Abstract
Learning to write legibly and efficiently is an important occupation for school-age children. The quality of handwriting may influence academic outcomes, with higher marks assigned for neatly written papers (Sweedler-Brown, 1992). The nature and extent of instruction affects a student’s ability to perform skilled handwriting, with explicit instruction increasing the legibility and efficiency of written work (Goldberg & Simner, 1999), as well as improving composition skills (Berninger et al., 1997; Graham, Harris, & Fink, 2000). Cursive script, in which the letters within words are connected by joined strokes, has been traditionally reported to be the form of writing preferred for efficiency in legibility and speed, although controversy does exist (Graham, Weintraub, & Berninger, 1998).
Legibility of cursive writing has been evaluated using global rating scales that compare the individual’s performance with a series of model specimens (Feder, Majnemer, & Synnes, 2000). Legibility has also been judged in terms of components: slant, alignment, spacing, size, and letter formation (Feder & Majnemer, 2007). Speed has been evaluated using letters per minute, but the use of varied instructions and task demands confounds the ability to determine typical writing speeds across ages (Graham, Berninger, Weintraub, & Schafer, 1998).
Some students struggle in their written work, and teachers typically estimate that 12% of children in their class have difficulties with handwriting (Barnett, 2005). Boys have primarily been identified with poor handwriting skills (Berninger et al., 1997; Graham et al., 2000), as well as children with motor coordination or learning disabilities (Jongmans, Linthorst-Bakker, Westenberg, & Smits-Engelsman, 2003; Missiuna et al., 2008). Graham and Harris (2005) indicated that inadequate instruction results in writing challenges. Handwriting difficulties may lead children to avoid writing (Berninger et al., 1997; Graham et al., 2000) and may affect a student’s motivation, self-image, and academics (Piek, Baynam, & Barrett, 2006).
Handwriting difficulties are cited as one of the most frequently mentioned reasons for student referral to school-based occupational therapists (Feder et al., 2000; Missiuna et al., 2008). Studies have found that handwriting can be improved after 9–10 hr of individualized occupational therapy intervention (Case-Smith, 2002; Peterson & Nelson, 2003). Feder et al. (2000) found that 56% of occupational therapists surveyed provided handwriting remediation sessions on a weekly basis.
Several studies have described features of remediation programs to help develop handwriting skills. Laszlo and Broderick (1991) reported that kinesthetic instruction had a positive effect on the handwriting of students who had difficulties, whereas Sudsawad, Trombly, Henderson, and Tickle-Degnen (2002) found that handwriting difficulties could not be improved by means of kinesthetic training. Studies have supported reintroduction of the letter forms explaining each form visually and verbally (Karlsdottir, 1996), self-instruction (Jongmans et al., 2003), random practice (Ste.-Marie, Clark, Findlay, & Latimer, 2004), and model/review of stroke direction before practice (Berninger et al., 1997). Little consensus has been reached on the most effective strategy for remediation, and further research on current remediation practices has been recommended (Asher, 2006; Bonney, 1992).
Many occupational therapists recommend the program Loops and Other Groups: A Kinesthetic Writing System (Benbow, 1990). This program focuses on the motor aspects of cursive handwriting, with letters taught in groups that share movement patterns. It combines sensorimotor techniques, along with letter formation practice, and includes modeling and verbal analysis of letters, motor learning through tracing, revisualization, verbal self-guidance, handwriting from memory, and self-assessment of letters most accurately produced. We could not find any peer-reviewed studies evaluating the effectiveness of this program; therefore, the need to evaluate its effectiveness is evident.
Objectives
Our primary objective was to determine whether children who participate in a kinesthetic writing intervention would improve significantly in legibility of cursive handwriting. Secondary objectives were to determine whether children who participate in a kinesthetic writing intervention would (1) improve in speed of handwriting at 4 mo postintervention and (2) improve in their personal satisfaction with handwriting.
Method
Research Design
A repeated measures design, with four time points, was used to evaluate change in legibility, speed, and personal satisfaction with handwriting over time. The study received ethics approval from the health region and two school district ethics review boards.
Participants and Recruitment
A convenience sample was obtained by contacting 72 schools in Calgary and outlining the study objectives and protocol. Participants were identified by their teachers as having handwriting problems, determined on the basis of the student’s handwriting performance within the classroom setting. We screened the students through parent and caregiver contact. Students who had moderate to severe physical limitation, identified cognitive impairment, or persistent behavioral concerns; were receiving occupational therapy services for handwriting difficulties; or whose parents did not give informed consent were excluded.
Instruments
Handwriting Skill.
Three handwriting samples were collected from each participant to measure skill:
Copying: the phrase “The quick brown fox jumped over the lazy dogs” (Sovik, 1975);
Composition: the Handwriting Subtest from the Test of Written Language (Hammill & Larsen, 1983); and
Alphabet samples: both connected and unconnected.
Handwriting Quality and Speed.
We used the following methods of rating the quality and speed of the handwriting samples:
The Test of Written Language (TOWL) Handwriting Subtest Rating Scale rates global legibility of a student’s composition by comparing it with a series of graded specimens with a value ranging from 0 to 10, which is then standardized on a scale ranging from 0 to 20. Test–retest reliability is reported at .89, interscorer reliability at .76, and criterion validity with teacher ratings at .46.
The Handwriting Evaluation Scale (HES; Malloy-Miller, 1985) evaluates legibility by direct analysis of handwriting errors, on the basis of definitions for seven components of spacing within words, spacing between words, size of letters within words, size between words, baseline orientation, closure, and line quality. A percentage correct score (0%–100%) is obtained for each of the components. This instrument is reported to have face validity by using handwriting errors directly related to in-school handwriting tasks (Malloy-Miller, Polatajko, & Ansett, 1995). Interrater reliability of the HES has been reported at r = .94 (Roberts & Samuels, 1993).
Speed was evaluated on the copying sample, with the student writing the phrase repeatedly until 2 min had elapsed (Sovik, 1975). Letters per minute were calculated.
Personal Satisfaction With Handwriting.
The Attitude Scale was developed specifically for this study to measure personal satisfaction on the basis of the semantic differential, as outlined by Mueller (1986). It consists of seven evaluative adjective pairs specific to handwriting. Statements are provided, and the student places an X on one of seven spaces on a visual analog scale that has an adjective at each end. An example statement is, “I feel my handwriting is . . .” with the evaluative pair ugly and beautiful at either end of the scale. The total score on this measure ranges from 7 to 49 points. We piloted the scale with 122 Grade 5 and Grade 6 students, and a descriptive analysis of the data, using histograms, successfully showed a range in responses for each question. The Student Inventory (Alberta Children’s Hospital, 2001) is a nonstandardized attitude scale that includes eight questions, each with five pictures that range from a picture of a very happy dog (rated as 1) to one of a very unhappy dog (rated as 5). An example of a question is, “How do you feel about how neatly you write?” The total score ranges from 8 to 40 points.
Parent and Teacher Reports.
Questionnaires for parents and teachers were designed for this study. Open-ended questions determined the type and amount of instruction and handwriting practice in which students participated at home and school. Nine questions regarding attitude, speed, and legibility were rated on a 10-point visual analog scale; 0 indicated a poor response, and 10 indicated a good response (e.g., “How would you rate the legibility of your child’s [this student’s] written work?”). The teacher questionnaire included a 10th question asking whether it was expected that written work be completed in cursive handwriting.
Procedure and Data Collection
Testing Procedure.
Testing consisted of three handwriting samples and the two measures of personal satisfaction and took place 4 times during this study. Parents and teachers completed the questionnaires concurrently.
Intervention Procedure.
The remediation program was based on the Loops and Other Groups program and was provided by an occupational therapist and a therapy assistant. The program was extended from a 6-wk intervention to a 7-wk intervention by dividing one session, in which several new letters were introduced, into two sessions. Small groups of students were seen once a week for an hour after school in a quiet room in the therapy area of a local children’s hospital. The students were evaluated to ensure correct seating at a table as well as functional finger and wrist posture. Pencil grips were provided as required. Hand and arm activities occurred for 10 min at the beginning of each session in preparation for handwriting. The letters of the lowercase alphabet were taught in four groupings according to shared movement patterns. These included “clock climbers” (a, d, g, q, c), “kite strings” (i, u, w, t, j, p, r, s, o), “loop group” (h, k, b, f, l, e), and “hills and valleys” (n, m, v, y, x, z). During the remediation program, the student received homework sheets to be completed nightly with their parents for 15–20 min.
Data Collection.
After Test 4, the composition samples from each of the test sessions (with the date of testing excluded) were provided to the caregiver, the teacher, and an occupational therapist to rate global legibility using the TOWL. Scoring instructions were provided. The occupational therapist involved in the data collection was not the same occupational therapist who administered the intervention.
Interrater reliability of the HES was completed by two occupational therapists who independently scored the same three handwriting samples, after agreeing on scoring criteria. Pearson’s product–moment correlation coefficients ranged from .78 to .96 for the seven components; the overall r = .86. The two therapists, blind to the time of testing, scored half of the copying, composition, and alphabet samples independently, calculating the percentage of correct responses for each of the seven legibility components.
Sample Size.
The sample size was estimated using the paired t test of mean differences to allow for analysis of within-participant changes preintervention and postintervention. Using an α of .05 and power of 80%, a sample size of 30 was chosen because it would detect a minimum change of 3.0 standard scores in the TOWL (standard deviation [SD] = 2.0), detect a minimum change of 5.0 points in the HES (SD = 10.0), and allow for stratification by grade and gender.
Data Analysis.
Data were entered into Microsoft Excel 2003 spreadsheets, scored with a higher score relating to a more positive outcome, and then transferred into Stata S/E, Version 9.2 (StataCorp, College Station, TX) for analysis. Data were checked and corrected for entry errors by producing the range, frequency, and histogram for each variable; no data required elimination as a result. Descriptive analyses were performed to examine the characteristics of the children who participated in the study and all outcome measures. Categorical variables were expressed as frequencies and percentages, and continuous variables were reported as means with standard deviations.
For the primary objective, graphical displays of the median value at each time point were constructed for the TOWL and the HES for composition, copying, and alphabet samples. The occupational therapist’s TOWL scores were used because the return rates from the caregiver and teacher ratings were too small to be included. The distribution of the TOWL and all components of the HES scores were each assessed individually for normality using the Wilk–Shapiro test (Shapiro & Wilk, 1965). The normality assumption was violated; therefore, the nonparametric Wilcoxon signed-rank test (Wilcoxon, 1945) was used to examine the paired data for all outcome measures (i.e., TOWL and HES).
Four separate statistical comparisons were carried out to test the impact of the handwriting intervention: (1) Test 1 results were compared with Test 2 results to establish any natural improvement in handwriting technique over time without the intervention, (2) Test 2 results were compared with Test 3 results to determine whether the intervention was effective, (3) Test 3 results were compared with Test 4 results to assess long-term improvement after the handwriting intervention, and (4) Test 4 results were compared with Test 1 results to determine improvement from the start of the study to 4 mo postintervention.
To define a clinical indicator of success in handwriting improvement, the five graded specimens provided in the TOWL manual were examined a priori, and an increase of 3.0 points on the TOWL standard scale (8–11 yr) was set as a clinically important change from a less to a more legible specimen. Differences in demographic variables (e.g., gender, grade), attendance at intervention, and reported homework practice were examined using Fisher’s exact test (Agresti, 1992; Fisher, 1922) for a difference in proportions from Test 1 to Test 4 on the TOWL.
For the secondary objectives, speed and satisfaction with handwriting were both examined using the nonparametric Wilcoxon signed-rank test at the various time points, and graphical displays of the median change over time were constructed. Speed by grade level, as well as parent and teacher reports, were examined descriptively with the mean and standard deviation reported for each time point.
All statistical tests for both primary and secondary objectives were two-tailed, with p values of .05 considered to be statistically significant. Children’s samples were excluded from all analyses if students did not attend a minimum of four treatment sessions. Analyses were based on available data from all instruments completed in a valid manner. Several students printed, scribbled, or left blanks during the testing sessions; therefore, the denominator varied for each measure at each time point.
Results
Recruitment, Participation, and Baseline Characteristics
Forty-two students from 28 schools were recruited to the study, and 32 students attended four or more of the treatment sessions. Six group interventions occurred, with a range of 3–7 students per group. There were more boys (84%) than girls (16%), with an average age of 10.5 yr (range = 8 yr, 7 mo, to 11 yr, 11 mo). One-third of the students were in each of Grades 4 (n = 11), 5 (n = 11), and 6 (n = 10). Although 25 students completed the postintervention testing session, only 18 of those students produced valid handwriting samples.
Primary Outcome: Legibility of Handwriting
Global Legibility.
Table 1 and Figure 1 show the statistical comparisons and median scores, respectively, on the TOWL. Scores increased during the intervention period, and gains were maintained. There were improvements in handwriting from Test 1 to Test 4. TOWL scores showed that 39% of students (7 of 18) improved by a score of 3 or more from Test 1 to Test 4. Seventy-five percent of Grade 6 students (6 of 8), 20% of Grade 5 students (1 out of 5), and 40% of Grade 4 students (2 of 5) improved in their ratings on the TOWL. However, statistically significant differences were not found in grade-level proportions (p = .18), gender (boys = 53%, girls = 33%, p = 1.00), attendance at intervention (5 sessions = 100%, 6 sessions = 57%, 7 sessions = 40%, p = .65), or amount of homework practice reported (<30 min per week = 60%, ≥30 min per week = 45%, p = 1.00).
Global ratings of handwriting legibility on the Test of Written Language Handwriting Subtest Rating Scale.
Statistical Comparisons Over Time for the Test of Written Language (TOWL), the Handwriting Evaluation Scale (HES), and Personal Satisfaction With Handwriting
Note. The nonparametric Wilcoxon signed-rank test was used for all paired comparisons.
Legibility Components—Alphabet Samples.
Table 1 and Figure 2 show the statistical comparisons and median scores, respectively, for letter formation and baseline orientation, closure, line quality, size, and space (only for connected alphabet) components using the HES. There were no changes in the median scores from Test 1 to Test 2, except in closure for the connected alphabet. For the intervention period, there were positive changes for all components in both alphabet samples. There was a decrease in several median scores of measures, including letter formation, size of letters, and line quality in both alphabet samples and in the closure variable for the unconnected alphabet from Test 3 to Test 4. From Test 1 to Test 4, there were significant increases in all components in the unconnected alphabet and in all but size of letters in the connected alphabet.
Component scores for the unconnected and connected alphabet.
Legibility Components—Copying and Composition.
Table 1 and Figure 3 show the statistical comparisons and median HES scores, respectively, for the copying and composition samples. In the copying samples, there were increases from Test 2 to Test 3 in all legibility components except word space. From Test 3 to Test 4, a decrease was noted in the size of letters within words, whereas an increase was noted in line quality. There were increases from Test 1 to Test 4 in the size of letters within words, baseline, and line quality components. In the composition sample, there were increases in all HES components from Test 2 to Test 3 except in letter space and word space. The size of letters within words demonstrated a decrease from Test 3 to Test 4. There was an increase from Test 1 to Test 4 for the closure component. Visual inspection of Figure 3 shows that scores for the size-within-words component were higher for the composition task, whereas scores for line quality were higher for the copying task.
Component scores for the copying and composition tasks.
Secondary Outcome: Speed of Handwriting.
Copying speed increased from Test 1 to Test 2 (t = 2.78, p < .01, N = 25) and from Test 2 to Test 3 (t = 1.98, p < .05, N = 25). From Test 3 to Test 4, gains were maintained, but there were no further increases in speed postintervention (t = –0.87, p = .38, N = 18). An increase was noted between Test 1 and Test 4 (t = 2.11, p = .04, N = 17). Mean letters per minute changed from 18.7 (SD = 16.0) at Test 1 to 23.8 (SD = 15.8) at Test 4.
An increase in average speed of writing between Test 2 and Test 3 was observed for Grades 4 and 5 but not for Grade 6. Increased writing speed was noted for all grades from Test 1 to Test 4, and the greatest gains were observed in Grade 4 students (Table 2).
Handwriting Copying Speed, by Grade
Note. The number of students at each test time varies because of absences or incomplete data. SD = standard deviation.
Secondary Outcome: Personal Satisfaction With Handwriting.
Student Inventory scores showed an increase in personal satisfaction over the preintervention phase. For both measures of personal satisfaction, there was an increase during the intervention period (Test 2 to Test 3) and overall (Test 1 to Test 4; Table 1 and Figure 4).
Attitude Scale and Student Inventory for personal satisfaction with handwriting.
Parent and Teacher Reports.
Classroom instruction and practice was reported by 9 of 24 teachers (38%), ranging from 1 to 5 times per week for 4–30 min per session. Handwriting practice at home was reported by 4 of 32 parents (13%), from 1 to 5 times per week for 5–25 min in length. Table 3 shows increased ratings in the areas of attitude, speed, and legibility on the parent and teacher questions, with the parents reporting greater preintervention to postintervention changes. Both parents and teachers reported the largest increase on the question that rated the child’s feelings toward handwriting. The teacher’s average score for student expectation to complete written work using cursive writing was 3.7 (N = 24) preintervention and 3.5 (N = 11) postintervention.
Parent and Teacher Report of Child’s Attitude, Speed, and Legibility of Written Work
Notes. Scores range from 0 = poor response to 10 = good response; Test 4 teacher report was omitted because of new school year with a new teacher. — = data missing because of administrative error.
Discussion
The greater proportion of boys than girls recruited was representative of the gender difference found in handwriting problems (Berninger et al., 1997). The Loops and Other Groups program, presented in a small-group format, was shown to be effective in improving the participants' skills. More than one-third of the students demonstrated improvement in global legibility of cursive handwriting in composition samples. A significant improvement in the cursive formations of individual letters reflected the primary strategy of the intervention program and is of great importance in legibility (Feder & Majnemer, 2007). Significant improvements were identified in the legibility components of baseline, closure, and line quality, which are reported to relate to kinesthetic feedback (Malloy-Miller, 1985). Parents and teachers reported improvements in aspects of legibility, speed, and attitude.
We found that students improved more in the copying task than in the composition task. In the copying samples, there were significant increases in six handwriting components during the intervention period and in three of the components from preintervention until 4 mo postintervention. In the composition sample, significant increases occurred in five handwriting components during intervention and in one component from preintervention to 4 mo postintervention. Composition is reported to involve “an integration and synthesis of cognitive skills more complex than those needed for a simpler writing task, such as copying” (Amundson, 1992, p. 67). When children focus on the mechanical aspect of handwriting, they may not be able to fully attend to the content of their work and vice versa (Berninger et al., 1997; Graham & Harris, 2005). The copying phrase was used several times in this study, and the students may have developed skill in the formations and joins in this task. The complexity of composition imposes challenges to optimal performance of the legibility components, and consideration should be given to students who have difficulties in this area. More time and instructional probes would be beneficial so that improved legibility could be incorporated gradually into compositional work.
Spacing between words did not improve significantly on either the copying or the composition tasks, a component that the Loops and Other Groups program does not address. Few errors were observed in the spacing-within-words component, and we question whether it should be evaluated when examining handwriting legibility (Roberts & Samuels, 1993). Immediately after treatment, the students wrote larger in the provided line space for the copy task (1.4 cm high), but initially and at 4 mo postintervention, they demonstrated a preference to write smaller, within half of this space, eliciting size errors. Fewer size errors occurred in the composition task in which the line spacing provided for handwriting was smaller (0.8 cm high). Studies of space size with younger students reported more correct letter strokes with wide-spaced paper than with normal-spaced paper (Hill, Gladden, Porter, & Cooper, 1982; Waggoner, LaNunziata, Hill, & Cooper, 1981). Benbow (1990), however, found that Grade 4, 5, and 6 students scored significantly better on the TOWL writing scale when using narrow-ruled paper.
The speed of writing improved initially with the students but did not continue to improve with time, as would be expected with practice of the skill. Even with the significant gains made from preintervention to postintervention, the students continued to be well below the norm for their grade on handwriting speed. The students had average speeds of 18–24 letters per minute, compared with norms for Grade 4 students of 34 letters per minute, Grade 5 of 39 letters per minute, and Grade 6 of 56 letters per minute (Coulter, Pollock, & Lockhart, 1992). This finding is in keeping with other researchers (Jongmans et al., 2003; Volman, van Schendel, & Jongmans, 2006), who found that children with handwriting problems had slower speeds of writing. On the basis of the premise that greater handwriting speed develops with practice (Graham, Berninger, et al., 1998), a period of >4 mo may be required to produce ongoing increases in the speed of handwriting. A reason for the lack of progress during the postintervention phase may be that these students were not expected to practice cursive writing skills in the context of classroom work (Asher, 2006). In this study, low teacher expectation was reported for written work to be completed in cursive handwriting, and the students may not have chosen to practice their newly learned skills.
A higher percentage of Grade 6 students improved in legibility than Grade 4 or 5 students, although this finding was not statistically significant, perhaps because of the relatively small sample size. Graham, Berninger, et al. (1998) found that children’s handwriting became more legible with advancing elementary grades. The Grade 6 students had slower speed of handwriting after the intervention. They may have taken more time to demonstrate improvement in the quality of handwriting. Weintraub and Graham (1998) found that when children were asked to write quickly, there was a corresponding decline in legibility; when asked to write neatly, their speed of handwriting decreased.
Personal satisfaction improved in the students after the intervention program. Moreover, parents and teachers perceived a noteworthy improvement in the students’ feelings toward handwriting. The cooperative, small-group format and the improvements in cursive handwriting skills may have contributed to this improved attitude (Townsend & Hicks, 1997). It is possible that the students reported improved attitudes toward handwriting because they knew that they were not alone in facing their handwriting challenges (Wentzel & Caldwell, 1997).
Future Research
Further research is needed to determine the intervention program—or combination of intervention components—and timing of the intervention that is most effective (Asher, 2006; Berninger et al., 1997). Teacher and student expectations regarding style and use of written output needs to be explored more fully. Examination of this program’s effectiveness when delivered by others, such as teachers, would be valuable. Explicit instruction and practice in the motor aspect of handwriting in the regular classroom has been recommended (Asher, 2006; Berninger et al., 1997; Ste.-Marie et al., 2004), including monitoring of students who present with handwriting challenges (Graham et al., 2000). It may be that some handwriting challenges would be reduced or eliminated with implementation of a program such as Loops and Other Groups in the classroom. Use of a structured program in the classroom may eliminate the need for referral and intervention by occupational therapy for some students. Limited resources could then focus on the students who would benefit most from this service (Asher, 2006).
Limitations
Several limitations, some inherent in intervention studies with children, may have affected the results of this study. The study used a small, convenience sample, which may have been biased by teacher and parent interest in the study. Some teachers may have identified their students as having handwriting difficulties because of problems observed in print style. The sample was not geographically representative of all children in Calgary. It is not known what influence teaching style and instruction in the classroom had on the performance of the participants. Data were incomplete, because not all of the children attended all of the intervention sessions, some children were lost to follow-up, and some test forms were not completed adequately for analysis. Hence, given the small sample size based on available data at each time point and multiple testing required, caution needs to be taken when interpreting the results.
Conclusions
The objective of this study was to determine whether Grade 4–6 students who participated in a kinesthetic writing intervention would improve in legibility, speed, and their personal satisfaction with handwriting. The group of students recruited was primarily boys with handwriting challenges. Global legibility and components of legibility improved, as well as speed of handwriting and personal satisfaction with handwriting. The Loops and Other Groups program appears to have had an impact on developing handwriting skills with these students, particularly in components that have been reported to be related to kinesthetic feedback, with better performance on copying tasks and with narrow-ruled paper. This program can be used and recommended by occupational therapists with greater confidence, with the provision of supplementary instruction in word spacing. Further study is required to confirm these findings. In addition, further evaluation of teacher expectations for cursive handwriting, the ongoing integration of skills into daily compositional tasks, and the use of this program in the classroom would be beneficial.
Footnotes
Acknowledgments
The study team acknowledges the children, parents, and teachers who participated in the study. We thank the Decision Support Research Team, particularly Suzanne Tough, Brenda Wilson, and Karen Tofflemire; intervention teams (Aynsley Wennberg, Chandra Kipfer); research staff (Brigitte Roy, Gina Blumes); and leadership support (Lori Craig, Darlene Winder, Jane Pollack). Financial support of the Alberta Children’s Hospital Foundation and the Occupational Therapy Service and Regional School Health Program of Alberta Health Services is acknowledged.