Faculty of Social and Behavioural Sciences
Graduate School of Child Development and Education
The influence of treatment dosage on
effects of a spelling intervention for
students with and without dyslexia: A
moderated mediation model
Research Master Child Development and Education Research Internship
Joyce C. A. Standaert (10503609) Supervisor: dr. M. van den Boer Second supervisor: dr. E.H. de Bree Date: 12-07-2020
Abstract
Both students with dyslexia and typically developing (TD) students continue to have difficulties with certain aspects of Dutch spelling in high school. In this study the effectiveness of a spelling intervention for verb inflection and visual imprint words was assessed. Additionally, the extent to which dosage variables mediated this effectiveness and the moderating effect of having dyslexia were considered. Participants were 153 eighth grade students (25 with dyslexia). Knowledge of verb inflection and visual imprint words was measured with a pre- and posttest. A repeated measures ANOVA indicated that the intervention was effective for visual imprint words, but not for verb inflection. A mediated moderation model was used to assess the dosage variables and influence of dyslexia. For verb inflection results showed that the percentage of completed exercises was associated with higher posttest scores. This is a surprising result, as the intervention for verb inflection was not effective. No significant relations between dosage variables and the posttest were found for visual imprint words. Dyslexia status did not moderate the relation between dosage variables and the posttest. Dosage variables thus did not impact the effectiveness of the intervention for the visual imprint words in this study. The intervention for verb inflection was not effective, which makes dosage not informative. For future research it is important to measure the effect of dosage variables in a spelling intervention with a longer duration.
Keywords. Spelling intervention, treatment dosage, dyslexia, verb inflection,
The influence of treatment dosage on effects of a spelling intervention for students with and without dyslexia: A moderated mediation model
Out of all school subjects, spelling is highly understudied in both elementary and high school. This is true for both students with and without learning difficulties.
Students’ general spelling abilities are good at the end of primary school, as students have then mastered most of the spelling rules. Yet, this knowledge is not maintained well in high school, because little time is spent on spelling instruction (Van de Gein, 2010). The knowledge of spelling rules seems insufficiently automatized in elementary school, and through the lack of maintenance, spelling ability becomes insufficient for many high school students (Van de Gein, 2010). While many students still experience problems with spelling in high school, few spelling interventions have been designed for high school students. Therefore, this study focuses on a spelling intervention for high school students with and without dyslexia.
Spelling interventions could be useful for all students, but especially for students with learning disorders such as dyslexia. Dyslexia is defined as a specific learning disability with a neurobiological origin (American Psychiatric Association, 2013). Students with dyslexia have problems with accurate and/or fluent word recognition and show poor spelling (American Psychiatric Association, 2013; Lyon, Shaywitz, &
Shaywitz, 2003). The reading and spelling difficulties do not respond well to intervention and are therefore persistent (American Psychiatric Association, 2013). Dyslexia may also influence reading comprehension and vocabulary (Lyon et al., 2003). Eighty percent of students with a learning disorder have dyslexia, making it the most common learning disorder (Shaywitz & Shaywitz, 2003). Worldwide, estimates of the prevalence of
dyslexia are between 5 and 10 percent (Peer, 2001; Peterson & Pennington, 2015; Shaywitz & Shaywitz, 2003). Typically developing (TD) students might experience difficulties with spelling due to a lack of maintenance of their spelling skills. For students with dyslexia, however, spelling difficulty is one of the key components of the disorder. Therefore, students with dyslexia continue to have difficulties with spelling in high school and later in life, even when they maintain their spelling skills (Swanson & Hsieh, 2009).
Dutch Spelling
Dutch is a consistent language when pronouncing written words: (combinations of) letters are mostly pronounced in the same way. However, when spelling spoken words, the Dutch language is more inconsistent, because there are different graphemes for the same phoneme(s) (Borgwaldt, 2003; Bosman & Van Orden, 1997). Therefore, Dutch is considered an asymmetric language, for which reading is easier to learn than spelling (Bosman & Van Orden, 1997). In first grade, students start to learn to spell in Dutch. They first learn the phoneme-grapheme connections. When students have
mastered this knowledge, they are able to spell phonologically consistent words, such as ‘kat’ (cat). Dutch students seem to keep relying on phonological spelling in later grades and keep generalizing this phonemic principle to new words, even if this is incorrect (Landerl & Reitsma, 2005). Most spelling mistakes thus are phonologically correct (e.g. gijt instead of geit (goat)).
In Dutch orthography there are multiple types of inconsistent words, such as analogy-based, rule-based, and visual-imprint words (Keuning & Verhoeven, 2008). These inconsistent words do not follow a phonetic pattern or can be logically deduced
from other known words. Two types of inconsistent words are discussed in this study: verb inflection (rule-based words) and imprint words. Verb inflection and visual-imprint words are taught through explicit instruction primarily in the upper elementary grades. Rule-based and visual-imprint words are among the most difficult spelling categories to learn in Dutch. Although students have good general knowledge of spelling rules at the end of primary school, only eighty percent of students have mastered rule-based words and only twenty percent of students have mastered visual-imprint words at the end of primary school (Keuning & Verhoeven, 2008).
Verb inflection refers to the process where a suffix is added to the stem of a word in order to create accordance with both the tense and person. An example is the verb ‘werken’ (to work), which changes into ‘ik werk’ (I work, stem) and ‘jij werkt (you work; stem + t). In the past tense the morphemes ‘te’ or ‘de’ can be added to the stem, which results in ‘hij werkte’ (he worked). Next to verbs that follow these rules, there are also many verbs that are exceptions to this rule (e.g. the verb ‘lopen’ (to walk) changes in ‘ik loop’ (I walk) and ‘ik liep’ (I walked)). Around 17% of the Dutch verbs are irregular (Tabak, Schreuder, & Baayen, 2005). As most of the Dutch verbs are regular, the intervention in this study focuses solely on regular verbs and the step-by-step plan that can be used to write the correct form. Verb inflection is considered to be one of the most difficult parts of Dutch spelling, mainly because of the presence of many homophones (such as ‘ik vind’ (I find) and ‘hij vindt’ (he finds)) (Assink, 1987).
Visual-imprint words on the other hand, do not follow spelling rules and thus need to be remembered. An example is the grapheme ‘ij’, which can be written as ‘ei’ or ‘ij’. Loan words are another form of visual-imprint words. In Dutch, a substantial number
of words originate from French, German, or English (Van der Sijs, 2009). Loan words retain the spelling from the foreign language, therefore do not follow the Dutch phoneme-grapheme associations, and as such are not phonologically consistent (Van der Sijs, 2009). Dutch examples are the French word ‘bureau’ (desk) and the English word ‘baby’.
Most spelling mistakes made by students with and without dyslexia concern verb inflection or visual imprint words (Desoete et al., 2010; Keuning & Verhoeven, 2008). Previous research found that students with dyslexia make the same mistakes as TD students, except they make significantly more mistakes compared to TD students and they make these mistakes for a longer period of time (Callens, Tops, & Brysbaert, 2012; Tops, Callens, Bijn, & Brysbaert, 2014; Tops, Callens, Van Cauwenberghe, Adriaens, & Brysbaert, 2013). The intervention in this study solely focuses on increasing verb
inflection and visual imprint word skills. Spelling Interventions
In general, there is not a lot of research available on what makes spelling
interventions effective. Next to this, most spelling interventions are designed for students in elementary school. Furthermore, research on literacy interventions for high school students is often focused on reading instead of spelling (Wanzek et al., 2006). Still, some research has been done on effective intervention components for students with spelling disabilities. Concerning the content of the intervention, two forms of intervention have been found to be the most effective: phonics instruction and the cover-copy-compare (CCC) technique. Providing phonics instruction significantly improved spelling skills (Galuschka, Ise, Krick, & Schulte-Körne, 2014). Phonics instruction entails focus on letter-sound connections and decoding. This is a technique that is primarily implemented
in elementary schools. With the CCC technique, a student receives a word list. The student studies a word, covers the word, writes the word on paper from memory, and finally checks the spelling with the original word list (Harris et al., 2017; Skinner,
McLaughlin, & Logan, 1997). CCC is especially effective when it is combined with other instructional components, such as repetition or offering rewards (Joseph et al., 2012). Finally, it seems that interventions offering teaching material through playing games is also an effective method to improve spelling skills, and also a good way to keep students motivated (Harris et al., 2017).
Next to the content of an intervention, some other criteria can make a spelling intervention effective. Explicit spelling instruction is a key component of learning to spell (Harris et al., 2017). With explicit instruction, students learn about the patterns and rules underlying the words they are learning. This makes it more easy to generalize new knowledge to other words. Furthermore, practice and repetition are necessary
components to learn spelling. Research found that it is better to practice short periods of time on multiple days than to practice a lot on one day (Brooks, 2007; Graham, 1999). Additionally, highly structured interventions seem to be important for improving spelling ability (Brooks, 2007). This means that interventions need to have predictable lessons, with the same routines and layouts. This needs to be adjusted to the way students prefer to learn. Whether the intervention is computer-based or provided in-class, does not seem to influence the intervention effect (Brooks, 2007).
In conclusion, research on spelling interventions is still upcoming. However, we do know some of the content and criteria that contribute to the effectiveness of a spelling
intervention. The spelling intervention in this study also contains some of these effective components, such as explicit instruction and repetition.
Treatment Fidelity
To measure the effectiveness of an intervention reliably, it is important that treatment fidelity is in order. Treatment fidelity is the degree to which an intervention is employed as it was intended (Gresham, MacMillan, Beebe-Frankenberger, & Bocian, 2000; Yeaton & Sechrest, 1981). There is some preliminary evidence that treatment fidelity consists of different components. From all available literature, four recurring components were identified, namely content (content of the intervention), quality (how well was the content implemented), quantity (how much of the intervention was
provided), and process (how was the intervention implemented; Sanetti & Kratochwill, 2009). In this study we specifically look at the quantity, also referred to as the dosage of the intervention (Jones et al., in Sanetti & Kratochwill, 2009) or exposure (Dane & Schneider, 1998). Important for the dosage are – amongst others – the number of sessions provided, the length of each of these sessions, and the time between the sessions.
Research on spelling interventions was already limited and therefore even less literature is available on dosage in relation to spelling interventions. However, some research exists on the relation between dosage and spelling outcomes. Prior research found an effect for the length of an intervention. Longer interventions seemed to be more effective than short-term interventions, with interventions between 15 and 34 hours being the most effective (Galuschka et al., 2014). However, no effect was found when an intervention lasted more than a year compared to less than a year (Stockard, Wood, Coughlin, & Khoury, 2018). In addition, no effect was found for spending more than an
hour a day on the intervention (Stockard et al., 2018), which is in line with the finding that practicing short periods of time is better than a lot on one day (Graham, 1999).
All in all, little research has been conducted on dosage variables for spelling interventions. It seems that longer interventions of a few months are most effective and that practicing short periods of time works better than practicing a longer period of time on the same day. Yet, no information is available on the effect of dosage related to spelling interventions in high school or computer-based interventions. Other dosage variables, such as the number of sessions, spending time on instruction, and time between sessions have not been priorly researched in relation to spelling outcomes.
Present Study
In the present study we looked at the spelling intervention ‘Muiswerk’. Muiswerk is a Dutch, computer-based program for all educational levels and provides exercises for reading, spelling, and mathematics. Muiswerk has several didactic core assumptions, which align with effective intervention components that have previously been described. First, spelling categories are repeated and offered in different forms, so that students learn the new spelling category in different situations (Brooks, 2009). Second, the program provides explicit instruction prior to the exercises (Harris et al., 20017). During the exercises immediate feedback is provided. Finally, the intervention is computer-based, which has been found to be just as effective as regular in-class interventions (Brooks, 2007).
Muiswerk offers many remedial spelling exercises. In this study, however, we
specifically looked at verb inflection and spelling of visual imprint words for high school students (eighth grade) with and without dyslexia. Muiswerk was provided during Dutch
classes in addition to the regular spelling classes. As Muiswerk is computer-based it was possible to look at different forms of treatment fidelity in learning behavior, specifically dosage. The effectiveness of Muiswerk has been studied before. Contradictory results have been found. One study among high school students (all educational levels) found that spelling scores did not increase compared to the control group (Meijer, van Eck, & Felix, 2008), while another intervention study found that students following
pre-vocational education improved when working on verb inflection exercises (Nierop, 2017). However, more research is necessary to prove effectiveness for this specific group (general secondary education and pre-university education) and intervention content (verb inflection and visual imprint words).
The present study had three different aims. First, we wanted to investigate whether the Muiswerk spelling intervention for verb inflection and visual imprint words was effective for Dutch eighth grade students. We also looked at whether there was an interaction effect for dyslexia status on the effectiveness of the intervention. The second aim of this study was to identify which elements of dosage mediate the effectiveness of practicing both verb inflection and visual imprint words. Therefore, we took into account the amount of the intervention that was completed, the total repetitions, the amount of explanation, total time spent on exercises, the number of sessions, and the time between sessions. The third and final aim was to look at whether the same dosage variables contributed to effectiveness for students with dyslexia and TD students.
We expected that the intervention Muiswerk had a positive effect on the spelling scores on the posttest for both verb inflection and visual imprint words. We also expected that all seven dosage variables that were examined in this study mediated the effect
between the pre- and posttest. Prior literature showed that for school interventions it is important to complete the intervention (Dane & Schneider, 1998). Therefore, it is expected that the higher the number of sessions and the higher the percentage of completed exercises, the more effective Muiswerk was. Furthermore, the literature was not unambiguous about the best length of an intervention (Galuschka et al., 2014; Stockard et al., 2018). Additionally, no prior research described the best time between sessions. However, as Graham (1999) found that spelling interventions were more effective when practiced short periods of time on multiple days, compared to a long period of time on one day, we expected that it was more effective to keep the time between sessions short. Next to this, we expected that students benefited from more time listening to explanations and watching more explanation screens, as explicit instruction was found to be one of the key components of learning to spell (Harris et al., 2017). The seventh dosage variable was the number of repeated exercises. As prior literature has repeatedly found that repetition is important to acquire spelling skills (Graham, 1999; Harris et al., 2017), it was expected that students who repeated more exercises, would have higher spelling scores. Finally, we expected that there was a difference in the relation between the dosage variables and the posttest scores for students with dyslexia and TD students. As students with dyslexia students seem to make the same spelling mistakes as TD students, but make these mistakes for a longer period of time (Callens et al., 2012; Tops et al., 2014; Tops et al., 2013), it was expected that the relations between the dosage variables and the posttest were stronger for TD students than for students with dyslexia. Students with dyslexia most likely need to invest more time to gain the same effect as TD students.
Method Participants
A total of 153 eighth grade students participated in the study (59 boys, 94 girls). Nine classes from three Dutch high schools were recruited and consented to participating in the study. Students either attended general secondary education (35.5%) or
pre-university education (64.7%). The mean age of the students was 13.44 years (SD = 1.17 years; range 11.49 – 15.45 years). All students in all nine classes joined the intervention, which took place in the classroom. Active informed consent for processing and analyzing the intervention data for the current study was obtained from parents.
Twenty-five of the 153 participants were considered dyslexic (10 boys, 15 girls). Students were defined as dyslexic in this study when one of their parents reported that their child was in possession of an official dyslexia diagnosis (n = 19) or when parents had clear reasons to suspect dyslexia (e.g. child was going to be assessed in the near future; n = 6). Of the students with dyslexia 40% attended general secondary education, while 60% attended pre-university education. The mean age of the students with dyslexia was 13.0 years (SD = 2.62). Students with and without dyslexia did not differ
significantly in gender (t (151) = 0.160, p = .873), but TD students were significantly older than students with dyslexia (t(151) = 2.089, p = .038).
Measures
Intervention. Students completed the computer-based intervention Muiswerk. This intervention contained 27 assignments (7 sessions) for verb inflection and 22 assignments (5 sessions) for visual-imprint words. Verb inflection assignments consisted of the following topics: the finite verb (in past and present tense), the subject, and the past
participle (as adjective). The visual-imprint assignments consisted of two main topics. First, Dutch homophones were considered, namely au/ou, ei/ij, and g/ch. Each of these grapheme pairs are pronounced in the same manner. As the correct spelling cannot be determined based on the pronunciation of the word, these words need to be remembered. Second, French and English loan words were practiced. Within a session students either worked on verb inflection assignments or visual-imprint assignments. These were never mixed within one session.
Prior to the exercises, students could study three to five explanation screens per exercise. An explanation screen contained explanation about the spelling rule that was presented in the exercise. Within each of these explanation screens it was also possible to listen to explanation about the spelling rule. Consequently, to complete an exercise students needed to answer more than 75% of the questions correctly. If they did not achieve this percentage the first time they made the exercise, students had to repeat the assignment until they reached the minimum score. Furthermore, for the visual-imprint words one lesson required students to repeat earlier exercises. Therefore, the number of repeated exercises was higher in the visual-imprint measure. When students did not finish all assignments during the Dutch lesson they could finish the assignments at home.
Verb inflection. Students’ knowledge on verb inflection was tested with an online test provided by Muiswerk. Students answered 36 multiple-choice questions in which they had to choose the correctly spelled verb. Questions contained sentences in which a missing word had to be chosen from three or four answer options. Missing verbs could be in present (e.g., houdt (love)) or past tense (e.g., verdwaalde (got lost)) and past principles (e.g., geverfd (painted)). Furthermore, some questions asked students to identify whether
a word was correctly spelled by choosing between true/false. The percentage correctly answered questions was analyzed.
Visual-imprint words. Students’ knowledge of visual-imprint words was tested with a dictation task. Twenty-four words were dictated by the teacher, both separately and in a sentence (e.g., “Joris is allergisch voor garnaal. Schrijf op: garnaal” (“Joris is
allergic to shrimp. Write down: shrimp”)). The test contained several loan words (e.g., trottoir), and words with the similar sounding graphemes au/ou (e.g., gebouw (building),
kabeljauw (codfish)), ch/g (e.g., richel (ledge), agent (police officer)) and ei/ij (e.g., chagrijnig (grumpy), omheining (fence)). Students entered the dictated word in an online student response system (i.e. Socrative). The percentage correctly spelled words was registered.
Dosage. Several dosage factors were measured in the online Muiswerk
intervention program. Each of the following measures was reported separately for verb inflection and visual imprint words. First, the number of completed exercises was
measured as the percentage of the total exercises that was completed. Second, the number of repeated exercises was measured as the total number of times a student repeated an exercise during the intervention. Third, the number of explanation screens was a count of how many explanation screens a student on average looked at per exercise. The fourth measure was the time listening to explanation, which was coded as the average number of seconds a student listened to explicit instruction per exercise. Fifth, the average time spent on exercises was measured by the average number of minutes students spent on an exercise. Sixth, the number of sessions was measured by the number of days students
worked on exercises. Finally, the time between sessions was measured in the number of days between the sessions on average.
Procedure
A pretest-posttest-follow-up design was used in this study to evaluate the effect of the intervention. Prior to the start of the intervention, students completed a pretest. In this pretest student’s knowledge of visual-imprint words and verb inflection was tested. Additionally, the pretest contained a writing assignment, which was not included in the present study. After the pretest students started with the intervention Muiswerk. The intervention was implemented three days a week for six weeks during Dutch lessons. In each Dutch lesson 15 minutes were spent on the intervention. The first and last session were the pre- and posttest. In week 1 the third lesson of the week entailed an introduction to visual-imprint words. In week 3 and 5 the third lesson of the week entailed a writing assignment in which a few sentences were given and students were asked to finish the story. In the third lesson from week 4 students were asked to study and correct the story of a classmate. Data from these assignments were not taken into account in the present study. The additional two weekly sessions and the third session in week 2 contained
Muiswerk exercises.
Immediately after the six week intervention, students were tested again on their knowledge of verb inflection and visual-imprint words with the same experimental tests that were used in the pretest. Six weeks after the posttest a follow-up measurement was conducted using the same tests. The follow-up measurement was not used in this study. This study was approved by the ethics committee of the University of Amsterdam (project number 2019-CDE-10437).
Analysis
To know whether the Muiswerk intervention was effective in improving students’ spelling skills, a repeated measures ANOVA was conducted. Dyslexia status was added as a moderator to also look at the influence of having dyslexia on the improvement in spelling skills. The second goal of this study was to see to what extent the Muiswerk intervention was mediated by different dosage factors. Therefore, path analysis was used to analyze the data. Finally, we wanted to investigate whether these indirect effects were moderated by being dyslexic. Therefore, the path model was extended to a partial
moderated mediation model (see Figure 1 for the path model). Ten variables were included in this model. All variables were treated as continuous in this study, except for dyslexia status.
The analyses were carried out in RStudio (version 3.5.1; R Core Team, 2018) using the lavaan package (version 0.6-5; Rosseel, 2012). For both models (verb inflection and visual-imprint) correlated residuals were used as we were only interested in the direct effect of pretest on posttest and the indirect effects of pretest to posttest via the different dosage variables. Therefore, the models were just-identified and had perfect fit. Missing data was handled with Full Information Maximum Likelihood (FIML), in which all available data is used to estimate the model (Enders & Bandalos, 2001).
Figure 1. Hypothesized path model for both verb inflection and visual imprint words.
Correlated residuals are included, but not depicted in the figure for clarity reasons. Results
Data description
Prior to the analysis, data was screened. First, pre- and posttest data was deleted for students who did not complete at least one session of the intervention for verb inflection or visual-imprint words separately. Furthermore, a data point was considered an outlier when its standardized (z-)value was below -3.3 or above 3.3. All outliers were removed from the dataset. For verb inflection twelve outliers (one in pretest, eleven in dosage variables) were removed for five of the nine variables. These twelve outliers comprised 0.92% of the data. For visual imprint words five outliers (all in dosage variables) were removed for five out of the nine variables. These five outliers comprised 0.39% of the data. After removing the outliers, the data was checked for normality. For verb inflection three variables (number of exercises repeated, time listening to
explanation, and time between sessions) were not normally distributed, while one variable (time between sessions) was not normally distributed for visual-imprint words. As the data was not completely normally distributed, and Maximum Likelihood (ML) estimation is only robust to minor deviations in normality, robust ML was used in the analysis. Robust ML corrects the standard errors according to the Huber-White method (Huber, 1967).
In Table 1 descriptive statistics for the main variables are presented for verb inflection. Looking at the data, students on average completed two sessions less than was prescribed. This caused students to have less separate time points to practice the teaching materials. Furthermore, on average around 70% of the intervention was completed. However, the variance was large (25-30%). A lot of the students did thus not complete a large part of the intervention. Additionally, students repeated few exercises and the variance was small, indicating that while it is likely that not all students reached 75% on all exercises, they did not repeat these exercises to obtain a higher score. For both
spelling categories however, averages on time listening to explanation, total time spent on exercises, and time between sessions were as originally intended for the intervention. Results are displayed separately for students with and without dyslexia. As can be expected, students with dyslexia had a significantly lower score on the pre- and posttest than TD students. On most dosage variables students from both groups did not differ significantly, except for time spent on exercises. Students with dyslexia on average spent more time on an exercise than TD students.
Table 1
Descriptive Statistics of Verb Inflection Exercises for Students with Dyslexia and Typically Developing Students.
Students with dyslexia Typically developing students
N Mean (SD) Range N Mean (SD) Range t
Pretest 22 66.41 (11.22) 50.00-90.32 104 75.25 (12.47) 44-100 3.071** Posttest 21 64.66 (9.85) 47.22-83.33 85 76.64 (12.61) 38.89 – 97.22 4.055*** % exercises completed 25 70.86 (23.53) 21.43-96.43 128 70.28 (29.54) 3.57 – 96.43 0.106 Exercises repeated 25 0.92 (1.29) 0-5 125 1.02 (1.62) 0-8 0.279 Explanation screens viewed per exercise 25 3.77 (0.99) 2.07-5.50 127 3.47 (1.12) 2.00-7.25 1.259 Time listening to explanation per exercise (sec.) 23 24.80 (13.34) 6.46-65.41 127 22.86 (12.17) 5.50-71.00 0.694 Time spent on an exercise (min.) 25 4.64 (1.14) 2.48-7.37 127 4.13 (1.01) 2.34 – 7.72 2.270* Total number of sessions 25 5.60 (2.14) 2-9 128 5.27 (2.14) 1-10 0.714 Time between sessions (days) 25 5.88 (2.42) 2.67-14.00 121 6.56 (3.34) 0-22 0.980 Note. * < .05, ** < .01, *** <.001
Descriptive statistics for visual-imprint words are listed in Table 2. For visual imprint words, the same results stood out as for verb inflection. Students on average spent two sessions less than prescribed on the intervention. Furthermore, there was a large variance in the percentage of exercises that had been completed, which could influence the interpretability of the results of the effectiveness of the intervention. Moreover, students with dyslexia scored significantly lower than TD students on the pre- and posttest. The only dosage variable on which students with dyslexia differed from TD
students is time spent on exercises. Students with dyslexia on average spent more time on an exercise than TD students.
Table 2
Descriptive Statistics of Visual Imprint Word Exercises for Students with Dyslexia and Typically Developing Students.
Students with dyslexia Typically developing students
N Mean (SD) Range N Mean (SD) Range t
Pretest 20 47.55 (13.47) 21.00-71.00 92 77.21 (12.07) 46-100 9.752*** Posttest 19 63.90 (17.29) 37.50-95.83 89 86.15 (11.39) 42-100 5.366*** % exercises completed 24 69.01 (27.12) 12.50-100.00 111 80.80 (27.52) 6.25-100.00 1.908 Exercises repeated 24 6.13 (2.70) 1-12 110 6.95 (2.65) 0-14 1.370 Explanation screens viewed per exercise 23 3.45 (1.09) 2.00-5.50 111 3.26 (0.98) 2.00-5.75 0.819 Time listening to explanation per exercise (sec.) 24 15.43 (7.61) 4.00-33.50 110 15.81 (9.11) 5.33–44.50 0.189 Time spent on an exercise (min.) 24 3.25 (0.72) 1.88-4.37 111 2.77 (0.62) 1.61-4.81 3.369** Total number of sessions 24 3.13 (1.45) 1-7 111 3.57 (1.38) 1-6 1.412 Time between sessions (days) 20 8.20 (5.31) 3-28 101 8.12 (4.30) 0-27 0.074 Note. * < .05, ** < .01, *** <.001
A 2x2 repeated measures ANOVA was conducted for both word types to see whether the intervention had improved the spelling scores from pretest to posttest and whether there was a difference between students with and without dyslexia. No
difference in mean scores between pretest and posttest scores on verb inflection was found, F(1, 90) = 0.004, p = .951, indicating that the intervention was not effective. However, the change in spelling scores between pre- and posttest differed significantly for students with and without dyslexia, F(1, 90) = 4.596, p = .035. TD students appeared to experience some growth, while scores for students with dyslexia slightly decreased from pretest to posttest. For visual imprint words students improved their spelling scores from pretest to posttest, F(1, 88) = 51.334, p <.001. However, dyslexia status did not moderate the relation between pretest and posttest, F(1, 88) = 2.576, p = .112, meaning that TD and students with dyslexia did not differ significantly in their growth on spelling scores.
Correlations between all variables included in the model are displayed in Table 3 for both verb inflection and visual-imprint words. The correlation patterns mostly seemed to be similar for verb inflection and visual imprint words. The relation between the pretest and posttest was positive and medium strong for both word types. Furthermore, the pre- and posttest did not correlate with the dosage variables, except for a small to medium, negative correlation between the pretest and posttest and time spent on exercises, indicating that when students spent more time on exercises this went along with lower scores on the pre- and posttest. This was a surprising correlation. Furthermore, dyslexia status only correlated with the pre- and posttest and time spent on exercises. Finally, some of the dosage variables were highly correlated. For both verb inflection and visual imprint words a large, positive correlation was found between number of sessions and the percentage of exercises that was completed. Additionally, for visual imprint words a large, positive correlation was found number of sessions and exercises repeated.
As this correlation was extremely large, there was multicollinearity. Therefore, for visual imprint words the variable ‘exercises repeated’ was removed.
Table 3
Correlation Matrix for Predictors and Outcome Variables
Note. Visual-imprint words above diagonal, verb inflection below diagonal.
* < .05, ** < .01, *** <.001
Initially, we wanted to look at whether the same dosage variables contributed to the effectiveness of Muiswerk for students with and without dyslexia. However, as most of the conducted t-tests between students with and without dyslexia were not significant and most dosage variables did not correlate with the pre- and posttest, the results did not give cause for conducting moderator analyses. To be certain that there was no difference
1 2 3 4 5 6 7 8 9 10 1. Dyslexia status - .681*** .562*** .163 .118 -.071 .016 -.280** .122 -.007 2. Pretest .266** - .764*** .224* .123 -.108 -.063 -.365*** .159 .069 3. Posttest .369*** .570*** - .079 .025 -.163 -.069 -.344*** -.018 -.031 4. Percentage exercises completed -.007 .214* .491*** - .910*** -.086 .086 -.002 .810*** -.189* 5. Exercises repeated .023 .029 .128 .258** - -.156 .016 -.031 .785*** -.148 6. Explanation screens viewed -.102 -.138 -.185 -.151 -.123 - .250** .369*** -.056 -.127 7. Time listening to explanation -.057 -.036 -.112 -.107 -.139 .525*** - .437*** .082 -.043 8. Time spent on exercises -.182* -.155 -.183 .093 -.161 .000 .472*** - .024 -.009 9. Number of sessions -.058 .188* .351*** .830*** .274** .266** .003 .110 - -.249** 10. Time between sessions .081 .007 -.167 -.230** -.095 .069 .074 .036 -.295*** -
between students with and without dyslexia on the dosage variables, correlations between pre- and posttest and the dosage variables were compared for students with dyslexia and TD students. For the verb inflection exercises there were no large differences in
correlations between students with dyslexia and the TD group. Therefore, no indication was found that moderator analyses were necessary for verb inflection. However, for visual imprint exercises, differences were found between the correlations for students with and without dyslexia. This correlation matrix is reported in Appendix A. As
differences were found, moderator analyses were conducted for the visual-imprint words. Verb Inflection
Figure 2. Final path model used for verb inflection. Correlated residuals are included, but
The impact of the dosage variables for verb inflection was tested in a mediation model. Figure 2 shows the model that was used for verb inflection. This model differs from the model previously described, as there was no indication that moderator analyses were necessary. The pretest had a significant positive direct effect on the posttest, = 0.440, p < .001. This means that students kept their position relative to other classmates on their spelling scores. Students who obtained a high pretest score, in general also obtained a high posttest score. Furthermore, one of the indirect effects was significant. The posttest was significantly and positively predicted by the pretest via the percentage of exercises that has been completed, = 0.129, p = .015. This result is surprising, as the intervention was not effective. Dosage is not expected to significantly influence a non-effective intervention. None of the other indirect effects were significant (see Table 4). This model explained 50.7% of the variance in the posttest.
Table 4
Total, Total Indirect, Indirect, and Direct Effects on the Posttest (Verb Inflection)
Note. CI = confidence interval, = standardized coefficient.
*= <.05, ** = <.01, *** = <.001
Visual Imprint Words
Path p-value 95% CI
Pretest → posttest 0.440 <.001** 0.310; 0.611
Pretest → % completed → posttest 0.129 .015* 0.027; 0.244
Pretest → repeated → posttest -0.001 .848 -0.007; 0.006
Pretest → explanation screens → posttest 0.016 .324 -0.017; 0.051 Pretest → listening to explanation → posttest 0.001 .842 0.015; 0.001 Pretest → total time → posttest 0.014 .490 -0.027; 0.056
Pretest → sessions → posttest -0.042 .289 -0.127; 0.038
Pretest → time between sessions → posttest 0.001 .891 -0.010; 0.011
Total indirect 0.119 .008 0.032; 0.216
For visual imprint words a moderated mediation model was used (see Figure 1). Therefore, dyslexia status was added to the model as a moderator on all relations between the dosage variables and the posttest. The posttest was directly predicted by the pretest, = 0.426, p < .001. This effect was positive, which means that students kept their relative position to other students from pre- to posttest. None of the indirect effects between pretest and posttest via the dosage variables was significant. Looking at the moderated mediation, none of the relations between the dosage variables and the posttest were moderated by dyslexia status (see Table 6). This means that for the effectiveness of the intervention it did not matter how much of the intervention students completed, how much explicit instruction they received, how long they worked on the assignments and how many sessions and how much time was between sessions. Furthermore, no
difference was found for students with dyslexia and TD students. This model explained 61.3% of the variance in the posttest.
Table 6
Moderation Effects of Dyslexia (Dyslexia = 0, No Dyslexia = 1) on Paths Between Dosage Variables and the Posttest
Path p-value 95% CI
% completed → posttest -0.205 .491 -0.361; 0.173
Explanation screens → posttest -0.334 .166 -9.544; 1.639 Listening to explanation → posttest 0.276 .122 -0.129; 1.096
Total time → posttest 0.193 .528 -6.158; 12.014
Sessions → posttest 0.197 .447 -3.041; 6.895
Discussion
The present study looked at the Dutch spelling intervention Muiswerk, which contained assignments for verb inflection and visual imprint words and was provided to eighth grade students with and without dyslexia. The study had three different aims. The first aim was to investigate whether the spelling intervention was effective in improving spelling skills. The second aim was to identify to which extent dosage variables
influenced the spelling scores on the posttest. The final aim was to look at whether the same dosage elements contributed to effectiveness of the intervention for students with dyslexia and TD students. Results are discussed separately for verb inflection and visual imprint words.
The intervention Muiswerk did not appear to be effective for verb inflection, as the intervention did not lead to a significant increase in spelling scores. However, the change in spelling scores from pretest to posttest differed significantly between students with dyslexia and TD students. TD students seemed to slightly increase their spelling scores, while students with dyslexia slightly decreased their spelling scores. There was only one dosage variable that mediated the relation between pretest and posttest: the percentage of exercises that was completed in the intervention. This is an interesting finding, as the intervention was not found to be effective. It was expected that the dosage variables would not be significant mediators, as spending more time on an intervention that is already not effective is likely not useful. A possible explanation for the significant mediation effect of percentage completed exercises, is that the variation in this variable was large. As many students did not complete a significant part of the intervention, it cannot be expected that the intervention improves spelling skills significantly. Possibly
the mediating effect of percentage completed exercises indicates that when students would complete the full intervention, the intervention would become more effective.
While the intervention did not have a significant effect on verb inflection, it did seem like the intervention significantly improved students’ knowledge on visual imprint words. The effect of the intervention was not different for students with and without dyslexia. They improved in the same amount. This result should be interpreted with caution, however, as no control group was used in this study. Therefore, we cannot conclude that the progress in visual imprint words is exclusively due to the intervention. Surprisingly, none of the dosage variables significantly mediated the relation between pre- and posttest. Moreover, the influence of dosage variables did not differ between students with and without dyslexia in this study. It was expected that TD students would benefit more from spending more time on the exercises and from the explicit instruction than students with dyslexia, as previous research found that students with dyslexia need more repetition, time, and explanation in order to improve their spelling skills compared to TD students (e.g. Tressoldi, Vio, & Iozzino, 2007). In this study students with dyslexia did not repeat more exercises and did not look at the explanations more than TD students. Therefore, it is interesting that the influence of dosage variables did not differ between students with and without dyslexia in this study. These results show that the intervention most likely improves visual imprint skills in students with and without dyslexia, however it does not appear to matter how much time, repetition, and explanation is invested in the intervention. Furthermore, all results were equal for students with and without dyslexia, meaning that it improves skills in both groups, even if they spend the same amount of time on exercises and instruction.
A possible explanation for the difference in effectiveness between visual imprint words and verb inflection is that the intervention for visual imprint words was more structured than the intervention for verb inflection. For verb inflection many different topics were discussed (the finite verb, past principle, the subject, present tense, past tense and past principle as adjective were considered). Each lesson students were confronted with a new step-by-step plan to learn about new rules. Most of these topics were only discussed during one lesson and then mixed together with all other rules in the final lesson. For visual imprint words, however, a select amount of teaching material was offered (loan words and homophones) which was repeated every other lesson. It might be that the verb inflection intervention was too short and contained too little repetition for students to remember all information that was offered in the intervention (Galuschka et al., 2014). Repetition, and especially practicing short periods of time on multiple days with the same material has been found essential to improving spelling skills (Graham, 1999; Harris et al., 2017). This is more in line with the content of the intervention for visual imprint words.
Apart from one variable for verb inflection, none of the dosage variables were significant mediators between the pre- and posttest for verb inflection and visual imprint words. To interpret these findings, it is important to also take a closer look at the dosage variables. For both verb inflection and visual imprint words an interesting finding is that students on average completed two sessions less than was prescribed. Furthermore, for both word types, on average around 70% of the intervention was completed, with a large variance (25-30%). This might also have an influence on the intervention not having the expected effect, as a lot of students thus did not complete a large part of the intervention.
The available research shows that for school interventions completing the intervention is an important aspect of effectiveness (Dane & Schneider, 1998). For some variables (e.g., explanation screens viewed, time spent on an exercise, total number of sessions) the variance was quite small, which could also explain that no dosage effects were found. The differences between students could have been too small to find a meaningful effect. In addition, the intervention was relatively short. Missing a part of the intervention then quickly had a large influence.
Strengths of the current study are that Muiswerk is a computer-based intervention. Therefore, it was possible to get reliable information the dosage. Furthermore, the extent to which dosage influences spelling outcomes has hardly been studied before. Knowledge on this is necessary to improve current spelling interventions. However, there are some problems that might restrict the generalizability of the findings in this study. It would be interesting to replicate the study with more participants, especially the group of students with dyslexia was quite small in the current study. Therefore, results might not be representative for the population of students with dyslexia and might be less reliable. Furthermore, the current intervention was quite short. Therefore, it was not always possible to get a lot of variance in the dosage variables, which might be the reason why not been many significant effects were found. Future research could investigate the influence of dosage variables in a longer intervention. For future research it might also be interesting to extend the group of participants. Participants in this study followed either general secondary education or pre-university education, an educational level only 35% of the Dutch population reaches. Previous research found that students following pre-vocational education (the level 65% of Dutch population reaches) experience more
problems with verb inflection during and after high school (Van den Bergh, Van Es, & Spijker, 2011). Possibly the intervention has a different effect on this group compared to the current participants. Finally, there was an age difference between students with and without dyslexia in the current study. However, this does not have to pose a problem for interpreting the results as for spelling the number of years of education is more important than the age of the students. Most of the spelling students learn, is through explicit instruction during the spelling lessons at school and these are similar for all students (Kemper, Verhoeven, & Bosman, 2012).
This study is one of the few studies looking at the influence of dosage on the effectiveness of a spelling intervention. The current study has some implications for clinical practice and future research. The Dutch verb inflection intervention Muiswerk appears not to be effective for eighth grade students. However, it seems to be effective for visual imprint words. As it is a short intervention, it might be useful to implement it for visual imprint words in more high schools. Pretest scores showed that especially students with dyslexia seem to have difficulties with visual imprint words. This fits earlier findings in which it became clear that visual imprint words are mastered by less Dutch students than rule-based words in sixth grade (Keuning & Verhoeven, 2008). Therefore, especially students with dyslexia could benefit from the visual imprint intervention. TD students showed a much higher score on the visual imprint words pretest. However, for students who have received multiple years of spelling instruction, an even higher score was expected. Students with and without dyslexia had sufficient scores on the pretest for verb inflection. However, verb inflection is also a topic that has been extensively covered in the previous years of spelling instruction. It was thus
expected that students would have a higher pretest score on verb inflection as well. Therefore, it can be concluded that spelling interventions for students in high school are necessary and useful.
Dosage in relation to interventions is a difficult field to study. We do not know a lot about dosage and the role of dosage in effectiveness of interventions yet. Furthermore, it is sometimes hard to define expectations. Hypotheses can go both ways. For example, when students spent much time om an intervention, this can be because they take the assignment seriously and learn the skill that is taught in the exercise. However, when a student spends little time on an exercise, it could also mean that a student already knows the teaching material and thus obtained a high score on the assignment. Sometimes, both of these situations occur within the same intervention and then the dosage variable can show no effect, while actually there would be an effect if interactions with scores on assignments were taken into account. In future research, these interaction effects need to be taken into account. All in all, dosage is a relatively new research field that needs to be extended in relation to effectiveness of interventions.
References
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.
Assink, E. M. H. (1987). Algorithms in spelling instruction: The orthography of Dutch verbs. Journal of Educational Psychology, 79, 228–235. Retrieved from
https://www.apa.org/pubs/journals/edu/
Borgwaldt, S. R. (2003). From onset to entropy: spelling-pronunciation patterns in six
Netherlands). Retrieved from https://dare.uva.nl/search?identifier=9ae9c0e2-ba39-42fa-9d5d-ddc354fdad67
Brooks, G. (2007). What works for pupils with literacy difficulties? The effectiveness of
intervention schemes (Report No. 380). Retrieved from
https://www.icommunicatetherapy.com/wp-content/uploads/2012/09/Literacy-difficulties.pdf
Callens, M., Tops, W., & Brysbaert, M. (2012). Cognitive profile of students who enter higher education with an indication of dyslexia. PLoS ONE, 7: e38081.
doi:10.1371/journal.pone.0038081
Dane, A. V., & Schneider, B. H. (1998). Program integrity in primary and early secondary prevention: Are implementation effects out of control? Clinical
Psychology Review, 18, 23–45. Retrieved from
https://www.journals.elsevier.com/clinical-psychology-review
Desoete, A., Brysbaert, M., Tops, W., Callens, M., Lange, C. de, & Hees, V. van (2010).
Studeren met dyslexie. Gent, Belgium: University of Gent.
Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full information maximum likelihood estimation for missing data in structural equation models.
Structural Equation Modeling, 8, 430–457. doi:10.1207/S15328007SEM0803_5
Galuschka, K., Ise, E., Krick, K., & Schulte-Körne, G. (2014). Effectiveness of treatment approaches for children and adolescents with reading disabilities: A meta-analysis of randomized controlled trials. PLoS One, 9, e89900.
Gein, J. van de (2010). Komd een kind van de basisschool. Onderwijscommissie onderschat spelvaardigheden basisscholieren. Onze Taal, 79, 228–230. Graham, S. (1999). Handwriting and spelling instruction for students with learning
disabilities: A review. Learning Disability Quarterly, 22, 78–98. doi:10.2307/1511268
Gresham, F. M., MacMillan, D. L., Beebe-Frankenberger, M. E., & Bocian, K. M. (2000). Treatment integrity in learning disabilities intervention research: Do we really know how treatments are implemented? Learning Disabilities Research &
Practice, 15, 198–205. Retrieved from
https://onlinelibrary.wiley.com/journal/15405826
Harris, K. R., Graham, S., Aitken, A. A., Barkel, A., Houston, J., & Ray, A. (2017). Teaching, spelling, writing, and reading for writing. Powerful evidence-based practices. Teaching Exceptional Children, 59, 262–272.
doi:10.1177/0040059917697250
Huber, P. J. (1967). The behavior of maximum likelihood estimates under non-standard conditions. In: Proceedings of the Fifth Berkeley Symposium on Mathematical
Statistics and Probability (pp. 221–233). Berkeley, CA: University of California
Press.
Joseph, L. M., Konrad, M., Cates, G., Vajcner, T., Eveleigh, E., Fishley, K. M. (2012). A meta-analytic review of the cover-copy-compare and variations of this
self-management procedure. Psychology in the Schools, 49, 122–136. doi:10.1002/pits.20622
Kemper, M. J., Verhoeven, L., & Bosman, A. M. T. (2012). Implicit and explicit instruction of spelling rules. Learning and Individual Differences, 22, 639–649. doi:10.1016/j.lindif.2012.06.008
Keuning, J., & Verhoeven, L. (2008). Spelling development throughout the elementary grades: The Dutch case. Learning and Individual Differences, 18, 459–470. doi:10.1016/j.lindif.2007.12.001
Kuhlemeier, J. B., & Van den Bergh, H. (1989). Periodieke peiling van het
onderwijsniveau in het voortgezet onderwijs PPON – VO: Proefpeiling Nederlands 1988. Retrieved from https://doi.org/10.17026/dans-xmr-gmyb
Landerl, K., & Reitsma, P. (2005). Phonological and morphological consistency in the acquisition of vowel duration spelling in Dutch and German. Journal of
Experimental Child Psychology, 92, 322–344. doi:10.1016/j.jecp.2005.04.005
Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). Defining dyslexia, comorbidity, teachers’ knowledge of language and reading. Annals of Dyslexia, 53, 1–14. doi:10.1007/s11881-003-0001-9
Meijer, J., Van Eck, E., & Felix, C. (2008). Leren met meer effect; rapportage van het
onderzoek. Retrieved from
http://www.fisme.science.uu.nl/staff/christianb/downloads/EindrapportageLMME .pdf
Nierop, J. (2017). Omgaan met verschillen op het vmbo. Een onderzoek naar de invloed
van effectief en gedifferentieerd werkwoordspellingsonderwijs (Bachelor’s thesis,
Hogeschool Rotterdam, Rotterdam, The Netherlands). Retrieved from http://cdn.muiswerk.nl/omgaan_met_verschillen_op_het_vmbo.pdf
Peer, L. (2001). Dyslexia and its manifestations in the secondary school. In L. Peer & G. Reid (Eds.) Dyslexia – Succesful inclusion in the secondary school (pp. 1–10). London, UK: David Fulton Publishers.
Peterson, R. L., & Pennington, B. F. (2015). Developmental dyslexia. Annual Review of
Clinical Psychology, 11, 283-307. doi:10.1146/annurev-clinpsy-032814-112842
R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Retrieved from: https://www.R-project.og/
Rosseel, Y. (2012). Lavaan: An R package for structural equation modeling. Journal of
Statistical Software, 48, 1–36. doi:10.18637/jss.v048.i02
Sanetti, L. M. H., & Kratochwill, T. R. (2009). Toward developing a science of treatment integrity: Introduction to the special series. School Psychology Review, 38, 445– 459. Retrieved from https://naspjournals.org/loi/spsr
Shankweiler, D., Lundquist, E., Dreyer, L. G., & Dickinson, C. C. (1996). Reading and spelling difficulties in high school students: Causes and consequences. Reading
and Writing: An Interdisciplinary Journal, 8, 267–294. Retrieved from
https://rw.org.za/index.php/rw
Shaywitz, S. E., & Shaywitz, B. A. (2003). Dyslexia (specific reading disability).
Pediatrics in Review, 24, 147–152. doi:10.1542/pir.24-5-147
Skinner, C. H., McLaughlin, T. F., & Logan, P. (1997). Cover, copy, and compare: A self-managed academic intervention effective across skills, students, and settings.
Journal of Behavioral Education, 7, 295–306. Retrieved from
Stockard, J., Wood, T. W., Coughlin, C., & Khoury, C. R. (2018). The effectiveness of direct instruction curricula: A meta-analysis of a half century of research. Review
of Educational Research, 88, 479–507. doi:10.3102/0034654317751919
Swanson, H. L., & Hsieh, C. J. (2009). Reading disabilities in adults: A selective meta-analysis of the literature. Review of Educational Research, 79, 1362–1390. doi:10.3102/0034654309350931
Tabak, W. M., Schreuder, R., & Baayen, H. (2005). Lexical statistics and lexical processing: Semantic density, information complexity, sex, and irregularity in Dutch. In S. Kepser, & M. Reis (Eds.), Linguistic Evidence – Empirical,
Theoretical, and Computational Perspectives (pp. 529–555). Den Haag, The
Netherlands: De Gruyter Mouton.
Tops, W., Callens, C., Bijn, E., & Brysbaert, M. (2014). Spelling in adolescents with dyslexia: Errors and modes of assessment. Journal of Learning Disabilities, 47, 295–306. doi:10.1177/0022219412468159
Tops, W., Callens, C., Van Cauwenberghe, E., Adriaens, J., & Brysbaert, M. (2013). Beyond spelling: The writing skills of students with dyslexia in higher education.
Reading and Writing, 26, 705–720. doi:10.1007/s11145-012-9387-2
Tressoldi, P. E., Vio, C., & Iozzino, R. (2007). Efficacy of an intervention to improve fluency in children with developmental dyslexia in a regular orthography. Journal
of Learning Disabilities, 40, 203–209. doi:10.1177/00222194070400030201
Van den Bergh, H., van Es, A., & Spijker, S. (2011). Spelling op verschillende niveaus: Werkwoordspelling aan het einde van de basisschool en het einde van het
voortgezet onderwijs. Levende Talen Tijdschrift, 12, 3–14. Retrieved from http://www.lt-tijdschriften.nl/
Van der Sijs, N. (2009). Loanwords in Dutch. In M. Haspelmath & U. Tadmor (Eds.),
Loanwords in the World’s Languages: A Comparative Handbook (pp. 338–359).
Den Haag, The Netherlands: De Gruyter Mouton.
Wanzek, J., Vaughn, S., Wexler, J., Swanson, E. A., Edmonds, M., & Kim, A. (2006). A synthesis of spelling and eading interventions and their effects on the spelling outcomes of students with LD. Journal of Learning Disabilities, 39, 528–543. doi:10.1177/00222194060390060501
Yeaton, W. H., & Sechrest, L. (1981). Critical dimensions in the choice and maintenance of successful treatments: Strength, integrity, and effectiveness. Journal of
Consulting and Clinical Psychology, 49, 156–167. Retrieved from
Appendix A: Correlation Table Table A1
Correlation Matrix for Visual Imprint variables
Note. Students with dyslexia above diagonal, typically developing below diagonal.
* < .05, ** < .01, *** <.001 1 2 3 4 5 6 7 8 9 1. Pretest - .805*** -.398 -.478* -.025 -.472* -.478* -.486* .172 2. Posttest .497*** - -.284 -.392 -.067 -.328 -.487* -.412 -.156 3. Percentage exercises completed .329** .092 - .932*** .082 .305 .132 .812*** -.131 4. Exercises repeated .245* .098 .904*** - -.102 .116 .157 .835*** -.056 5. Explanation screens viewed -.089 -.165 -.111 -.161 - .494* .364 .054 -.358 6. Time listening to explanation .073 -.041 .045 -.005 .209* - .450* .121 -.231 7. Time spent on exercises -.104 -.155 .026 -.037 .363*** .467*** - .127 .052 8. Number of sessions .259* -.002 .806*** .770*** -.073 .074 .044 - -.164 9. Time between sessions .100 -.058 -.204* -.170 -.081 -.005 -.027 -.270** -
