was supported by the Institute of Education Sciences, U.S.
Department of Education, through Grant No. R305K05157 to the University at Buffalo, State
University of New York, D. H. Clements and J. Sarama, ―Scaling Up TRIAD: Teaching Early
Mathematics for Understanding with Trajectories and Technologies.‖ The opinions expressed are
those of the authors and do not represent views of the U.S. Department of Education.
Race, maternal education and Bus Story retells 30
References
Abdi, H., Edelman, B., Valentine, D., & Dowling, W. J. (2009). Experimental Design & Analysis
for Psychology. New York: Oxford.
Altman, D. G. (2000). Appendix 1: Confidence intervals. In D. L. Sackett, S. E. Straus, W. S.
Richardson, W. Rosenberg & R. B. Haynes (Eds.), Evidence-based medicine: How to
practice and teach EBM. London: Churchill Livingstone.
Anderson, J., Anderson, A., Lynch, J., & Shapiro, J. (2003). Storybook reading in a multicultural
society: Critical perspectives. In A. van Kleeck, S. A. Stahl & E. Bauer (Eds.), On
reading to children: Parents and teachers (pp. 203-230). Mahwah, NJ: Lawrence
Erlbaum Associates.
Anderson, V. A., Morse, S. A., Catroppa, C., Haritou, F., & Rosenfeld, J. (2004). Thirty month
outcome from early childhood head injury: A prospective analysis of neurobehavioural
recovery. Brain, 127, 2608-2620.
Anderson, V. A., Morse, S. A., Klug, G., Catroppa, C., Haritou, F., Rosenfeld, J., et al. (1997).
Predicting recovery from head injury in young children: A prospective analysis. Journal
of International Neuropsychological Society, 3, 368-580.
Anderson-Yockel, J., & Haynes, W. O. (1994). Joint book-reading strategies in working-class
African American and white mother-toddler dyads. Journal of Speech and Hearing
Research, 37, 583-593.
Bishop, D. V. M., & Edmundson, A. (1987). Language-impaired 4-year-olds: Distinguishing
transient from persistent impairment. Journal of Speech and Hearing Research, 52, 156-
173.
Bloome, D., Katz, L. & Champion, T. (2003). Young children‘s narratives and ideologies of
language in classrooms. Reading & Writing Quarterly, 19, 205 – 223.
Bornstein, M. H., Hahn, C.-S., Suwalsky, J. T. D., & Haynes, O.M. (2003). Socioeconomic
status, parenting, and children development: The Hollingshead four-factor index of social
status and the socioeconomic index of occupations. In M. H. Bornstein & R. H. Bradley
(Eds.), Socioeconomic status, parenting, and child development (pp. 29 – 82). Mahwah,
NJ: Erlbaum.
Botting, N. (2002). Narrative as a tool for assessment of linguistic and pragmatic impairments.
Child Language Teaching & Therapy, 18(1), 1-22.
Botting, N., Conti-Ramsden, G., & Crutchley, A. (1997). Concordance between teacher/therapist
opinion and formal language assessment scores in children with language impairment.
European Journal of Disorders of Communication, 32, 317-327.
Briscoe, J., Gathercole, S. E., & Marlow, N. (1998). Short-term memory and language outcomes
after extreme prematurity at birth. Journal of Speech, Language, and Hearing Research,
41(41), 654-666.
Brooks-Gunn, J., Klebanov, P. K., Smith, J., Duncan, G. J., & Kyunghee, L. (2003). The black-
white test score gap in young children: Contributions of test and family characteristics.
Applied Developmental Science, 7(4), 239-252.
Campbell, M. J., Bell, S. K., & Keith, L. K. (2001). Concurrent validity of the Peabody Picture
Vocabulary Test – Third Edition as an intelligence and achievement screener for low SES
African American children. Assessment, 8(1), 85-94.
Race, maternal education and Bus Story retells 31
Chamberlain, C., & Madeiros-Landurand, P. (1991). Practical considerations for assessment of
LEP students with special needs. In E. V. Hamayan & J. S. Damico (Eds.), Limiting bias
in the assessment of bilingual students (pp. 111 - 156). Austin, TX: Pro-Ed.
Champion, T. B., Hyter, Y. D., McCabe, A., & Bland-Stewart, L. M. (2003). A matter of
vocabulary: Performances of low-income African American Head Start children on the
Peabody Picture Vocabulary Test-III. Communication Disorders Quarterly, 24, 121-127.
Connor, C. M., Son, S., Hindman, A., & Morrison, F. J. (2005). Teacher qualifications,
classroom practices, family characteristics and preschool experience" Complex effects on
first graders' vocabulary and early reading outcomes. Journal of School Psychology, 43,
343 - 375.
Conover, W. J. (1999). Some methods based on ranks. In Practical nonparametric statistics (pp.
269-427). New York: John Wiley & Sons, Inc.
Conti-Ramsden, G., & Botting, N. (1999). Characteristics of children attending language units in
England: A national study of 7-year-olds. International Journal of Language &
Communication Disorders, 34(4), 359-366.
Conti-Ramsden, G., Crutchley, A., & Botting, N. (1997). The extent to which psychometric tests
differentiate subgroups of children with SLI. Journal of Speech, Language, and Hearing
Research, 40, 765-777.
Cortina, J. M., & Nouri, H. (2000). Effect size for ANOVA designs. Thousand Oaks, CA: Sage.
Craig, H. K., & Washington, J. A. (2006). Malik goes to school: Examining the language skills
of African American students from preschool–5th
grade. Mahwah, NJ: Lawrence Erlbaum
Associates.
Dockrell, J. E., & Lindsay, G. (2001). Children with specific speech and language difficulties –
The teachers‘ perspective. Oxford Review of Education, 27(3), 369-394.
Dollaghan, C. A., Campbell, T. F., Paradise, J. L., Feldman, H. M., Janosky, J. E., Pitcairn, D.
N., et al. (1999). Maternal education and measures of early speech and language. Journal
of Speech, Language, and Hearing Research, 42, 1432-1443.
Duncan, L., Yeung, W. J., Brooks-Gunn, J., & Smith, J. R. (1998). How much does childhood
poverty affect the life chances of children? American Sociological REview, 63, 406 - 423.
Dunn, L. M. (1980). Peabody Picture Vocabulary Test-Revised (PPVT-R). Circle Pines, MN:
American Guidance Service.
Dunn, L. M., & Dunn, L. M. (1997). PPVT-III: Peabody Picture Vocabulary Test (3rd
ed.).
Circle Pines, Minnesota: American Guidance Service.
Feagans, L., & Appelbaum, M. I. (1986). Validation of language substyles in learning disabled
children. Journal of Educational Psychology, 78, 358-364.
Federal Interagency Forum on Child and Family Statistics (2009). America’s children: Key
national indicators of well-being, 2009. Washington, D.C.: U.S. Government Printing
Office.
Flowers, L. A. (2007). New directions in research recommendations for research to improve
reading achievement for African-American students. Reading Research Quarterly, 42,
424 - 428.
Galdone, P. (1969). The Monkey and the Crocodile. New York: Houghton Miflin.
Gallagher, A., Frith, U., & Snowling, M. J. (2000). Precursors of literacy delay among children
at genetic risk of dyslexia. Journal of Child Psychology and Psychiatry, 41, 203-213.
Race, maternal education and Bus Story retells 32
Girolametto, L., Wiigs, M., Smyth, R., Weitzman, E., & Pearce, P. S. (2001). Children with a
history of expressive vocabulary delay: Outcomes at 5 years of age. American Journal of
Speech-Language Pathology, 10, 358-369.
Glasgow, C., & Cowley, J. (1994). Renfrew Bus Story (North American Edition). Centreville,
DE: Centreville School.
Gray, S. (2004). Word learning by preschoolers with specific language impairment: Predictors
and poor learners. Journal of Speech, Language, and Hearing Research, 47, 1117-1132.
Griffin, T., Hemphill, L., Camp, L., & Wolf, D. (2004). Oral discourse in the preschool years and
later literacy skills. First Language, 24, 123-147.
Haccoun, R. R. (2001). Review of The Renfrew Bus Story. In B. Plake & J. C. Impara (Eds.),
Fourteenth Mental Measurement Yearbook (pp. 1005-1008). Lincoln, NE: Buros Institute
of Mental Measurements.
Hauser, R. (1994). Measuring socioeconomic status in studies of child development. Child
Development, 65, 1541 – 1545.
Hinkle, D. E., Wiersma, W., & Jurs, S. G. (2003). Applied statistics for the behavioral sciences
(5th ed.). New York: Houghton Mifflin Company.
Hoff, E., Laursen, B., & Tardif, T. (2002). Socioeconomic status and parenting. In M. H.
Bornstein (Ed.), Handbook of parenting (2nd
ed., pp. 231 – 252). Mahwah, NJ: Erlbaum.
Hohen, B., & Stevenson, J. (1999). The structure of genetic influences on general cognitive,
language, phonological, and reading abilities. Developmental Psychology, 35(2), 590-
603.
Hughs, C., Dunn, J., & White, A. (1998). Trick or Treat?: Uneven understanding of mind and
emotion and executive dysfunction in ―hard-to-manage‖ preschoolers. Journal of Child
Psychology & Psychiatry, 39(7), 981-994.
Huston, A. C., McLoyd, V.C., & Garcia Coll, C. (1994). Children and poverty: Issues in
contemporary research. Child Development, 65, 275 – 282.
Kennedy, C. R., McCann, D. C., Campbell, M. J., Law, C. M., Mullee, M., Petrou, S., et al.
(2006). Language ability after early detection of permanent childhood hearing
impairment. New England Journal of Medicine, 354, 2131-2141.
Kovas, Y., Hayiou-Thomas, E., Oliver, B., Bishop, D., Dale, P. S., & Plomin, R. (2005). Genetic
influences in different aspects of language development: The etiology of language skills
in 4.5-year old twins. Child Development, 76(3), 632-651.
Kulkofsky, S., & Klemfuss, J. Z. (2008). What the stories children tell can tell about their
memory: Narrative skill and young children‘s suggestibility. Developmental Psychology,
44(5), 1442-1456.
Lange, A. A. (2010). Sales Report for 2009 for Learning Tools LLC [Report]. Buffalo,
NY: Author. Leys, C., & Schumann, S. (2010). A nonparametric method to analyze interactions: The adjusted
rank transform test. Journal of Experimental Social Psychology, 46, 684-688.
Lilliefors, H. W. (1969). On the Kolmogorov-Smirnov test for the exponential distribution with
mean unknown. Journal of the American Statistical Association, 64, 387-389.
Mackie, C., & Dockrell, J. E. (2004). The nature of written language deficits in children with
SLI. Journal of Speech, Language, and Hearing Research, 47, 1469-1483.
Malakoff, M. E., Mayes, L. C., Schottenfeld, R., & Howell, S. (1999). Language production in
24-month-old inner-city children of cocaine-and-other-drug-using mothers. Journal of
Applied Developmental Psychology, 20(1), 159 - 180.
Race, maternal education and Bus Story retells 33
McGregor, K. K. (2000). The development and enhancement of narrative skills in a preschool
classroom: Towards a solution to clinician-client mismatch. American Journal of Speech-
Language Pathology, 9(1), 55-71.
Menyuk, P., Chesnick, M., Liebergott, J. W., Korngold, B., D‘Agostino, R., & Belanger, A.
(1991). Predicting reading problems in at-risk children. Journal of Speech and Hearing
Research, 34, 893-903.
Mercy, J. A., & Steelman, L. C. (1982). Familial influences on the intellectual attainment of
children. American Sociological Review, 47, 532 – 542.
Miller, J. F., Heilman, J., Nockerts, A., Iglesias, A., Fabiano, L., & Francis, D. J. (2006). Oral
language and reading in bilingual children. Learning Disabilities Research & Practice,
21(1), 30 – 43.
Nathan, L., Stackhouse, J., Goulandris, N., & Snowling, M. J. (2004). The development of early
literacy skills among children with speech difficulties: A test of the ―critical age
hypothesis‖. Journal of Speech, Language, and Hearing Research, 47, 377-391.
Newcomer, P., & Hammill, D. (1997). Test of Language Development-Primary. Austin, TX:
PRO-ED, Inc.
Norbury, C. F., & Bishop, D. V. M. (2003). Narrative skills of children with communication
impairments. International Journal of Language & Communication Disorders, 38(3),
287-313.
O‘Brien, R. G. (1979). A general ANOVA method for robust test of additive models for
variance. Journal of American Statistical Association, 74, 877-880.
O‘Brien, R. G. (1981). A simple test for variance effects in experimental designs. Psychological
Bulletin, 89, 570-574.
O‘Neill, D. K., Pearce, M. J., & Pick, J. L. (2004). Preschool children‘s narratives and
performance on the Peabody Individualized Achievement Test-Revised: Evidence of a
relation between early narrative and later mathematical ability. First Language, 24, 149-
183.
Pankratz, M. E., Plante, E., & Insalaco, D. M. (2007). The diagnostic and predictive validity of
The Renfrew Bus Story. Language, Speech, and Hearing Services in Schools, 38, 390-
399.
Paul, R. (1996). Clinical implications of the natural history of slow expressive language
development. American Journal of Speech-Language Pathology, 5, 5-21.
Paul, R., Hernandez, R., Taylor, L., & Johnson, K. (1996). Narrative development in late talkers:
Early school age. Journal of Speech and Hearing Research, 39, 1295-1303.
Paul, R., & Smith, R. (1993). Narrative skills in 4-year-olds with normal, impaired, and late-
developing language. Journal of Speech and Hearing Research, 36, 592-598.
Petersen, D. B., Gillam, S. L., & Gillam, R. B. (2008). Emerging procedures in narrative
assessment: The Index of Narrative Complexity. Topics in Language Disorders, 20, 111-
126.
Preschool Curriculum Evaluation Research Consortium. (2008). Effects of Preschool Curriculum
Programs on School Readiness (NCER 2008-2009). Washington, D.C.: National Center
for Education Research, Institute of Education Sciences, U.S. Department of Education.
Price, J. R., Roberts, J. E., & Jackson, S. C. (2006). Structural development of the fictional
narratives of African American preschoolers. Language, Speech, and Hearing Services in
Schools, 37, 178-190.
Race, maternal education and Bus Story retells 34
Qi, C. H., Kaiser, A. P., Milan, S., & Hancock, T. (2006). Language performance of low-income
African American and European American preschool children on the PPVT-III.
Language, Speech, and Hearing Services in Schools, 37(1), 5-16.
Qi, C. H., Kaiser, A. P., Milan, S., Yzquierdo, Z., & Hancock, T. B. (2003). The performance of low-income, African American children on the Preschool Language Scale-3. Journal of Speech, Language, and Hearing Research, 46(3), 576-590.
Reese, E., Suggtate, S., Long, J., & Schaughency, E. (2009). Children‘s oral narrative and
reading skills in the first 3 years of reading instruction. Reading and Writing,
Renfrew, C. E. (1969). The Bus Story: A test of continuous speech. North Place, Old Headington:
Oxford.
Restrepo, M. A., Schwanenflugel, P. J., Blake, J., Neuharth-Pritchett, S., Cramer, S. E., &
Ruston, H. P. (2006). Performance on the PPVT-III and the EVT: Applicability of the
measures with African American and European American preschool children. Language,
Speech, and Hearing Services in Schools, 31, 17-27.
Roth, F. P., Speece, D. L., & Cooper, D. H. (2002). A longitudinal analysis of the connection
between oral language and early reading. Journal of Educational Research, 95, 259-272.
Scarborough, H. S. (2001). Connecting early language and literacy to later reading (dis)abilities:
Evidence, theory, and practice. In S. B. Neuman & D. K. Dickinson (Eds.), Handbook of
Early Literacy Development (pp. 97-110). New York: Guilford Publications.
Sénéchal, M., & LeFevre, J. A. (2002). Parental involvement in the development of children‘s
reading skill: A five-year longitudinal study. Child Development, 73, 445-460.
Serpell, R., Baker, L., & Sonnenschein, S. (2005). Becoming literate in the city: The Baltimore
Early Childhood Project. Cambridge: Cambridge University Press.
Snow, C. E., Tabors, P. O., Nicholson, P. A., & Kurland, B. F. (1995). SHELL: Oral language
and early literacy skills in kindergarten and first-grade children. Journal of Research in
Childhood Education, 10, 37-48.
Snyder, L. S., & Downey, D. M. (1991). The language-reading relationship in normal and
reading-disabled children. Journal of Speech and Hearing Research, 34, 129-140.
Stein, N., & Glenn, C. (1982). Children‘s concept of time: The development of a story schema.
In W. Friedman (Ed.), The developmental psychology of time (pp. 255-282). New York:
Academic Press.
Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related precursors to reading:
Evidence from a longitudinal structural model. Developmental Psychology, 38(6), 934-
947.
Stothard, S. E., Snowling, M. J., Bishop, D. V. M., Chipchase, B. B., & Kaplan, C. A. (1998).
Language-impaired preschoolers: A follow-up into adolescence. Journal of Speech and
Hearing Research, 41, 407-418.
Tabors, P. O., Snow, C. E., & Dickinson, D. K. (2001). Homes and schools together: Supporting
language and literacy development. In D. K. Dickinson & P. O. Tabors (Eds.), Beginning
literacy with language (pp. 313-334). Baltimore: Paul H. Brookes.
Talbert-Johnson, C. (2004). Structural inequities and the achievement gap in urban schools.
Education and Urban Society 37(1), 22 - 36.Thomas-Tate, S., Washington, J., Craig, H.,
& Packard, M. (2006). Performance of African American preschool and kindergarten
students on the expressive vocabulary test. Language, Speech, and Hearing Services in
Schools, 37, 143 - 149.
Race, maternal education and Bus Story retells 35
Thompson, G. L. (1991). A note on the rank transform for interactions. Biometrika Trust, 697-
701.
Trujillo-Ortiz, A., & Hernandez-Walls, R. (2003). Obrientest: O‘Brien‘s test for homogeneity of
variances. A MATLAB file. [WWW document], from
http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=3335&obje
ctType=FILE
van Kleeck, A. (2007). SLPs‘ foundational role in reading comprehension: A response to Kamhi.
The ASHA Leader, 12 (10), 32 – 33.
van Kleeck, A. (2008). Providing preschool foundations for later reading comprehension: The
importance of and ideas for targeting inferencing in book-sharing interventions.
Psychology in the Schools, 45(6), 627-643.
van Kleeck, A. & Schwarz, A. L. (under review). Making ―school talk‖ explicit: A critical
preschool foundation for later reading comprehension and general academic success.
Swiss Journal for Educational Research.
Vellutino, F. R., Tunmer, W. E., Jaccard, J. J., & Chen, R. (2007). Components of reading
ability: Multivariate evidence for a convergent skills model of reading development.
Scientific Studies of Reading, 11, 3-32.
Vernon-Feagans, L., Hammer, C., Miccio, A., & Manlove, E. (2001). Early language literacy
skills in low-income African American and Hispanic children. In S. B. Neuman & D. K.
Dickinson (Eds.), Handbook of early literacy development (pp. 192-210). New York:
Guilford Publications.
Washington, J. A., & Craig, H. K. (1999). Performance of at-risk, African American
preschoolers on the Peabody Picture Vocabulary Test-III. Language, Speech, and
Hearing Services in Schools, 30, 75-82.
Wayne, A. J., & Youngs, P. (2003). Teacher characteristics and student achievement gains: A
review. Review of Educational Research, 73(1), 89 - 122.
Williams, K. T. (1997). Expressive Vocabulary Test. Circle Pines, MN: American Guidance Service.
Wyatt, T. (2002). Assessing the communicative abilities of clients from diverse cultural and
language backgrounds. In D. E. Battle (Ed.), Communication disorders in multicultural
populations (3rd ed.) (pp. 415 - 458). Boston: Butterworth-Heinemann.
Zimmerman, I., Steiner, V., & Pond, R. (1992). Preschool Language Scale - 3. San Antonio, TX: The Psychological Corporation, Harcourt, Brace, Jovanovich.
Race, maternal education and Bus Story retells 36
Table 1
Frequency Counts of Children Who Received the TRIAD Treatment by Race and Maternal Education Categories and Sample
Selection Procedure for the Current Study
TRIAD Maternal
Education Categories 8
th Grade
High School
(GED)
Vocational /
Technical
1 to 2 years
College
Associate‘s
Degree
Bachelor‘s
Degree
Master‘s
Degree Total
African American (AA) 26 96 8 77 28 27 5 267
European American (EA) 13 28 1 14 6 19 15 96
Total 39 124 9 91 34 46 20 363
Quasi-Random Sample Selection for the Current Study
African American (AA) 13 28 1 14 6 19 5 86
European American (EA) 13 28 1 14 6 19 5 86
Total 26 56 2 28 12 38 10 172
Re-categorization of Maternal Education Categories for the Current Study
High School or Less (≤ HS) More than High School (> HS) Total
African American (AA) 41 45 86
European American (EA) 41 45 86
Total 82 90 172
Race, maternal education and Bus Story retells 37
Table 2
Sample characteristics by group
Classification Initial N N Less Outliers Boys Bostona
AA ≤ HS 41 40 21 7
AA > HS 45 42 21 9
EA ≤ HS 41 39 18 5
EA > HS 45 42 20 9
TOTAL 172 163 80 30
a = Data meeting our race and maternal education search criteria were collected at a site in
Buffalo, New York and in Boston, Massachusetts
Race, maternal education and Bus Story retells 38
Table 3
Analysis of gender differences on each dependent measure
Variable
Female
(n = 83)
Male
(N = 80)
Test
Statistic
Cohen’s
d
Confidence
Intervals
Mean SD Mean SD
Two-Tailed t-test Results (t statistic) df = 161
Information
Score
21.11 7.74 20.54 6.15 0.57 0.08
-0.04, 0.20
Two-Tailed Mann Whitney Test Results (U statistic)
Sentence
Length
8.95 2.32 8.63 1.99 3,063.50 0.15
0.07, 0.23
Complexity 2.46 1.35 2.39 1.31 3,225.00 0.05
-0.02, 0.13
Independence 39.48 4.98 39.86 5.85 3,579.00 -0.07
-0.18, 0.03
* p < .0125, **p <.001
Race, maternal education and Bus Story retells 39
Table 4
Analysis of site differences on each dependent measure
Variable
Buffalo
(n=133)
Boston
(n=30)
t-test
values
Cohen’s
d
Confidence
Intervals
Mean SD Mean SD
Two-Tailed t-test Results (t statistic) df = 161
Information
Score
20.77 6.81 21.07 7.83 -0.21 0.04
-0.12, 0.20
Two-Tail Mann Whitney Test Results (U statistic)
Sentence
Length
8.83 2.19 8.64 2.05 2,043.50 -0.09
-0.21, 0.04
Complexity 2.45 1.33 2.30 1.29 2,123.50 -0.11
-0.23, 0.00
Independence 39.52 5.53 40.33 4.89 1,813.50 0.15
0.00, 0.30
* p < .0125, **p <.001
Race, maternal education and Bus Story retells 40
Table 5
Renfrew Bus Story – North American Edition Means, Standard Deviations, and Ranges for Measures from the Test Manual
Information
Raw Score
Sentence
Length
Complexity
Independence
Raw Score
Race Maternal
Education
Mean
(SD)
Range Mean
(SD)
Range Mean
(SD)
Range Mean
(SD)
Range
AA High school or less 17.33
(6.10)
7.00 to
32.00
8.00
(2.14)
5.40 to
12.40
1.85
(1.05)
1.00 to
4.00
36.90
(7.52)
17.00 to
44.00
More than high school 19.95
(5.97)
10.00 to
32.00
9.26
(2.61)
5.60 to
16.40
2.64
(1.50)
1.00 to
7.00
42.17
(2.57)
36.00 to
44.00
EA High school or less 20.92
(6.13)
10.00 to
34.00
8.56
(1.51)
6.00 to
12.20
2.36
(1.22)
1.00 to
6.00
38.95
(5.11)
30.00 to
44.00
More than high school 24.95
(7.54)
11.00 to
39.00
9.29
(2.02)
6.00 to
15.60
2.81
(1.31)
1.00 to
6.00
40.48
(4.00)
29.00 to
44.00
Note: SD = standard deviations; Information = total score for amount of key information used in the story retell. Sentence length =
Average length in words of 5 longest utterances; Complexity = number of utterances with one or more subordinate clause;
Independence = inverse of number of prompts needed to complete retell. AA = African American; EA = European American
Race, maternal education and Bus Story retells 41
Table 6
Two-way fixed effect analysis of variance results
ANOVA Results using Raw Data
MS F
(1,159) Cohen‘s d
Confidence
Intervals
Information Score
Maternal Education 450.96 10.75* 0.49 0.47, 0.52
Race 752.36 17.94* 0.65 0.63, 0.68
Maternal Education x Race 20.00 0.48
ANOVA Results using Ranked Data
MS X
2
(1,159) a
Cohen‘s d
Confidence
Intervals
Sentence Length
Maternal Education 16,348.99 7.72* 0.47 0.45, 0.50
Race 5,974.36 2.82 0.14 0.11, 0.16
Maternal Education x Race 1,388.69 0.66
Complexity
Maternal Education 17,951.20 9.17* 0.48 0.46, 0.51
Race 8,246.40 4.21 0.26 0.24, 0.29
Maternal Education x Race 1,229.10 0.63
Independence Score
Maternal Education 25,194.50 13.02* 0.66 0.64, 0.69
Race 579.30 0.30 0.03 0.00, 0.05
Maternal Education x Race 10,594.40 5.47
Note: * = significant at the corrected alpha (Bonferroni) of 0.0125, a = the results are reported as
X2 because ANOVAs conducted on ranked data conform to a non-central X
2 distribution
(Thompson, 1991).