The Effects of a Computer-Based Early Reading Program on the Literacy Skills of Kindergarten Students

Abstract

This research examines the effects of a computer reading program as a supplement to classroom instruction for Kindergarten students across Study 1 and Study 2 (N = 55 and 88, respectively). Students attended schools in suburban settings and all students received their typical balanced literacy programming. Students in the treatment condition used the Ooka Island computer program (Scholastic F.I.R.S.T.^TM, 2017) as a supplement to classroom instruction. The computer program targets phonological awareness, word reading, and reading comprehension. Across both studies, students in the treatment group (n = 26 and 49 for Study 1 and 2, respectively) had better outcomes for phonological blending, word reading, reading comprehension, and book-reading levels than the business-as-usual control group (n = 29 and 39). There were no group differences for phonological segmentation. The findings demonstrate that early-reading computer programs, such as Ooka Island, have the potential to supplement classroom instruction toward building foundational reading skills.

Keywords:

• Computer-based instruction
• reading instruction
• early literacy
• Kindergarten students

Too many elementary school children across North America are not learning to read well (National Assessment of Educational Progress, 2019). Across several Atlantic Canadian Provinces, the context for this study, approximately one-third of students in grades 2 and 3 are not meeting grade-level expectations in reading (Nova Scotia Department of Education & Early Childhood Development, 2019; Prince Edward Island Department of Education, 2018). At the same time, research has shown that targeting foundational reading skills can prevent many elementary-school reading difficulties (Foorman et al., 2016; Savage et al., 2018). Computer-based early reading programs have the potential to provide an engaging learning environment to target decoding-related skills, can provide scaffolded practice with text reading, and can include activities to potentially enhance comprehension. This study examined the effects of Ooka Island (Scholastic F.I.R.S.T.^TM, 2017) on Kindergarten children’s reading skills.

Reading instruction for young children

Reading comprehension is the primary goal of elementary-school reading programs (Foorman et al., 2016). Skilled readers have accurate and quick word reading skills that allow attention to be directed to aspects of comprehension, such as using prior knowledge to understand and interpret words, sentences and longer segments of text (Stanovich, 2000). Developing skilled word reading is thus one important aspect of a complete early reading program (Al Otaiba & Foorman, 2008). Critical to building word reading skills, are children’s abilities to manipulate individual sounds in spoken words (i.e., phonemic awareness) and to use letter-sound correspondences to decode words (Ehri, 2005; Foorman et al., 2016). Kindergarten children who struggle with these skills have difficulties in learning to read (Lonigan & Shanahan, 2009; Moats, 2020).

Instruction in grapheme-phoneme correspondences and how to use these to decode written words (i.e., phonics instruction) is also critical to developing word reading skills (National Reading Panel, 2000). Teaching and practice of sounding out words supports children to become accurate and quick at reading words (Brady, 2020; National Reading Panel, 2000). Instruction targeting phonemic awareness, grapheme-phoneme knowledge and decoding strengthens reading comprehension and helps prevent reading disabilities (Al Otaiba & Torgesen, 2007).

There is an ongoing debate concerning how young children are best taught to read (Moats, 2020; Spear-Swerling, 2019). Balanced literacy approaches emphasize learning from classroom reading experiences, using multiple cues to solve for unknown words in text (meaning, grammar, and letters), and progressing through a series of leveled readers (e.g., Fountas & Pinnell, 2009). A structured literacy approach includes one focus on directly and systematically teaching decoding-skills (phonemic awareness, grapheme-phoneme correspondences, and left-to-right sounding out and spelling of written words; Spear-Swerling, 2019). Knowledge of how to teach foundational word reading skills with structured literacy is often not emphasized in teacher education programs (Ontario Human Rights Commission, 2022) and has been found to be an area of relative weakness for North American teachers (Joshi et al., 2009). Balanced literacy approaches characterize instruction the Canadian Atlantic provinces, the context of the current study. One critical question, therefore, is whether a supplemental computer program, like Ooka Island, can deliver the structured approach recommended in research to facilitate young children’s early skill development within classrooms using a balanced literacy approach.

Complete reading programs in the early elementary years also need to include ample opportunities to read connected text and engage with stories (Duke & Pearson, 2009; Foorman et al., 2016). Familiarity with stories and understanding story structures are important for reading comprehension as these help children follow and recall the sequence of events (Duke & Pearson, 2009). Moreover, adequate time for engaging with interesting texts is related to young students’ reading achievement and to their intrinsic motivations for reading (Gambrell, 2011). For these reasons, most early reading approaches incorporate an abundance of storybook reading and a focus on building vocabulary in early literacy classrooms (Lonigan & Shanahan, 2009).

Across North America, children’s reading progression is frequently assessed by informal reading inventories that determine the book reading-level for each child within a published system (Fountas & Pinnell, 2007; Parker et al., 2015). Teachers listen to a child read and record errors, fluency, and responses to orally posed comprehension questions. These assessments are often used to gauge individual progress and screen for those who require interventions, as well as for district level monitoring (National Reading Panel, 2000; Prince Edward Island Department of Education, 2018). Although one of the most frequently used assessment methods in North American classrooms (Parker et al., 2015), these measures have been largely excluded in research studies examining reading outcomes to classroom interventions. The current study addresses this limitation by examining teachers’ assessed levels as one outcome measure.

Computer-assisted early reading instruction

Computer technology has been explored as one possible delivery system to supplement classroom reading instruction in the early elementary years (Jamshidifarsani et al., 2019), a pressing need given the high number of students failing to meet expectations across North America. Chambers et al. (2008) noted that computer-assisted technology might be particularly beneficial for decoding-related skills, as these programs are often engaging for young students, target areas for which the student needs more practice, and advance at a personalized pace. Computer programs can thus accommodate diverse learning needs of students, at the same time as providing immediate feedback and rewards.

Studies examining the effectiveness of computer-assisted instruction have largely focused on lower-level foundational skills, such as phonemic awareness, letter-sound knowledge, and word reading (Jamshidifarsani et al., 2019). Improvements in preschoolers’ and Kindergarteners’ phonological awareness was observed with computer intervention in one study (Foster et al., 1994), and Kindergarten students’ phonological awareness and grapheme-phoneme correspondence knowledge showed improvements compared to unseen control groups (Macaruso & Rodman, 2011; Macaruso & Walker, 2008). For second grade students, phonological awareness and word reading skills improved with programs targeting grapheme-phoneme correspondences or correspondences at the rime-level (Kyle et al., 2013; Saine et al., 2011). Similarly, positive findings have been reported in randomized control trials for the ABRACADABRA program on aspects of Kindergarten and first-grade students’ phonological awareness and word-decoding related skills (Piquette et al., 2014; Savage et al., 2009). ABRACADABRA is an interactive literacy software program that targets phonics, phonemic awareness and word reading, as well as fluency and comprehension in young students. Researchers have largely attributed the positive effects of these computer reading programs to the content and sequence of presentation aligning with non-computerized evidence-based programs (Foster et al., 1994; Kyle et al., 2013; Macaruso & Rodman, 2011; Macaruso & Walker, 2008; Saine et al., 2011) and to high usability and engaging game-like activities (Kyle et al., 2013; Saine et al., 2011).

Fewer studies have examined the benefits of computer-assisted instruction on reading comprehension during the early elementary grades (Schechter et al., 2015). Nonetheless, first graders who received facilitator-led instruction with the ABRACADABRA program had higher reading comprehension than a business-as-usual control group (Savage et al., 2009). In contrast, other studies and major reviews have concluded that computer-assisted early reading programs are not beneficial to overall reading achievement (Paterson et al., 2003). A large-scale randomized controlled trial involving grade 1 classrooms from 43 schools, found a nonsignificant mean effect size across five computer reading-programs (Dynarski et al., 2007). Similarly, a meta-analysis and two research syntheses reported small effect sizes associated with computer-based reading programs (Blok et al., 2002; Cheung & Slavin, 2012; Cheung & Slavin, 2013). Such mixed results concerning the effects of computer-based programs on a variety of reading outcomes have led to questions of their potential for producing educationally meaningful effects.

One difficulty has been that reviews typically examined the effects of programs on the overall reading achievement of students, whereas most programs specifically target phonological awareness and word reading (Wood et al., 2017). More comprehensive programs targeting decoding-related skills and reading comprehension have not been extensively studied (Schechter et al., 2015). A recent study revealed that targeting phonological awareness, word reading, and comprehension with the addition of a technology component improved Kindergarten reading achievement compared to a business-as-usual comparison group (Wilkes et al., 2020) and there were similar findings for grade 1 and 2 students (Schechter et al., 2015). Our study also addresses this gap in the literature by examining the effects of a comprehensive computer reading-program on Kindergarten students’ early word reading and reading comprehension outcomes.

The current study

The current study set out to examine the effects of the Ooka Island program on young students’ reading skills. This program was developed by Dr. Kay MacPhee, who also developed SpellRead™ (2012), an evidence-based reading intervention program (Metsala & David, 2017). Wood et al. (2017) rated Ooka Island favorably as the instruction covered 8 of their 9 identified broad reading skills (i.e., print concepts, alphabet knowledge, phonological awareness, letter-sound knowledge, phonics, decoding, reading fluency and comprehension; but not syntactic awareness); however, there is no empirical research examining the effects of the program. The game-like and reading activities that take place on the magical Ooka Island target reading skills aligned with recommendations of the National Reading Panel (2000) to target phonological awareness, phonics and word reading, vocabulary and reading comprehension (Scholastic Canada, 2016).

This study is a field study with teachers implementing the Ooka Island program in their own classrooms, and the research questions and to some extent the potential study design, originated with education administrators. The first author, a university faculty member, was contacted by the team from the school district about conducting this research. The team included district administrators and representatives of the Ooka Island program. The authors have no affiliation with Scholastic and report there are no competing interests to declare.

Research questions

1. Compared to a business-as-usual control group, do students who used the Ooka Island program as a supplement to their regular, balanced-literacy classroom instruction have higher outcome scores on measures of word-reading and related skills; that is, on phonological blending and segmenting, pseudoword and word reading accuracy, and single word reading efficiency?

2. Compared to a business-as-usual control group, do students who used the Ooka Island program as a supplement to their regular, balanced-literacy classroom instruction have higher outcome scores on a measure of reading comprehension?

3. Compared to a business-as-usual control group, do students who used the Ooka Island program as a supplement to their regular, balanced-literacy classroom instruction have higher outcomes on teacher-assessed measure of book-reading level?

For both studies, we hypothesized affirmative findings to each of our research questions; that is, students who use the Ooka Island program would have better outcomes on each measure than the business-as-usual comparison group. Study 1 examined a cohort in Year 1 and Study 2 was with a separate cohort in Year 2 (see Figure 1 for a timeline, the number of classrooms each year and the number of participants in each group).

Study 1

Method

Participants

Students from four kindergarten classes within the same school in an Atlantic Canadian province participated in Study 1. All students in two of the classrooms, based on teacher self-selection, used the Ooka Island reading program from February to the end of May of the academic year. Students who returned consent forms signed by a guardian participated in this study. There were 26 participants (13 males; 13 females) in the Ooka Island group and 29 participants (15 males; 14 females) in the control group. Three students in each group had a first language other than English, as reported by their guardians. The groups did not differ on their mean age at pretest, t(53) = .88, p > .10 (M = 67.12 mos [SD = 3.14] and 68.03 mos [SD = 4.42] for the Ooka Island and comparison group, respectively). The participants were all from the same school, located in a suburban setting, ensuring comparability in socio-economic backgrounds. In general, the school served children from working- and middle-class families. The students attended full-day Kindergarten programs for 5-days a week in their local K-6 public school and the province has an overall inclusive educational philosophy and policies.

The Ooka Island computer reading program

One focus of the Ooka Island program are the game-like activities that teach the decoding- related skills of phonological awareness, letter-sound correspondences, word decoding and irregular word reading. A phonological blending activity, for example, requires students to listen to a series of phonemes and then to choose the correct response from one of several pictured items. This requires the students to blend the sounds together in order to choose the correct picture. This blending activity is situated within the context of a moving company that has accidently spilled the household contents into the river, and the child needs to choose the correct picture before the items go over a waterfall. Similar game-like settings, with responses given by a selection between several options, are used to identify rhyming syllables and words, to segment spoken-words into sounds, and to listen and identify vowel sounds in initial, medial or final position. Children are encouraged by the characters to vocalize aloud the target sound unit when appropriate (e.g., word, syllable, phoneme), although there is no voice monitoring system used in the program. Learning letter-sound correspondences includes activities requiring students to match spoken phonemes with their corresponding grapheme. This is set within the “Bubbly Trubbly” game, in which sounds have been stolen from the cake factory and dumped in a fountain. The graphemes representing the sounds come up in bubbles and children need to choose the correct grapheme for the spoken phoneme, so they can be taken back to the cake factory. Beginning levels are heavily scaffolded (e.g., first only one bubble with the correct grapheme appears, then two bubbles with one other grapheme that is also being learned). As they progress through the levels, students move from simple grapheme-phoneme correspondences (e.g., s, n, l, short _u_) to consonant diagraphs (e.g., ch, ch, th) and final e vowels (e.g., i_e, a_e,), all the way to complex vowel patterns (e.g., ow, ar). In similar game-like activities with differing premises, the word-reading activities have students identifying high frequency words and applying grapheme-phoneme correspondence knowledge to sound out syllables and words.

The second focus of the Ooka Island program addresses word reading, vocabulary, story comprehension, and text-reading fluency through the 85 electronic storybooks and associated learning activities. Children listen to books in sequence in the “Popcorn Library”. The books increase in complexity along several dimensions (e.g., word decoding, length, vocabulary, syntactic structures, and the concepts that are introduced). The books are read to the children, with each word turning from blue to black as it is read to help direct the child’s attention to the print. The initial reading of each book is heavily scaffolded by ‘Auntie Kay’ (a character present throughout most activities), much as teachers might do in class. Pre-reading observations from the cover picture and title are made, and from this information questions posed concerning what might happen in the story. Throughout the book, brief observations or summary statements are made after reading each page, and further questions posed concerning what might happen next. After a book is read the first time, children engage in activities which require them to: i) answer spoken questions focused on the who, what, and where of the story by choosing among about 3 responses (for earliest books, the answers are represented by pictures and in later books, are in words and phrases/sentences) and ii) choose the correct written form of a spoken word that was presented in the story. The book is read again to the child, this time without any commentary. Children then complete activities that focus on word recognition or story sequencing; that is, the child sequences pictures in the earliest books and later sentences as they progress, consistent with the order of occurrence in the story. Further activities have students choose a word that represents a synonym (and later antonym) for a word from the story; the correct answer was also introduced in the story (e.g., a word that means the same in a story about children making snowmen is building, which was also introduced in the story).

In this way, the story reading is designed to boost many aspects of students’ comprehension (e.g., familiarity with story sequencing, printed words, vocabulary, and concepts). The amount of time that each child has spent engaged with the program is automatically recorded by the program, and included in teacher reports, alongside additional information about students’ progress. Further information is given by Scholastic and one can register for a demonstration of the reading program (Scholastic F.I.R.S.T.^TM, 2017).

Procedure

Classroom instruction

The typical English Language Arts curriculum continued in all of the classrooms, thus fully comprising the literacy programming in the business-as-usual comparison group. This instruction is best characterized as balanced literacy (Spear-Swerling, 2019). The curriculum emphasizes using contextual cues (e.g., pictures and story context, grammar, letters) to problem solve unknown words and children advance through a series of leveled readers (Fountas & Pinnell, 2007), participating in both whole class instruction, small-group guided reading, and independent reading. Consistent with balanced literacy approaches, the curriculum included phonics instruction as more ad-hoc than systematic, and patterns of letters in words are discovered or brought to the students’ attention in their reading and writing. The teachers did not report using a supplemental synthetic phonics program in the classrooms.

Pretest and post-test phonemic awareness

Two subtests of the Comprehensive Test of Phonological Processing (CTOPP; Wagner et al., 1999) were administered at pretest and again at post-test. First, for the Blending subtest the child hears a series of phonemes and is required to blend these into a real word. Testing is discontinued after 3 consecutive errors and test-retest reliability is reported to be .88.

The second phonemic awareness test was the Elision subtest, which measured participants’ phonological segmentation abilities. Students are asked to repeat back a spoken word without a given segment (e.g., Say subway without the/way/; Say/sat/without the/s/). Testing is discontinued after 3 consecutive errors and test-retest reliability is .88 (Wagner et al., 1999).

Pretest word reading

Students completed the Word Identification subtest of the Woodcock Reading Mastery Tests - Revised (WRMT-R; Woodcock, 1987). Students read aloud a list of words that increase in difficulty and testing is discontinued after 6 consecutive errors. The manual reports the Spearman-Brown split-half reliability coefficient to be .98.

Post-test word reading

At post-test, students completed the Word Identification subtest of the WRMT-III (Woodcock, 2011). Students read increasingly more difficult words and the test is discontinued after 4 consecutive errors. The manual reports reliability to be .95.

Post-test pseudoword reading

The Word Attack subtest of the WRMT-III was administered at post-test. Students read a list of increasingly more complex pseudowords and the test is discontinued after 4 consecutive errors. The manual reports the reliability of this subtest to be .89 (Woodcock, 2011).

Pretest-oral language comprehension

Oral language comprehension at pretest was assessed by orally administrating the Passage Comprehension subtest of the WRMT-R (Woodcock, 1987). The examiner read aloud each sentence or short passage in turn, and the student was asked to supply the missing word. The test was discontinued after 6 consecutive errors. Cronbach’s alpha for this non-standardized use of this measure was calculated to be .92 for our sample.

Post-test reading comprehension

The Passage Comprehension subtest of the WRMT-III was administered. For this test, children were required to read a sentence or short passage silently and then supply the missing word in the passage. The test is discontinued after 4 consecutive errors. Alternate form reliability for the passage comprehension scale is reported to be .74 (Woodcock, 2011).

Leveled-book reading assessment

At post-test, as part of their classroom practices, teachers completed an assessment to determine each child’s instructional reading level. The teachers followed the procedures of the Benchmark Assessment System (Fountas & Pinnell, 2007). For students in Kindergarten, the instructional level is the highest book-level at which the student read the words with at least 90 − 94% accuracy and showed good comprehension in response to the questions following the book. At the end of the year, Kindergarten students who are at level C or above are judged to be meeting grade-level expectations (i.e., the third level). For the purposes of analyses, lettered levels submitted by the teachers were transformed into numbered levels (A = 1, B = 2 …E = 5, etc.). Alternate forms reliability is reported by the publisher to be .94 (Fountas & Pinnell, 2007).

Results

Participants in the Ooka Island condition used the program an average of 41.7 hours (SD = 7.77; range = 30.5 − 59.3 hours), with no difference between the two Kindergarten classes using the program (M = 39.9 and 43.6 hours, t(24) = 1.21, p = .24). The entire Ooka Program takes about 80 hours to complete.

Raw score distributions for all tasks were examined for normality. Log transformations were applied to correct for mild positive skewness for the measures of Word Reading, Pseudoword Reading, and Reading Comprehension (Tabachnick & Fidell, 2007). Analysis was performed on these transformed scores and raw scores for untransformed variables. All hypotheses were unidirectional and thus p-values are reported for one-tailed, Independent Samples t-tests. There were 7 data points missing, and these are reflected in the reported degrees of freedom. The data were analyzed using IBM SPSS Statistics (27; IBM Corp., 2020).

Phonological awareness and word-level reading

There were no group differences on the pretest measures of Blending, Elision, and Word Identification, t(53) = 1.25, 1.10, 1.26, p > .10, respectively (see Table 1 for group means and standard deviations for each measure in Study 1). Therefore, post-test measures of decoding-related skills were directly compared. On post-test phonological measures, there was a significant difference between the groups on Blending, t(53) = 2.93, p < .05, but no group difference on Elision t(52) = 1.2, p > .10. Students who used the Ooka Island program performed better at blending phonemes into words, and this was associated with a large effect size (Cohen’s d = .79).

Table 1. Study 1: Raw score means and standard deviations by group.

	Ooka Island		Comparison
Pretest Measure	M	SD	M	SD
Blending	6.50	3.54	5.67	2.58
Elision	4.35	2.74	3.59	4.35
Word ID	7.56	11.55	6.59	10.21
Oral Lang.	16.85	8.98	15.34	7.54
Post-test Measure	M	SD	M	SD
Blending*	9.73	3.24	7.17	3.22
Elision	5.88	2.76	5.00	2.45
Word Reading*	7.92	6.70	5.38	5.16
Word Attack	2.35	4.16	1.72	3.8
Passage Comp*	4.20	3.91	2.83	3.74
Reading Level*	4.50	1.30	3.80	1.16

Note. An *indicates that group comparison was significant at p < .05; Blending = CTOPP Blending Subtest, Elision = CTOPP Elision Subtest; Passage Comp. = Passage Comprehension Subtest.

At post-test, there was a significant difference between the groups for Word Identification, t(50) = 1.75, p < .05. The Ooka group outperformed the control group, and this difference was associated with a medium effect size (Cohen’s d = .48). There was no group difference on the Word Attack measure, t(53) = .91, p > .10 (see Table 1).

Text reading and reading comprehension

There was no difference between the groups on pretest word reading (noted above) or the oral language comprehension measure t(53) = .67, p > .10. There was a statistically significant differences between the groups on post-test Passage Comprehension, t(53) = 2.04, p < .05. The Ooka Island participants scored higher on this test than the control participants, a difference associated with a medium effect size (Cohen’s d = .55). There was also a significant difference on mean book reading-level between the groups, t(50) = 1.99, p < .05. Students in the Ooka program had higher reading levels than those in the control group and this was associated with a medium effect size (Cohen’s d = .57; see Table 1).

Study 2

Method

Participants

In the following year, three schools in this same school district, including the school from Year 1, participated in this study. For each class in a school that used the Ooka Island program to supplement the balanced literacy program, there was another class in the same school that participated in the study and did not use the Ooka Island program. In total, there were 4 classes that used the Ooka Island program and 4 that did not. Altogether, 49 students in the treatment condition (24 males; 25 females) and 39 (20 males; 19 females) in the nontreatment condition returned consent forms and participated in this study. The groups did not differ on their mean age at pretest, t(86) = 1.412, p = .161 (M = 63.51 mos [SD = 3.74] and 64.62 mos [SD = 3.52] for the Ooka Island and comparison group, respectively). In terms of students who learned English as an additional language, there were 14 in the treatment condition and 10 in the control condition. The students used the program from mid-November to the end of May, with pre- and post-testing completed in the two weeks before and after the use of the program, respectively.

Procedure

In order to reduce initial testing time for individual students, only two measures were collected at pretest. All measures in Study 2 were also used and described in Study 1, with the addition of a post-test measure of word reading efficiency (see below).

Pretest and post-test phonemic awareness

At pretest and post-test, the CTOPP Blending subtest (Wagner et al., 1999) was administered. At post-test, the CTOPP Elision test was also administered (Wagner et al., 1999).

Pretest word reading

Students completed the Word Identification subtest of the WRMT-R (Woodcock, 1987).

Post-test word and pseudoword reading

Students completed the Word Identification and Word Attack subtests of the WRMT-III (Woodcock, 2011).

Post-test word reading efficiency

The word reading subtest of the Test of Word Reading Efficiency (Torgesen et al., 1999) was added in Study 2 to have a measure of speeded word reading. This subtest measures the number of words the student reads from a list of words in 45 seconds. The manual reports test-retest reliability greater than .90.

Post-test reading comprehension

The Passage Comprehension subtest of the WRMT-III (Woodcock, 2011) was administered to measure reading comprehension.

Leveled-book reading assessment

Teachers’ assessment of students’ book reading levels were collected at post-test (Benchmark Assessment System; Fountas & Pinnell, 2007).

Results

Participants in the Ooka Island condition used the program an average of 44.4 hours (SD = 11.5 hrs; range = 35 − 70 hours). Raw score distributions for all tasks were examined for normality and log transformations were applied to correct for mild positive skewness for the measures of Word Reading and Pseudoword Reading (Tabachnick & Fidell, 2007). As in Study 1, other than these transformed scores, all analyses were completed on participants’ raw scores on each measure. There were 2 data points missing for the word reading efficiency test, and these missing scores are reflected in the degrees of freedom.

Univariate analyses of covariance (ANCOVAs) were conducted for each post-test measure with the pretest covariate of the same or most closely related skill. All hypotheses remained unidirectional as for Study 1. Means and standard deviations for Study 2 measures are presented in Table 2.

Table 2. Study 2: Raw score means and standard deviations by group.

	Ooka Island		Comparison
Pretest Measure	M	SD	M	SD
Blending	2.33	1.96	3.00	2.43
Word ID	.47	1.93	.44	1.33
Post-test Measure	M	SD	M	SD
Blending*	7.65	3.36	6.54	3.14
Elision	5.43	3.76	5.46	3.47
Word Reading*	7.65	3.36	6.54	3.14
Word Attack	2.12	3.11	1.56	2.53
SWRE*	16.52	13.24	12.08	8.65
Passage Comp*	6.24	4.04	4.21	2.68
Reading Level*	4.32	2.30	3.50	1.67

Note. An *indicates that post-test group comparison was significant at p < .05; Blending = CTOPP Blending Subtest, Elision = CTOPP Elision Subtest; SWRE = Single Word Reading Efficiency Subtest; Passage Comp. = Passage Comprehension Subtest.

Phonological awareness and word-level reading

An ANCOVA was conducted on post-test Blending scores with pretest Blending scores as the covariate. Results showed that participants in the Ooka condition performed better than those in the control condition, F(1, 85) = 5.93, p = .009, reflecting a medium effect size (ηp² = .065). The same ANCOVA with post-test Elision as the dependent variable and pretest Blending as the covariate did not show a difference between the groups (p = .165).

An ANCOVA conducted on post-test word reading, with pretest word reading as a covariate, showed that the Ooka group outperformed the control group, F(1, 85) = 4.93, p = .015. This was associated with a medium effect size (ηp² = .057). For a separate ANCOVA with post-test word attack as the outcome and pretest word reading as the covariate, the Ooka group performed better than the control group, F(1, 85) = 3.18, p = .039 and this was associated with a small effect size (ηp²= .025). Finally, the ANCOVA on post-test word reading efficiency, with pretest word reading as the covariate, showed that the Ooka group outperformed the control condition, F(1, 83) = 6.60, p = .006), and this was associated with a medium effect size (ηp² = .056).

Text reading and reading comprehension

An ANCOVA for the post-test passage comprehension measure, with pretest word reading as the covariate, showed that the Ooka group performed better than the control group, F(1, 85) = 10.10, p = .001). This difference was associated with a medium effect size (ηp² = .094).

An ANCOVA on the post-test book reading-levels with pretest word reading as a covariate showed an effect of group, F(1, 85) = 4.07, p = .024. Participants in the Ooka condition had a higher mean book reading level than those in the control condition, and this was associated with a small effect size (ηp² = .036).

Reading outcomes and English learner status

There was a larger group of students who were learning English as an additional language (EAL) in this second year of the study. To investigate whether the supplementary program was differentially effective for students who were learning English as an additional language vs. those who learned English as their first language, we examined whether interaction effects were significant for each of the measures that showed an effect of the program above. The ANCOVAs had two between participants variables with 2 levels each – Condition (Ooka, control) x English Language status (EAL, English first language). None of the interactions were significant (p’s = .151 − .485), indicating that differential effects by EAL status were not observed in this study.

Discussion

Educational administrators are tasked with making decisions concerning how to raise young students’ foundational reading skills. The current study addressed this applied question; Is the use of the Ooka Island program as a supplement to classroom balanced literacy instruction effective at increasing Kindergarten students’ reading and reading related skills? Previous findings on the effectiveness of computer-assisted instruction have been inconsistent, perhaps somewhat due to a mismatch between the skills targeted and the outcomes measured (Archer et al., 2014). Across two studies, we found that there were benefits for the students who used the Ooka Island program on important decoding-related skills and on students’ text comprehension and book reading-level.

Phonological awareness is critical to developing word reading (Foorman et al., 2016). Students in the treatment condition outperformed their peers on a post-test measure of blending phonemes into spoken words, and this was associated with a large and a medium effect size in Study 1 and Study 2, respectively. Games in the program engage children directly in blending individual sounds to identify a word. These blending skills are critical to sounding out words, which in turn leads to building up large numbers of sight words (i.e., word recognized automatically). This sounding out process serves a self-teaching function in building skilled word reading (Share, 1995).

Skilled word reading is necessary for optimal reading comprehension and deficits are a defining feature of reading disabilities or dyslexia (Foorman et al., 2016; Vellutino et al., 2004). Across both cohorts, students in the Ooka Island condition outperformed control students on a standardized test of word reading in the spring of Kindergarten and differences were associated with medium effect sizes. Additionally, the treatment group performed better on word reading efficiency and this was associated with a medium effect size. Thus, there were clear advantages for the development of word reading accuracy and automaticity for students who used the Ooka Island program. Strong early word reading skills are important toward preventing future reading disabilities (Foorman et al., 2016) and the pace of acquiring early word reading skills appears critical to life-long reading practices (Cunningham & Stanovich, 1998). This is an area of skill development most heavily criticized as absent in balanced literacy programs (Moats, 2020); therefore, that a supplementary computer reading program can potentially boost early word reading is meaningful in the context of this research to practice gap.

The program did not boost the students’ segmentation across both cohorts. Segmentation skills are more difficult for young students than blending (Anthony & Francis, 2005). A previous study also found that a computer intervention for grade 1 students had positive effects on blending, but not on segmentation (Savage et al., 2009). Segmentation skills may take longer to influence, or may rely more on production skills (voice and spelling output), which are difficult to monitor in computer programs. Pseudoword decoding showed a significant, small effect only in Study 2. This may be because the students used the Ooka program, on average, for more time in Study 2 than Study 1. Pseudoword reading is difficult for Kindergarten students and may respond only after more time is spent on the decoding skills.

Students who used the Ooka Island program showed higher outcome scores on reading comprehension than students who did not use the program – differences associated with medium effect sizes across the two cohorts. Relatively fewer computer-based programs target reading comprehension than decoding related skills, which may explain why overall reading achievement measures have not shown consistent improvements (Jamshidifarsani et al., 2019). More specifically, programs have not always shown benefits for Kindergarten (Cassady et al., 2018; Savage et al., 2013) or grade 1 students’ comprehension (Piquette et al., 2014; Savage et al., 2013). The Ooka program includes 85 electronic storybooks and learning activities, such as those focused on vocabulary and story sequencing. Ooka Island is thus a more complete reading program than those that primarily target decoding related skills. Computer programs may need to integrate decoding and comprehension activities to directly impact reading comprehension.

There has been a lack of research on early reading that include outcomes measures of word reading, reading comprehension, and book reading-levels. Across both cohorts, students who used the Ooka Island program had higher teacher-assessed book reading-levels; findings associated with small to medium effect sizes. Furthermore, teachers reported that the program was easily incorporated into their classrooms. Previous research has shown that the extent to which teachers are educated about the computer-assisted program and supported in program implementation is related to program effectiveness (Archer et al., 2014). One session with the Ooka Island team for technical support at startup and availability for any arising difficulties was adequate for teacher implementation. The Ooka Island activities were designed to be self-explanatory for young children to use with minimal supervision (Scholastic F.I.R.S.T.^TM, 2017).

Limitations

The results of the current study need to be interpreted within the context of its limitations, some of which arose given that the studies took place in schools. Classrooms were not randomly assigned. Principals deemed that it would be fair for the teachers who responded to an invitation first to get the use of the Ooka Island program. Nonrandom assignment and small n’s raise concerns about generalizability. It is reassuring that the findings across the two years were consistent, however, the findings need further replication. An additional limitation was the business-as-usual control group, similar to most comparable studies with young children (Kyle et al., 2013; Macaruso & Rodman, 2011; Macaruso & Walker, 2008; Paterson et al., 2003; Piquette et al., 2014; Savage et al., 2009). It is not always possible in field research to implement a second intervention for comparison groups. This study design addressed the question of whether computer-based programs are effective supplements for early elementary classrooms that follow balanced literacy approaches. Future research will need to address whether the benefits are sustained over time and whether computer reading programs would have similar benefits in the context of classroom approaches that teach foundational word reading skills through a structured literacy approach. Another limitation was that the number of students who were learning English as an additional language was relatively small. Although differences in response to the program were not found in this study, participant numbers were too small to make conclusions about the effects of the program on English language learners.

Conclusion

In summary, this was the first examination of student outcomes with the Ooka Island computer reading program. This study is consistent with some previous literature, adding to the evidence that computer-based reading programs can have positive effects on the phonological and word reading skills of young students (Savage et al., 2009; Savage et al., 2013). Extending previous research, the Kindergarten students who used the computer-assisted reading program also showed advantages on reading comprehension and on teacher-assessed book reading-levels. These results were found within the context of classrooms that used a balanced literacy approach to English Language Arts instruction. The Ooka Island program follows recommendations for effective programming concerning systematic presentation of content, familiar instructional routines, immediate feedback, and inclusion of the foundational aspects of reading (Scammacca et al., 2007) and includes one focus on story book reading. The results of this study support the position that comprehensive early-reading computer programs may be beneficial for Kindergarten students when used as a supplement to regular classroom instruction.

Acknowledgements

The authors wish to thank for teachers and children who took part in this study, and the education administrators who facilitated the completion of this research in their schools.

Disclosure statement

The authors report there are no competing interests to declare. The authors do not have any affiliation or financial interest with Scholastic or with the Ooka Island/Scholastic F.I.R.S.T. program.

Additional information

Funding

This research was made possible through Mount Saint Vincent Research funds for the activities of the Jarislowsky Research Chair in Learning Disabilities. Jarislowsky Foundation.