Article annotation
Badri Joe
KinderArticle.pdf
Evidence of alphabetic knowledge in writing: connections to letter and word identification skills in preschool and kindergarten
Victoria J. Molfese • Jennifer L. Beswick • Jill L. Jacobi-Vessels •
Natalie E. Armstrong • Brittany L. Culver • Jamie M. White •
Melissa C. Ferguson • Kathleen Moritz Rudasill • Dennis L. Molfese
Published online: 25 September 2010
� Springer Science+Business Media B.V. 2010
Abstract The writing skills of 286 children (157 female and 129 male) were studied by comparing name writing and letter writing scores from preschool to
kindergarten with letter and word reading scores over the same time period. Two
rubrics for scoring writing were compared to determine if scores based on multiple
components (i.e., letter formation, orientation on the vertical axis, left–right orien-
tation, and correct letter sequencing) would better reflect differences in children’s
writing knowledge in preschool and kindergarten than rubrics composed of one
component (i.e., letter formation only). While developmental changes in writing
scores were found, little additional information was provided by multiple component
scoring rubrics compared to the single component rubric. Letter writing scores were
more strongly related to letter and word reading scores than name writing scores but
neither writing score was predictive of growth. Implications of the findings for
intentional/systematic writing instruction in preschool curricula are discussed.
Keywords Letter and word identification skills � Letter writing � Name writing � Rubrics for scoring writing
Introduction
Research studies have confirmed the importance of alphabetic skills (knowledge of
letter names and sounds) for the development of reading. Indeed, Denton et al.
(Denton & West, 2002; West, Denton, & Germino-Hausken, 2000) report that
children who are proficient in identifying letters (naming upper and lower case
letters, recognizing beginning and ending word sounds) at entry into kindergarten
V. J. Molfese (&) � J. L. Beswick � J. L. Jacobi-Vessels � N. E. Armstrong � B. L. Culver � J. M. White � M. C. Ferguson � K. M. Rudasill � D. L. Molfese University of Louisville, Louisville, KY, USA
e-mail: [email protected]
123
Read Writ (2011) 24:133–150
DOI 10.1007/s11145-010-9265-8
show stronger skills at the end of kindergarten and in first grade on measures of
phonological processing and word reading compared to children who are not
proficient. The National Early Literacy Panel’s (2008) meta-analysis of the research
studies investigating relations between emergent literacy skills in the preschool
period and reading skills at school age identified alphabetic skills as strong predictors
(r = .48–.54) of decoding, comprehension, and spelling. Not only are alphabetic skills strong predictors of reading skills in English speaking samples, they are also
strong predictors of reading in non-English speaking children (Lyytinen et al., 2004;
Muter & Diethelm, 2001). The focus of this paper is on the alphabetic knowledge
reflected in children’s writing skills. Specifically, this study investigated the
assumption that children’s writing facilitates the development of alphabetic
knowledge (Bus et al., 2001; Juel, 2006) and, as such, facilitates the development
of letter identification and word reading skills in preschool and kindergarten children.
Research on young children’s writing consistently reports medium to large
correlations between writing skills and knowledge of letter names even when
different measures of writing skills are used. For example, many studies have used
some type of name writing task. Diamond, Gerde, and Powell (2008) report medium
correlations (r = .40–.45) across the preschool year between children’s name writing skills and knowledge of letter names. Riley (1995) reported large
correlations (r = .60 and .54, respectively) between kindergarten children’s name writing skills and their letter naming skills with reading skills measured at the end of
kindergarten. Weinberger (1996) reported correlations of similar magnitudes
(r = .51 and .53, respectively) between name writing and letter naming skills at 5 years with subsequent reading skills at 7 years. Similar associations have emerged
with studies using invented spelling (i.e., attempts of children to spell a word before
they acquire conventional spelling rules) as measures of writing. Sulzby and
colleagues (Bus et al., 2001; Sulzby, Barnhart, & Hieshima, 1989) found strong
associations (r = .74) between 4- and 5-year-old children’s use of invented spelling in written work and their knowledge of letter names.
Although writing is assumed to reflect alphabetic skills, this assumption may not be
true. For example, writing activities in which children write their first names are
commonly found in preschool curricula (e.g., morning sign-in, writing center
materials, small group activities). Name writing activities are popular because of the
special significance and relevance of personal names to young children and the ease of
engaging children in the activity (Treiman & Broderick, 1998). The rationale
underlying name writing activities in preschool classrooms is described in research as
well as practitioner publications (e.g., Bloodgood, 1999; Diamond et al., 2008; Green,
1998; Haney, Bissonnette, & Behnken, 2003; Levin, Both-DeVries, Aram, & Bus,
2005). For many children, name writing experiences in preschool classrooms begin
with tracing or copying the letters in their first names and progress toward writing
some or all of the letters in their first names, either unassisted or partially assisted.
However, many of the activities associated with name writing seem to be targeting
procedural knowledge (e.g., motor skills, mechanics of letter formation, perceptual
features of writing [letters arranged on a horizontal line, sequential arrangement])
rather than conceptual knowledge (e.g., graphic representations of letter names or
sounds, the communicative role of writing, language and speech conventions).
134 V. J. Molfese et al.
123
The abilities of children to write their names may not reflect the same
understanding of emergent literacy concepts, such as phonological processing,
concepts about print and alphabetic knowledge, that are assumed to underlie name
writing (Clay, 2001; Welsch, Sullivan, & Justice, 2003) compared to skills reflected
in other writing tasks. For example, Molfese, Beswick, Molnar, and Jacobi-Vessels
(2006) studied 4- and 5-year-old children’s skills in name writing compared to their
skills in writing letters to dictation. These two measures of writing skills were
compared to skills in identifying letters by name and word reading. Modest
correlations (.49 and .24, respectively) were reported between name writing skills
and identification of letters and words on the Wide Range Achievement Test
(WRAT, Wilkinson, 1993) compared to stronger correlations (.77 and .48,
respectively) between skills in writing dictated letters and identification of WRAT
letters and words. Shatil, Share, and Levin (2000) also report stronger relations
between kindergarten children’s letter knowledge skills and scores on writing
dictated words compared to their name writing scores. Attention to letter name and
letter sound relations may not be explicit in the name writing activities that are
commonly found in preschool classrooms. Indeed, Treiman and Broderick (1998)
report that 4 and 5 year old children’s letter name knowledge was linked to the first
letter of their first names but not to the other letters and not to the first letter of their
last names. Letter sound knowledge could not be linked to initial letters of the first
names. Thus, the assumption that having young children learn to write their names
will result in strong links between writing skills, alphabetic skills, and reading skills
may not be true.
Complicating the understanding of how writing may influence the development
of reading skills is the use of different scoring rubrics for writing skills across
studies. Many studies use letter formation as the basis for scoring writing. For
example, the scoring system for kindergarten children devised by Sulzby et al.
(1989) differentiated drawing and scribbling from writing letter strings, invented
spelling, and conventional writing. Welsch et al. (2003) devised a scoring system
for 4 year olds to separate writing elements from drawing elements. Levin et al.
(2005) differentiate graphic and writing-like writing from symbolic writing in an
investigation of the writing skills of children 2–5 years of age. Molfese et al. (2006)
used a rubric derived from the work of Sulzby et al. (1989) in which writing scores
were based on letter formation, with scores ranging from no attempt or refusal to
write, to writing some or all letters regardless of correct formation to writing all
letters in recognizable and correct formation. To these elements, Haney et al. (2003)
added components of letter size and spacing, fine motor control, and letters written
on a horizontal line as parts of a Name Writing Scale for use with kindergarten
children. However, there are other scorable components in writing. In name writing,
scorable components can include orientation of letters on the vertical axis, left to
right orientation of the written letters, letters written in the correct sequence,
appropriate use of upper and lower case letters, and complexity of the name (e.g.,
number of letters or inclusion of punctuation marks, such as hyphens or
apostrophes, along with letters). In letter writing, both letter formation and
orientation of the letters on the vertical axis can be scored. With the use of many
different scoring rubrics across studies it is difficult to compare findings.
Evidence of alphabetic knowledge 135
123
In the current study, we investigated the development of name and letter writing
skills in children from preschool and kindergarten. We were specifically interested
in the relations between writing scores derived from different scoring rubrics that
have often been used in studies of early writing skills and the relation of the writing
scores to letter identification and word reading skills. We hypothesized that scores
from rubrics composed of more components (i.e., letter formation, orientation on the
vertical axis, left–right orientation, and correct letter sequencing) would better
reflect differences in children’s knowledge of writing names and letters in preschool
and kindergarten than rubrics composed of one component (i.e., letter formation
only). We also studied the relation between the scoring rubrics applied to name and
letter writing and children’s performance on letter and word identification tests. If
knowledge of writing names and letters reflect children’s emergent literacy skills,
especially alphabetic skills, we hypothesized that writing scores and letter and word
identification scores would be correlated and that scores based on rubrics composed
of more components would be more sensitive to differences in letter and word
identification scores. Further, if the preschool emphasis on procedural knowledge in
learning writing skills is stronger than the presumed conceptual knowledge, we
hypothesized that letter writing skills would be stronger correlates of letter and word
identification scores. Finally, we studied the relation between the scoring rubrics
applied to name and letter writing and the growth of children’s letter and word
identification scores from fall of preschool to spring of preschool and to mid-year of
Kindergarten. Consistent with the findings reported by Diamond et al. (2008) that
growth of letter knowledge in a preschool sample was influenced by writing skills,
we hypothesized that growth of letter and word identification scores from preschool
to kindergarten would be influenced by name and letter writing scores.
Method
Participants
Participants were recruited from state-funded and Head-Start preschool classrooms
in two mid-western states. Teacher reports were used to identify children as normally
developing and speaking English as their primary language. Children participated in
the longitudinal study from preschool to kindergarten. At the start of the project, there
were 377 children (213 female and 164 male) with an average age of 53.56 months
(SD = 3.47, range = 47–59 months). Ninety-one children were excluded due to
missing data at one or more of the three time points. The remaining 286 children
(157 female and 129 male) with complete data on all measures at each time point
were included in this study. T tests were conducted to determine if there were differences between the included and excluded groups. No significant differences
were found in age at the start of the project (t375 = -.935, p [ .05), or on Woodcock– Johnson Letter Word Identification performance at the start of the project
(t375 = .398, p [ .05). The average age of participants at pretest was 53.62 months (SD = 3.53, range = 47–62 months). The Racial composition of the participants was
83.6% Caucasian, 8.4% African American, 4.2% Hispanic, and 3.8% Multi-racial.
136 V. J. Molfese et al.
123
Children were distributed across 46 preschool classrooms (26 Head Start), with
variable numbers of children per classroom. By kindergarten, children were dispersed
into 119 classrooms. Over half (n = 62) of these classrooms had only 1 study child, whereas 16 classrooms had 2 study children, 17 classrooms had 3 study children,
4 classrooms had 4 study children, 7 classrooms had 5 study children, 6 classrooms had
6 study children, 4 classrooms had 7 study children, 1 classroom had 8 study children,
and 2 classrooms had 9 study children. An intra-class correlation coefficient showed
that approximately 17% of children’s letter–word identification scores in kindergarten
could be explained by their kindergarten classroom placement.
Measures of writing skills
Writing skills: name writing
Children were asked to write their name on a sheet of unlined paper. Two
independent researchers scored each written name and differences in scores were
reconciled by discussion. Two methods were used to score the name writing: letter
formation and multiple components, which included letter formation, vertical
orientation, left-to-right orientation, and letter sequence. No corrections were
applied to control for differences in the number of letters or the presence of non-
letters (e.g., hyphens or apostrophes) in the name. The scoring rubric for name
writing is shown in ‘‘Appendix 1’’ and two name writing samples with scoring are
shown in Figs. 1 and 2.
Writing skills: letter writing
The Woodcock Johnson Tests of Achievement (3rd Edition; WJ-III; Woodcock,
McGrew, & Mather, 2001) Spelling subscale was used to obtain letter writing data.
The Spelling Subscale requires the child to write seven letters to dictation (items
8–14). The first five items could be written in upper or lower case: o/O, x/X, b/B,
c/C, u/U, and the remaining two items were requested to be written in lower case:
e and g. Two methods were used for scoring letter writing: letter formation and
multiple components. Multiple components included letter formation and vertical
orientation. The scoring rubric for letter writing is shown in ‘‘Appendix 2’’. Two
independent researchers scored each written name and the written letters for each
child at each measurement point. Any differences in scores were reconciled by
discussion. Less than 10% of the writing samples required discussion.
Fig. 1 Example of name writing scoring (Faith)—formation: 6; vertical orientation: 3; left–right orientation: 1; sequence: 2; total name writing score: 12
Evidence of alphabetic knowledge 137
123
Measures of non-writing skills
Letter naming and word reading
The Woodcock Johnson Tests of Achievement (3rd Edition; WJ-III; Woodcock,
McGrew, & Mather, 2001) Letter—Word Identification subscale was used to assess
letter naming and word decoding skills by asking the child to identify (point to or
name) a series of letters and words. Children are asked to point to six letters named by
the tester and to name seven more letters from an array of letters. Children are asked to
select the two named words from an array of words and to read as many of the
remaining words as they can until they have six incorrect responses. The WJ-III is
standardized for use with individuals age 2–90. Age-based standard scores in which
the mean score is 100 and the standard deviation is 15 and familiar to most readers
were used in analyses rather than the W score that is unique to the Woodcock Johnson.
Procedure
The University of Louisville Institutional Review Board approved this study.
Trained researchers conducted assessments of the preschoolers in the Fall
(September–October) and in the Spring (March–April). Kindergarten assessments
were completed in the Spring (January–March). Each assessment was conducted in
a quiet area within the school. The tasks were administered in the following order:
name writing, Letter–Word Identification subscale, and letter writing on the
Spelling subscale. Children received stickers and an age-appropriate toy at the
conclusion of the testing session.
Results
The descriptive statistics with means, standard deviations and ranges of the
participants’ ages as well as for the other measures used in the analyses are shown in
Table 1. As can be seen, there is a wide range of writing scores at all measurement
points. Children’s writing scores were better for name writing, with scores ranging
Fig. 2 Example of name writing scoring (Emilio)— formation: 2; vertical orientation: 1; left–right orientation: 1; sequence: 1; total name writing score: 5
138 V. J. Molfese et al.
123
Table 1 Means (M), standard deviations (SD), and range of scores (N = 286)
M SD Range Possible range
Fall Pre-K
Age in months 53.62 3.53 47–62
Name writing
Total score a
5.84 4.31 0–14 0–14
Letter formation b
3.06 1.92 0–7 0–7
Vertical orientation .95 1.10 0–3 0–3
Left–right orientation .61 .49 0–1 0–1
Correct letter sequence 1.20 1.18 0–3 0–3
Letter writing
Total score c
2.53 1.73 0–7 0–10
Letter formation d
2.23 1.30 0–5 0–7
Vertical orientation .30 .54 0–2 0–3
WJIII letter–word (t score) 46.23 7.30 28–65
Spring Pre-K
Age in months 60.34 3.67 51–67
Name writing
Total score a
10.26 3.45 0–14 0–14
Letter formation b
5.02 1.57 0–7 0–7
Vertical orientation 2.01 1.04 0–3 0–3
Left–right orientation .92 .27 0–1 0–1
Correct letter sequence 2.31 .97 0–3 0–3
Letter writing
Total score c
4.22 2.19 0–9 0–10
Letter formation d
3.41 1.53 0–6 0–7
Vertical orientation .79 .77 0–3 0–3
WJIII letter–word (t score) 49.00 7.62 24–70
Spring kindergarten
Age in months 70.13 3.55 63–76
Name writing
Total score a
12.84 1.60 0–14 0–14
Letter formation b
6.12 .88 0–7 0–7
Vertical orientation 2.80 .52 0–3 0–3
Left–right orientation 1.00 .06 0–1 0–1
Correct letter sequence 2.94 .30 0–3 0–3
Letter writing
Total score c
7.65 1.68 1–10 0–10
Letter formation d
5.54 .99 1–7 0–7
Vertical orientation 2.12 .79 0–3 0–3
WJIII letter–word (t score) 53.70 7.74 24–70
a Name writing total score: F(2,570) = 457.98, p = .000
b Name writing-letter formation score: F(2,570) = 425.44, p = .000
c Letter writing total score: F(2,570) = 1050.70, p = .000
d Letter writing-letter formation score: F(2,570) = 857.46, p = .000
Evidence of alphabetic knowledge 139
123
from zero to the maximum score, than for letter writing. However, many children
received low scores at all ages for both name and letter writing. The number of
children in the sample of 286 children at the fall of preschool, spring of preschool,
and spring of kindergarten whose name writing included no correct letters (only
scribbles, drawings, or random letters) or only one or two letters in their names was
131, 22, and 2, respectively. At the other end of the continuum, the number of
children at each measurement point whose name writing included all the letters in
their names written in correct form was eight, 25, and 93, respectively. Letter
writing is also revealing. The number of children at each measurement point whose
letter writing included no letters or only random letters was 110, 47, and two,
respectively, and the number of children at each measurement point with perfect
scores was nine, three, and 59, respectively.
The purpose of the study is to compare two scoring rubrics, one composed of a
single component (letter formation) compared to a second using multiple
components (i.e., letter formation, orientation on the vertical axis, left–right
orientation and correct letter sequencing), to determine if they reflect differences in
children’s knowledge of writing names and letters. Correlations were used to
examine relations between name and letter writing component scores (Table 2).
Zero-order correlations reflect medium to large correlations ranging from .32 to .85
(p \ .01) between writing component scores at each time point. As reflected in the correlations, letter formation, and correct letter sequence were stronger correlates of
the other elements at each measurement point than were vertical orientation and
left–right orientation. Repeated measures ANOVA was used to determine if single
component (letter formation) compared to multiple component total scores were
different at the three time points assessed: fall of preschool, spring of preschool,
Table 2 Correlations among Pre-K and kindergarten name and letter writing elements (N = 286)
Name writing
Fall Spring Kindergarten
1 2 3 4 1 2 3 4 1 2 3 4
Name writing
1 Letter formation .84** .64** .85** .80** .48** .78** .48** .41** .58**
2 Vertical orientation .55** .78** .43** .71** .32** .55**
3 Left–right orientation .80** .66** .59**
4 Correct letter
sequence
Letter writing
Fall Spring Kindergarten
Letter writing
1 Letter formation .74** .75** .78**
2 Vertical orientation
** p \ .01
140 V. J. Molfese et al.
123
and spring of kindergarten. Significantly larger scores were obtained at each time
point regardless of the scoring rubric used (shown in Table 1).
Correlations were used to determine whether scores based on single writing
components compared to multiple writing components were differentially related to
the children’s WJ-III Letter–Word Identification scores (Table 3). The correlations
Table 3 Correlations between writing skills and alphabetic skills (N = 286)
Fall Pre-K Spring Pre-K Spring kindergarten
WJ letter/word WJ letter/word WJ letter/word
Fall Pre-K
Name writing
Total score a
.42** .37** .21**
Letter formation b
.39** .36** .21**
Vertical orientation .39** .38** .18**
Left–right orientation .31** .23** .15**
Correct letter sequence .41** .33** .22**
Letter writing
Total score a
.51** .39** .31**
Letter formation b
.48** .37** .29**
Vertical orientation .49** .37** .29**
Spring Pre-K
Name writing
Total score c
.45** .29**
Letter formation d
.43** .24**
Vertical orientation .38** .26**
Left–right orientation .26** .22**
Correct letter sequence .44** .31**
Letter writing
Total score c
.60** .44**
Letter formation d
.59** .42**
Vertical orientation .58** .43**
Spring kindergarten
Name writing
Total score e
.22**
Letter formation f
.20**
Vertical orientation .23**
Left–right orientation .14**
Correct letter sequence .17**
Letter writing
Total score e
.44**
Letter formation f
.39**
Vertical orientation .43**
z test of proportions: a
p = .171; b
p = .186; c
p = .013; d
p = .010; e
p = .003; f
p = .013
** p \ .01
Evidence of alphabetic knowledge 141
123
ranged from .14 to .60 (p \ .01). z tests of proportions (shown in Table 3) revealed that correlations were larger between letter writing total scores and
letter writing formation scores with the WJ-III Letter–Word Identification scores
in spring of preschool and spring of kindergarten compared to correlations
between name writing total scores and name writing formation scores and the
WJ-III scores. Correlations for letter writing scores were not different from those
of name writing scores and WJ-III Letter–Word Identification scores in fall of
preschool.
Associations between measurements of children’s writing skills from the
beginning of preschool and their growth on scores from the W-J-III Letter–Word
Identification from the beginning of preschool were also examined. In
these analyses, latent growth curve modeling was used rather than hierarchical
linear modeling despite the presence of multiple classrooms across which the
children were distributed. This decision was made for two reasons. First, writing
activities were observed to be very similar across preschool classrooms,
reflecting common program elements characteristic of these public pre-k and
Head Start classrooms. Second, children were widely dispersed across kinder-
garten classrooms, resulting in situations where typically fewer than 5 children
were in each classroom.
A series of latent growth curve models were estimated to examine associations
between writing scores and W-J-III Letter–Word Identification scores. Hu and
Bentler (1999) and Marsh, Hau, and Wen (2004) recommend using a family of fit
indices to capture the most accurate assessment of fit. Therefore, we used NFI, RFI,
TLI, and CFI (values greater than .95 indicate good fit; Hu & Bentler, 1999), and
RMSEA (values under .05 indicate good fit; Hu & Bentler, 1999).
Latent growth curve modeling provides estimations of a latent intercept (initial
status) and slope (growth over time). In the current model, intercept and slope
were indicated by Letter–Word Identification scores in the fall of preschool,
spring of preschool, and spring of kindergarten. Given the uneven spacing across
time points, parameterizations for the slope manifest variables were set to 1, 1.5,
and 2.5. Children’s writing skills were hypothesized to predict the intercept and
slope for children’s Letter–Word Identification scores. To begin, two full models
were estimated. The first model included paths from children’s name writing
formation and letter writing formation scores in the fall of preschool to the latent
intercept and slope. The second model included paths from children’s name
writing total and letter writing total scores in the fall of preschool to the latent
intercept and slope. Next, four nested models were tested to determine the most
parsimonious models with the best fit, based on incremental changes in chi-square
values and model fit indices when comparing the two full models with their
respective nested models. Nested models each contained only one predictor. For
the first full model (total writing model), one nested model included only name
writing total scores, and the other included only letter writing total scores. For the
second full model (letter formation model), one nested model included only name
writing formation scores, and the other included only letter writing formation
scores.
142 V. J. Molfese et al.
123
Total writing model
The full model with both name writing total and letter writing total scores reflected a
good fit (NFI = .96, RFI = .94, IFI = .97, TLI = .95, CFI = .97, RMSEA = .10).
Parameter estimates are shown in Table 4. Consistent with findings from the first
model with both name letter formation and letter formation scores, the intercept was
statistically significant (p \ .001), indicating that children’s average Letter–Word Identification scores in the fall of preschool were 253.31. In addition, the variance of
the intercept (101.64) was also statistically significant (p \ .001), indicating that children’s initial Letter–Word Identification scores varied substantially. The
estimate for the slope indicated that children’s rate of growth in Letter–Word
Identification scores was positive and statistically significant (45.47, p \ .001). When applied to the parameterizations for growth across all three time points
(i.e., 1, 1.5, and 2.5), children’s Letter–Word Identification scores increased
(1.5 9 45.47) to 68.21 points from fall to spring of preschool, and (2.5 9 45.47) to
113.68 points from fall of preschool to spring of kindergarten. The variance of the
slope was also statistically significant (29.81), indicating that students varied widely
in their growth in Letter–Word Identification scores from fall of preschool to spring
of kindergarten.
Writing indicators for the full model
Standardized estimates of the direct effects of name writing total and letter writing
total scores on the intercept and slope factors are shown on the bottom of Table 4.
Both were significant and positively related to the initial status of Letter–Word
Identification, indicating that children with higher name writing total scores and
higher letter writing total scores were more likely to have higher Letter–Word
Identification scores in the fall of preschool. However, neither was significantly
related to children’s growth in letter–word identification across preschool and
kindergarten.
Table 4 Growth curve parameter estimates and standardized effects of name letter and letter formation scores on intercepts and slopes for letter–word identification
Full letter formation model Name letter formation only Letter formation only
Intercept
factor
Slope
factor
Intercept
factor
Slope
factor
Intercept
factor
Slope
factor
Intercept
estimate
250.59*** 45.43*** 258.37*** 45.31*** 255.73*** 44.70***
Predictors Standardized estimates
Name letter formation .41*** -.18 .66*** -.19 – –
Letter formation .49*** -.03 – – .71*** -.12
*** p B .001 (standardized estimates)
Evidence of alphabetic knowledge 143
123
Nested models for total writing scores
As stated above, a series of nested models, each with different combinations of
predictor variables, were tested to determine the most parsimonious model of
children’s Letter–Word Identification score growth in preschool and kindergarten.
The first nested model included only name writing total scores as a predictor of initial
status and growth in Letter–Word Identification. Fit indices (see Table 4) suggested
the model fit the data well (NFI = .97, RFI = .96, IFI = .98, TLI = .98, CFI = .98,
RMSEA = .08) and the chi-square difference score indicated this model provided
statistically significantly improved fit with the data compared to the full model
(v2D = 8.64, df D = 1, p \ .01). Name writing total scores were significant and positively related to children’s initial Letter–Word Identification scores (intercept),
but were unrelated to children’s growth in Letter–Word Identification (slope).
The second nested model included only letter writing total scores as predictors
of initial status and growth in Letter–Word Identification. Fit indices suggested
the model fit the data satisfactorily (NFI = .95, RFI = .94, IFI = .96, TLI = .96,
CFI = .96, RMSEA = .11), but the chi-square difference score indicated this
model provided significantly poorer fit with the data compared to the full model
(v2D = -.62, df D = 1, p [ .05). Even so, letter writing total scores were significant and positively related to children’s initial Letter–Word Identification
scores (intercept), but unrelated to children’s growth in Letter–Word Identification
(slope).
Letter formation model
The full model with both name writing formation and letter writing formation scores
provided good fit (NFI = .96, RFI = .94, IFI = .97, TLI = .95, CFI = .97,
RMSEA = .10). Parameter estimates are shown in Table 1. The intercept was
statistically significant (p \ .001), indicating that children’s average Letter–Word Identification scores in the fall of preschool were 250.59. In addition, the variance of
the intercept (117.09) was also statistically significant (p \ .001), indicating that children’s initial Letter–Word Identification scores varied substantially. However, of
key interest in latent growth curve modeling is the rate of growth. The estimate for
the slope indicated that children’s rate of growth in Letter–Word Identification was
statistically significant (45.43, p \ .001) and positive. When applied to the parameterizations for growth across all three time points (i.e., 1, 1.5, and 2.5), we
see that children’s Letter–Word Identification scores increased (1.5 9 45.43) to
68.15 points from fall to spring of preschool, and (2.5 9 45.43) to 113.58 points
from fall of preschool to spring of kindergarten. The variance of the slope was also
statistically significant (28.71), indicating that children varied widely in their growth
in Letter–Word Identification scores from fall of preschool to spring of kindergarten.
Writing indicators for the full model
Standardized estimates of the direct effects of name writing formation and letter
writing formation scores on the intercept and slope factors are shown on the bottom
144 V. J. Molfese et al.
123
of Table 4. Both were significant and positively related to the initial status of
Letter–Word Identification scores, indicating that children with higher name writing
formation and letter writing formation scores were more likely to have higher
Letter–Word Identification scores in the fall of preschool. However, none of the
predictor variables were significantly related to children’s growth in Letter–Word
Identification across the three measurement points in preschool and kindergarten.
Nested models for letter formation
As stated above, a series of nested models, each with different combinations of
predictor variables, were tested to determine the most parsimonious model of
children’s Letter–Word Identification score growth in preschool and kindergarten.
The first nested model included only name writing formation scores as a predictor of
initial status and growth in Letter–Word Identification scores. Fit indices suggested
the model fit the data well (NFI = .97, RFI = .96, IFI = .98, TLI = .97,
CFI = .98, RMSEA = .08) and the chi-square difference score indicated this
model provided statistically significant improved fit with the data compared to the
full model (v2D = 6.99, df D = 1, p \ .01). Name writing formation scores were significant and positively related to children’s initial Letter–Word Identification
scores (intercept), but were unrelated to children’s growth in Letter–Word
Identification (slope).
The second nested model included only letter writing formation scores as a
predictor of initial status and growth in Letter–Word Identification scores. Fit
indices suggested the model fit the data satisfactorily (NFI = .95, RFI = .94,
IFI = .96, TLI = .96, CFI = .96, RMSEA = .11), but the chi-square difference
score indicated this model provided significantly poorer fit with the data compared
to the full model (v2D = .26, df D = 1, p [ .05). Even so, letter formation was significant and positively related to children’s initial Letter–Word Identification
scores (intercept), but unrelated to children’s growth in Letter–Word Identification
(slope).
Discussion
There are three main findings from this study: children’s writing skills reflected a
wide range of scores across the three time periods under study, the different
components of the scoring rubrics applied to name writing and to letter writing were
highly correlated, and writing scores are related to the children’s scores on the
Woodcock–Johnson Letter–Word Identification subscale but not to the growth of
these scores.
Young children’s knowledge of writing was found to reflect a wide range of skills
in writing their names and in writing letters. This range in writing skills is reflected
in all of the writing components as well as the overall total scores, which are
summed component scores. The scores range from 0 to the maximum (or nearly to
the maximum) score possible and the wide range of scores is evident in the fall and
spring of preschool as well as in kindergarten, and for both name writing and letter
Evidence of alphabetic knowledge 145
123
writing. Although some children reached ceiling by kindergarten, the range of
scores at each measurement point reflects the usefulness of these name and letter
writing tasks for understanding children’s knowledge of writing even in kindergar-
ten. The mean scores for name writing total scores and letter formation scores and
the mean scores for letter writing total scores and letter formation scores reflect age
differences, where the scores at each older age are higher than those at the younger
ages, as would be expected. These developmental changes reflect the benefits gained
at least in part from exposures to classroom-based writing activities. All classrooms
had writing centers and name writing was a common activity, but the wide range of
skills reflected in the children’s performance, including observations of children’s
name writing and letter writing where no or few letters were written, is evidence that
not all children even at the mid-year point in kindergarten have good name or letter
writing knowledge.
The second main finding is the consistently strong correlations between the
components of name writing (letter formation, vertical orientation, left–right
orientation, and correct letter sequence) and between the components of letter
writing (letter formation and vertical orientation). It does not appear that scoring the
multiple components provides much in the way of additional information beyond
that provided by the letter formation component alone. Diamond et al. (2008)
reported similar findings for Head Start preschoolers. In that study, name writing
was scored on a 9-point scale ranging from 1 = scribbling to 6 = writing included
letters and letter like shapes to scores of 7–9 that reflected name writing composed
of only letters. Additional ‘‘process’’ aspects of name writing were scored on a 6-
point scale to reflect horizontal and linear orientation and case. The letter writing
and process scores were highly correlated (r = .77) suggesting that the scores based on the letter formation aspects of name writing and those based on other features are
highly overlapping components of name writing in this preschool aged group.
Our findings extend findings of Diamond et al. (2008) with preschoolers by
showing that scoring rubrics that include letter formation and other features of
writing are also strongly correlated in kindergarten. The different components of
name and letter writing included in the scoring rubrics used in many studies appear
to be part of children’s holistic repertoire of writing skills and the components do
not appear to represent independent features of writing knowledge nor do they
appear to develop separately. Bloodgood’s (1999) findings from a study of 3- to 6-
year-old children support this view. She used a 7-point scale to score name writing
skills and reported strong correlations at each age (r = .69–.92) between name writing scores obtained when children spontaneously wrote their names and when
asked to write each letter of their name as it was dictated. Thus, the complex,
multiple component scoring rubrics that have been used in some studies compared
to single component rubrics do not identify unique writing skills and do not add
much to our understanding of the development of young children’s writing
knowledge.
These studies share in common the use of tasks that specifically require children
to write their names or write letters rather than using a more general writing task,
such as having children write a word, sentence, or story, all of which have been
used with preschool and kindergarten children (Bloodgood, 1999; Bus et al., 2001;
146 V. J. Molfese et al.
123
Levin et al., 2005; Louvet-Schmauss & Preteur, 1993). However, the use of writing
tasks, such as writing letters or words on demand, that explicitly require children to
attend to the relation between the sounds of letters and letter combinations in
writing graphic representations of what they heard come closer to reflecting the
critical association between writing skills and alphabetic knowledge than do tasks,
including name writing tasks, that may only engage copying or rote memory skills.
The former tasks may be more useful than the latter tasks for gaining an
understanding about the role of writing in language, spelling, and reading.
The final main finding is that the writing scores are related to the children’s
scores on the Woodcock–Johnson Letter–Word Identification subscale in preschool
and kindergarten. There are larger correlations for letter writing scores compared to
name writing scores at all measurement points with concurrent measures of WJ-III
Letter–Word Identification. However, the growth curve analyses show that writing
scores are not predictive of the growth of Letter–Word Identification scores from
preschool to kindergarten. This latter finding is unexpected because other
researchers have reported bi-directional influences of letter writing and changes
in letter knowledge skills (Bloodgood, 1999; Diamond et al., 2008; Molfese et al.,
2006). However, the current study has a limitation in measurement time points.
With three measurement time points needed for the dependent variable in the
growth curve analyses, it required the use of writing scores from the fall of
preschool as predictors of letter/word identification scores. A stronger predictor
likely would have been the writing scores from the spring of preschool when the
children had a full year of opportunities to engage in the writing, language and
literacy activities included in their preschool curricula. Models using these spring
writing scores would require letter/word identification scores to be obtained from
the children in 1st grade to have sufficient measurement time points.
The fit statistics from the growth curve analyses provide additional information
on the scoring rubrics used for the name and letter writing tasks in this study. The fit
statistics for the two full models (total writing model and letter formation model) are
very similar for the naming writing and letter writing data and indicate that both
models fit the data equally well. However, the fit statistics of the nested models
show that the name writing scores fit the models slightly better than the letter
writing scores. This finding may reflect the special significance of the first name to
young children described in other studies (Green, 1998; Treiman & Broderick,
1998; Welsch et al., 2003) and the commonality of name writing tasks across
preschool classrooms that provide children with experience in writing their names.
However, we argue that letter writing, especially writing dictated letters rather than
copying letters, may be more advantageous to the development of alphabetic skills
because of the involvement of both letter name knowledge and letter sound
knowledge as described by Molfese et al. (2006) and Shatil et al. (2000).). Further
research is needed involving more than the WJ-III Letter–Word Identification
assessment. To better understand relations between young children’s knowledge of
letter writing and their alphabetic knowledge, assessments must include measures of
letter sound knowledge as well as letter name knowledge, such as the tasks used by
Diamond et al. (2008), and measures of word reading skills assessed with a larger
number of age-appropriate words in addition to basing the writing assessment on
Evidence of alphabetic knowledge 147
123
more writing tasks. Such assessments might provide stronger evidence for the
influence of writing on the growth of letter and word knowledge from preschool to
kindergarten, which could not be identified from the results of the present study.
In conclusion, the findings from the present study suggest that focusing on name
writing as the predominant writing activity in preschool classrooms may not benefit the
growth of young children’s alphabetic knowledge in the critical years of early
education. Current practice limits the benefits that young children can gain from writing
by focusing writing activities on procedural knowledge rather than on conceptual
knowledge gained from understanding how letter sounds are represented by written
letters and experience with the visual characteristics of written letters and words, such as
explored by Treiman, Cohen, Mulqueeny, Kessler, & Schechtman (2007).
Appendix 1: Scoring rubric for name writing
Letter Formation (without respect to orientation or sequence):
0 No attempt or refusal
1 Wrote something, including drawings, scribbles, or random letters
2 Wrote one or two letters of the name in a recognizable manner regardless of
form
3 Wrote one or two letters of the name with good form (easily recognizable
alphabet letter; each component correctly placed; e.g., vertical line of lower-case
d is oriented to the right, touching the loop not connected to the center or
displaced from the loop).
4 Wrote several letters of the name in a recognizable manner regardless of form
5 Wrote several letters of the name with good form
6 Wrote all letters of the name in a recognizable manner regardless of form
7 All letters of the name were written with good form
Letter Formation Total: ______
Orientation of name letters on the vertical axis (score only non-mirror letters):
0 None correct (reversed or upside down), no attempt or refusal
1 Wrote one or two letters without reversal or upside down
2 Wrote several letters without reversal or upside down
3 Wrote all letters of the name without reversal or upside down
Vertical Orientation Total: ______
Left to Right Orientation of name letters:
0 Wrote name from right to left, refusal, or scribble
1 Wrote name from left to right or wrote only first letter
Left/Right Orientation Total: ______
Sequence:
0 Scribble, first letter not written in first position, or refusal
1 First letter of name in 1st position on the left
148 V. J. Molfese et al.
123
2 More than the 1st letter in correct position, with no additional letters inserted
between
3 All letters of the name in correct position, with no additional letters inserted
anywhere
Sequence Total: ______
Name Writing Total Score ______
Appendix 2: Scoring rubric for letter writing
Letter Formation (without respect to orientation):
0 No attempt or refusal
1 Wrote something, including drawings, scribbles, or random letters
2 Wrote one or two letters in a recognizable manner regardless of form
3 Wrote one or two letters with good form
4 Wrote several letters in a recognizable manner regardless of form
5 Wrote several letters with good form
6 Wrote all letters in a recognizable manner regardless of form
7 All letters were written with good form
Letter Formation Total: ______
Orientation on the vertical axis (score only non-mirror letters):
0 No attempt or refusal, scribble, or random letters
1 Wrote one or two letters without reversal
2 Wrote several letters without reversal
3 Wrote all letters without reversal
Vertical Orientation Total: ______
Letter Writing Total Score: ______
References
Bloodgood, J. W. (1999). What’s in a name? Children’s name writing and literacy acquisition. Reading Research Quarterly, 34, 342–367.
Bus, A. G., Both-de Vries, A., de Jong, M., Sulzby, E., de Jong, W., & de Jong, E. (2001).
Conceptualizations underlying prereaders’ story writing (CIERA Report #2-015). Ann Arbor, MI: University of Michigan, School of Education, Center for Improvement of Early Reading
Achievement.
Clay, M. (2001). Change over time in children’s literacy development. Portsmouth, NH: Heinemann. Denton, K., & West, J. (2002). Children’s reading and mathematics achievement in kindergarten and first
grade. Washington, DC: US Department of Education, National Center for Education Statistics. Diamond, K., Gerde, H., & Powell, D. (2008). Development in early literacy skills during the pre-
kindergarten year in head start: Relations between growth in children’s writing and understanding of
letters. Early Childhood Research Quarterly, 23, 467–478. Green, C. R. (1998). This is my name. Childhood Education, 74, 226–231.
Evidence of alphabetic knowledge 149
123
Haney, M. R., Bissonnette, V., & Behnken, K. L. (2003). The relationship among name writing and early
literacy skills in kindergarten children. Child Study Journal, 33, 99–115. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. Juel, C. (2006). The impact of early school experiences on initial reading. In S. Neuman & D. Dickinson
(Eds.), Handbook of early literacy research (Vol. 2, pp. 437–447). New York, NY: Guilford Press. Levin, R., Both-DeVries, A., Aram, D., & Bus, A. (2005). Writing starts with own name writing: From
scribbling to conventional spelling in Israeli and Dutch children. Applied Psycholinguistics, 26, 463–477.
Louvet-Schmauss, E., & Preteur, Y. (1993). Conceptualization of the writing system and knowing how to
use a children’s book at preschool age as predictors of reading and writing acquisition in the first
year of primary school: A comparative study between France and Germany. European Journal of Psychology of Education, 8, 221–234.
Lyytinen, H., Aro, M., Eklund, K., Erskine, J., Guttorm, T. K., & Laakso, M. L. (2004). The development
of children at familial risk for dyslexia: Birth to school age. Annals of Dyslexia, 54, 185–220. Marsh, H. W., Hau, K. T., & Wen, Z. L. (2004). In search of golden rules: Comment on hypothesis testing
approaches to setting cutoff values for fit indexes and dangers in overgeneralising Hu & Bentler
(1999) findings. Structural Equation Modeling, 11, 320–341. Molfese, V., Beswick, J., Molnar, A., & Jacobi-Vessels, J. (2006). Alphabetic skills in preschool: A
preliminary study of letter naming and letter writing. Special Issue: Developmental Neuropsychol- ogy, 29, 5–19.
Muter, V., & Diethelm, K. (2001). The contribution of phonological skills and letter knowledge to early
reading development in a multilingual population. Language Learning, 51, 187–219. National Early Literacy Panel (Eds) (2008). Developing early literacy: Report of the National Early
Literacy Panel. Washington, DC: National Institute for Literacy. Riley, J. (1995). The relationship between adjustment to school and success in reading by the end of the
reception year. Early Child Development and Care, 114, 25–38. Shatil, E., Share, D., & Levin, I. (2000). On the contribution of kindergarten writing to grade 1 literacy: A
longitudinal study in Hebrew. Applied Psycholinguistics, 21, 1–21. Sulzby, E., Barnhart, J., & Hieshima, J. (1989). Forms of writing and rereading from writing: A
preliminary report. In J. Mason (Ed.), Reading and writing connections (pp. 31–63). Needham Heights, MA: Allyn and Bacon.
Treiman, R., & Broderick, V. (1998). What’s in a name? Children’s knowledge about the letters in their
own names. Journal of Experimental Child Psychology, 70, 97–116. Treiman, R., Cohen, J., Mulqueeny, K., Kessler, B., & Schechtman, S. (2007). Young children’s
knowledge about printed names. Child Development, 78, 1458–1471. Weinberger, J. (1996). A longitudinal study of children’s early literacy experiences at home and later
literacy development at home and school. Journal of Research in Reading, 19, 14–24. Welsch, J. G., Sullivan, A., & Justice, L. M. (2003). That’s my letter!: What preschoolers’ name writing
representations tell us about emergent literacy knowledge. Journal of Literacy Research, 35, 757–776.
West, J., Denton, K., & Germino-Hausken, E. (2000). America’s kindergarteners. US Department of Education, National Center for Education Statistics. NCES 2000-070. Retrieved March 8, 2004 from
http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2000070.
Wilkinson, G. (1993). The wide range achievement test-3. Wilmington, DE: Jastak Associates. Woodcock, R., McGrew, K., & Mather, N. (2001). Examiner’s manual, Woodcock–Johnson III tests of
cognitive ability. Itasca, IL: Riverside Publishing.
150 V. J. Molfese et al.
123
Copyright of Reading & Writing is the property of Springer Science & Business Media B.V. and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.
