Printer Friendly

CAN COLOR-CODED WORD-FAMILIES FLASHCARDS FACILITATE WORD RECOGNITION IN KINDERGARTENERS?

The Common Core State Standards (CCSS), adopted by 45 states, outline several 'foundational skills' in language arts that are essential for kindergarteners to be able to read, comprehend, and interpret grade-level texts. One standard in the Phonics and Word Recognition strand of the CCSS states that kindergarteners will be expected to apply knowledge of letter-sound correspondences to identify unfamiliar words. Mapping sounds onto printed letters allows children to learn to read many words that they already have in their oral vocabulary (Ehri, 1992; Share, 1995; Ziegler & Goswami, 2005). Additionally, decoding using letter-sound correspondences is an effective strategy for acquisition of word recognition skills in beginning readers (Adams, 1990; NICHD, 2000). However, decoding may not be the best approach for emergent readers in kindergarten. Contrary to the CCSS, many kindergarteners are unable to use letter-sound correspondences to acquire word recognition skill because they cannot isolate individual sounds called phonemes. The ability to isolate and manipulate phonemes within spoken words is a Common Core standard within the Phonological Awareness strand that kindergarteners are expected to achieve by the end of the school year. These two standards contradict what many teachers know as effective practice, and they ignore an established body of research indicating that many kindergarten-age children lack sufficient phonemic awareness to be able to isolate and manipulate phonemes and use letter-sound correspondences to decode words (Liberman, Shankweiler, Fischer, & Carter, 1974; Moustafa, 2014; Treiman, 1985; Treiman & Zukowski, 1990).

Most children do not fully develop phonemic awareness until the end of first grade as they learn letters and corresponding sounds, and as they acquire greater exposure to and practice with printed words during reading instruction (Goswami & Bryant, 1990; Torgesen & Mathes, 2000; Ziegler & Goswami, 2005). However, by age four or five, children have an awareness of onsets and rimes (Anthony, Lonigan, Driscoll, Phillips, & Burgess, 2003; Cisero & Royer, 1995; Ziegler & Goswami, 2005). They can identify the initial consonant sounds in a spoken syllable or monosyllabic word (e.g., the /t/ in tip), and they can recognize when spoken words share a rime, which is the vowel and remaining consonant sounds (e.g., the /Ip/ in tip; Goswami, 1995). Words rhyme when they share a rime. Later, as children develop phonemic awareness, they can recognize that the constituent phonemes of the rime are units themselves (e.g., the /Ip/ can be broken down into /I/ and /p/) and recognize the individual phonemes that comprise complex onsets, such as the /s/ /t/ and /r/ in struck.

This progression from simpler to more complex forms of phonological awareness suggests that kindergarteners may be more developmentally ready to learn to read words using a reading-by-analogy strategy (also called an orthographic rime analogy) because it capitalizes on their onset-rime awareness, allowing children to read an unfamiliar word (e.g., rip) by analogy to a familiar word with a spelling pattern representing the same rime (e.g., tip). Using onsets and rimes to identify unfamiliar words in English may be more useful than phonemes because they have fewer inconsistencies in their sound-to-print mappings than grapheme-phoneme correspondences (Ziegler & Goswami, 2005). Based on this research literature, some experts call for revising the CCSS standards to emphasize a focus on learning to read words by analogy using onsets and rimes (Moustafa, 2014).

Our intent is not to claim that there is a single or correct way to teach young children to read words. Rather, we suggest that some kindergarteners may benefit from reading-by-analogy using onsets and rimes if they are not developmentally ready to profit from decoding using grapheme-phoneme correspondences. Therefore, we have focused our research on a practical issue concerning whether prereaders could be taught a reading-by-analogy strategy in naturalistic settings, such as schools or at home, and whether this approach would facilitate development of their word recognition skill. To our knowledge, there are only two published studies on reading-by-analogy conducted in settings outside the laboratory. One is a study by Peterson and Haines (1992). Kindergarteners in an experimental group were trained to read words using a clue word analogy method (e.g., learning bold, colt, cone using cold as a clue word). Training occurred in seven, 15-minute sessions within a month. The authors used a clue-word test to assess word recognition, but the clue words and target words in the test differed from the clue and target words in the training condition. Children in the experimental group made significantly greater gains on reading test words by analogy than control group children, and this benefit was especially apparent for children with average or above-average onset-rime ability.

Walton, Walton, and Felton (2001, Experiment 2) compared rime analogy training to grapheme-phoneme correspondence training in prereading kindergarteners. Children were randomly assigned to a rime analogy group, letter recoding group (sounding out letter-by-letter), or control group (no contact with experimenters). In two 20-minute sessions per week over 11 weeks, children received 1 to 2 minutes of direct instruction in their assigned strategy then practiced the strategy by playing games in small groups. The researchers assessed word recognition with words that could be read by analogy, words that required recoding, and nonwords. Analogy and recoding test words were first presented randomly and then again using a clue word. Compared to the control group, children who received rime analogy training read more test words that were similar in spelling pattern to the training words, but only when given clue words. Also, they showed no advantage when reading words with spelling patterns that were different from the training words.

These school-based studies revealed inconclusive results about whether there is a benefit of reading-by-analogy training on word recognition. Even though Peterson and Haines (1992) found positive results, their testing approach calls into question whether children learned the rime analogy strategy and could use it spontaneously to identify new words without needing the clue word. A better approach would have been to provide trained words in isolation with new, untrained words having the same rime. Unlike Peterson and Haines, the Walton et al. (2001) did not find a positive benefit of rime analogy training. Further, the authors acknowledged that kindergarteners may require more extensive instruction and practice than what was provided.

To our knowledge, there are no recent studies in field settings examining the efficacy of reading-by-analogy training for teaching children to read words. Our aim was to create an evidence-based method for teaching young children to read words that could be easily used by adults with little formal training. Therefore, the primary goal of our research was to evaluate the efficacy of teaching kindergarteners in naturalistic settings (home and school) to read-by-analogy using color-coded word-families flashcards. To that end, our research implemented two important improvements over past studies in the literature.

Our first improvement was in the design of the intervention itself. Instead of clue-word training, direct instruction, or small-group games (e.g., Peterson & Haines, 1992; Walton et al., 2001), we used a simple flashcard method that made the onset-rime pattern salient. Borrowing from Bruck and Treiman (1992) who highlighted letters within words to indicate similar spelling patterns, we developed color-coded word-families flashcards. A word-family is a set of words that share a rime (e.g., bat, cat, fat, all share the -at rime). Word-families rely on a reading-by-analogy strategy; once children know that a-t says /aet/ they should be able to learn all the words in a family because they share the -at rime. In our approach, the rime of words in a word family was shown in black ink to indicate that this spelling pattern in all the words has a constant phonological representation (e.g., the "at" says /aet/), and initial consonants were shown in a unique color to indicate that this sound changes from word to word.

The use of word-families flashcards is consistent with the way children might learn to read words in naturalistic settings. Kindergarten teachers often use word-families to highlight consistencies between letter units and their sounds, and parents are familiar with rhyming words in children's poems, books, and songs. Flashcards are also ubiquitous in early childhood, being used by parents and teachers for all sorts of learning (e.g., alphabet letters, sight words, names for objects, numerals, math facts).

The second improvement was our testing approach. After an exhaustive search of standardized word recognition tests, we found there were none appropriate for beginning kindergarteners who may be younger than age five. Therefore, we developed our own test, which allowed us to systematically evaluate children's word-level skills in a way that was directly relevant to the word-families intervention. Our word recognition test consisted of 16 words all printed in black ink, where half of the words were practiced in the intervention and the other half were new words with the same rime units as flashcard words. Constructing the test in this way allowed us to assess whether children actually used the color-coded onsets and rimes in the flashcards to learn words by analogy rather than by memorizing words or learning associations between colors and letter patterns.

We present a set of three novel experiments with kindergarteners in three subsequent school years to evaluate the efficacy of reading-by-analogy using color-coded word-families flashcards. Experiment 1 involved an 8-week, at-home intervention compared to a wait-list control. Experiments 2 and 3 were conducted in schools to improve the intervention by reducing variability in the amount of practice and to make the intervention more standardized. The word-families intervention was compared to a vocabulary-building control group in Experiment 2 and a control group which listened to stories in Experiment 3. In these control groups, children received an equivalent experience that could help them acquire reading skills.

General Methodology

Participants

In Experiments 1 and 2, kindergarteners were from the two schools (A and B) within the same district. In Experiment 3, School B was not available, so School C was added. Random assignment to intervention and control groups occurred within classrooms to reduce any effect of classroom or school, which includes differences in literacy instruction among teachers. Due to small samples, we pooled the groups across the schools. Table 1, which shows demographic variables of schools for the three experiments, indicates that schools within each experiment were comparable on several key variables. Exceptions were in Experiment 2, where schools differed in gender composition, percent of students receiving free/reduced lunch, and percent receiving reading remediation, and in Experiment 3 where schools differed in percent of children speaking English at home. Only gender composition in Experiment 2 and percent of children speaking English at home were significant. Gender composition is not directly related to our hypotheses or outcome variable, and while home language is important, the schools had nearly identical phonological awareness performance (means and standard deviations) and were not significantly different on this variable which is directly related to the outcome of word recognition test performance. Therefore, in all experiments, the comparability of schools on variables directly relevant to the outcomes justified our decision to pool across schools.

Materials

Reading log. For Experiment 1, families were given a reading log with instructions for practicing flashcards. The first step involved pointing to the black-ink letters in each word and stating that 'all of the words rhyme because they say...' (pronouncing the rime unit). Then, parents pointed to the initial letter of each word and asked their child the sound of the letters (the onsets printed in colors). If children did not know, parents could tell them, and this step could be skipped if children knew the letter sounds (which was important to reduce repetition and maintain student interest later in the intervention). At this step, parents also pointed out digraphs (e.g., ch, kn, qu, sh, wh) if they were in the flashcard set and told children that these two letters together (always shown in blue) make one sound. Parents read each word first before asking children to read through the flashcards. The instructions suggested that parents shuffle the flashcards after every few trials so children were not memorizing word order. Parents were instructed not to coerce the children to practice each set a certain number of times. On each reading log, parents circled the number of times children practiced the flashcard set from Monday through Thursday. Materials were returned to school on Fridays. In Experiments 2 and 3, which had research assistants (RAs) conducting the intervention in schools, we trained RAs using the same instructions in the parent reading log.

Flashcards. We selected word-families containing 10 or more words with single consonant onsets or digraph onsets from the Reading Teacher's Book of Lists (Fry, Kress, & Fountoukidis, 1993) to form a pool of 17 word-families. For Experiments 1 and 3 involving an 8-week intervention, we created flashcard sets by randomly selecting eight word-families from the pool: -ail, -ag, -ap, -at, -et, -ill, -ot, -ug. For Experiment 2 involving a 16-week intervention, we randomly sampled eight word-families from the remaining pool: -ack, -ad, -ake, -ay, -ell, -est, -in, -ip.

We randomly selected 10 words from each word family to create the flashcard sets (-ake had only 10). Black ink was used for the rimes to indicate that this spelling pattern has a constant phonological representation for all the words. The initial consonant was shown in a unique color (e.g., b was sea green, c was sky blue, etc.) to indicate that this sound changes from word to word. Because there were more consonants in the alphabet than shades of colors in our word processing software, some colors were re-used, taking care not to use the same color for letters that are confusing to young children (e.g., b, p, d). Also, digraphs (e.g., ch) at the beginning of some words were always displayed as bright blue to indicate that the two letters together make one sound.

Measures

Survey. In all experiments, parents provided children's birthdate, gender, and primary language spoken at home. For Experiments 2 and 3, we asked whether the child received free/reduced lunch, which is the measure that school districts use to report socioeconomic status (SES) of its students. In Experiment 3, after posttesting, teachers also provided information on which children received remedial reading intervention.

Phonological awareness (PA). It was important to assess PA because of its relationship to the development of word recognition skill (Adams, 1990; Wagner, Torgesen, & Rashotte, 1994). We used the Sound Matching subtest of the Comprehensive Test of Phonological Processing (CTOPP; Wagner, Torgesen, & Rashotte, 1999) (1) which assesses beginning and ending sounds. The CTOPP battery has adequate evidence supporting its reliability and validity (Hurford, 2003; Wright, 2003). Total raw scores were converted to age-based standard scores (M = 10, SD = 3).

Word recognition test. In each experiment, the word recognition test consisted of 16 words. Experiments 1 and 3 used the same test because both used the same word-families for the intervention. To construct the test, eight words were randomly selected from the words practiced during intervention, and eight words were randomly selected from the words not used to make the flashcards but from the same word-families (i.e., unfamiliar words with the same rime). The order of words was randomized, and one order was used for all children. Four words appeared per page, centered and printed in 48-point Times New Roman black font. We used the same procedure to create the test for Experiment 2, randomly sampling eight practiced words from the 16 sets of flashcards and eight words from the list of unused flashcard words. The tests have high internal consistency with Cronbach's alpha coefficients ranging from .83 to .94 for all administrations across the three experiments. (2)

Children were given three practice trials with feedback to help them understand the task. They received 1 point for each word correctly identified or decoded, yielding a range of scores from 0 to 16. In Experiment 3, we also recorded response time on the test.

Procedure

Parents completed the survey before pretesting. RAs individually administered the CTOPP first, then the word recognition test several days apart. Our general procedure is to test all children within each classroom on one test (e.g., CTOPP) until everyone is done (which can take several days) and then begin administering the next test (e.g., word recognition) in roughly the same order of children (excepting absences) so that the interval between tests is similar for all children. Random assignment occurred after pretesting was completed. Post-testing began one or more days following the last day of intervention.

Experiment 1

Participants

Twenty-six kindergarteners from two schools (see Table 1) were randomly assigned within school to the intervention group (N = 8 and 6, respectively by school) or a wait-list control group (N = 7 and 5, respectively by school). Schools were combined to yield one intervention group (N = 14) and one control group (N = 12).

Procedure

Each child in the intervention group was randomly assigned a different order of the flashcard sets (-ail, -ag, -ap, -at, -et, -ill, -ot, -ug) and practiced a new set each week for 8 weeks. All students were post-tested in December and in April after the control group received the same 8-week intervention from February to April.

Results and Discussion

Children in the intervention group practiced an average of 9.13 times per week (range: 3-15). Table 2 (top section) displays word recognition performance of the intervention and control groups at pretest and December posttest. Because distributions were extremely skewed at pretest, two separate Mann-Whitney U tests were conducted to examine median score differences between groups at pretest and posttest, with alpha set at .025 one-tailed to control for Type I error. As shown in Table 2, the median number of words correctly identified at pretest was zero for both groups, U = 75, p > .05. However, the difference between groups at posttest was significant, U = 46, p = .04 (one-tailed), r = .347. The advantage of the intervention group is represented by the moderate effect size of .35 (Cohen, 1992). Furthermore, Figure 1 shows that the positively skewed distribution of the control group remains unchanged from pretest to posttest (panels c and d), whereas the posttest distribution for the intervention group changes from positively skewed (panel a) to bimodal (panel b).

As shown by the bimodal distribution in Figure 1, not all children in the intervention group made word recognition gains. PA is one factor that might have affected the results. Figure 2 shows word recognition performance over the three test administrations as a function of group and PA level (obtained by a median split of CTOPP pretest standard scores). Although sample sizes are too small for statistical analyses, the pattern suggests that children with higher PA benefit more from word-families practice than those with less PA knowledge. Students in the intervention group with higher PA showed a word recognition advantage over students with lower PA throughout the study. Students in the control group with higher PA made dramatic gains during their 8-week intervention, whereas those with lower PA made virtually no gain at all. This finding is consistent with the results of Peterson and Haines (1992) which showed that children with higher onset-rime awareness made significantly greater word learning gains compared to controls with similar levels of skill, whereas children with lower onset-rime skill showed no advantage over controls.

While these results are promising, we acknowledge limitations to our conclusions. First, there was wide variation in the amount of practice. Only 9 of 14 participants completed all 8 weeks of flashcards. Also, many parents did not return reading logs, so we cannot be certain whether their children were practicing the flashcards. Finally, the modest results might suggest the need for a longer intervention. Experiment 2 was designed to address these issues.

Experiment 2

Participants

Twenty-seven kindergarteners from two schools (see Table 1) were randomly assigned within classroom to the intervention or a vocabulary-building control group. Schools were combined to yield one intervention group (N = 14) and one control group (N = 13).

Materials

The vocabulary-building control condition utilized photo cards from the Elements of Reading curriculum (Beck & McKeown, 2004). One side showed a photo depicting a vocabulary word (e.g., comforting) and the other side showed a sentence using the word. Researchers followed the activities in the curriculum manual. Children were exposed to five new vocabulary words in each of the 16 activities, one for each week of the intervention. Inclusion of this control group improves upon the studies mentioned earlier which did not give control groups an equivalent activity (Peterson & Haines, 1992; Walton et al., 2001, Experiment 2).

Procedure

RAs worked individually with intervention and control children in a quiet corner outside their classrooms. The order of flashcard sets was standardized (-at, -ad, -ag, -ack, -ap, -ip, -in, -ill, -et, -ell, -est, -ot, -ug, -ail, -ake, -ay), and children in the intervention group practiced one flashcard set each week, 3 times in a session (taking about 5 minutes) for 3 days per week over 16 weeks. We chose this amount of practice because it was consistent with the average amount of practice from Experiment 1. Children in the control group received the same amount of time with the RA as the intervention group, practicing a set of 5 vocabulary cards each week for 3 days per week over 16 weeks. During posttesting in April, RAs who worked with children in one classroom/school during the intervention were switched to a different classroom/school to prevent any bias in test administration and scoring.

Results and Discussion

Table 2 (middle section) shows pretest and posttest scores on word recognition for both groups. Unlike Experiment 1, the control group made slightly greater word recognition gains than the intervention group. The control group also showed more variability at pretest. One student earned a pretest score of 15 (out of 16), but removal of this outlier did not change the results. Fewer students in the control group received free/reduced lunch (15.4%) compared to the intervention group (28.6%), suggesting that they may have had a higher socioeconomic status (SES).

Importantly, the intervention group did improve in word recognition. Modes of 3 and 10 in the posttest distribution indicated that some children performed comparable to the control group, but others did not improve much. To further explore this pattern, we divided the intervention group into high- and low-scorers based on a median split of posttest scores. The high-scorers performed significantly better on the CTOPP at pretest (M =11.86, SD = 1.95) compared to low-scorers (M = 9.17, SD = 1.72), t (11)= 2.61, p < .05, d = 1.46. Also, significantly more low scorers received free/reduced lunch (50%) compared to high scorers (0%), [X.sub.2] (1, N = 13) = 4.55, p < .05, [phi] = .59. These results suggest that intervention children in the high-scoring group had higher SES and PA than those in the low-scoring group at the outset of the study.

Taken together, these results suggest that SES and PA moderate the effectiveness of the word-families intervention. Control group children made greater gains than intervention children and may have had higher SES based on their free/reduced lunch status. Children in the intervention group who were high-scoring at posttest also appeared to have higher SES as well as better CTOPP performance at pretest, indicating higher initial PA. It is well documented in the research literature that SES is related to achievement, particularly reading achievement (Bowey, 1995; McLoyd, 1998; Sirin, 2005; White, 1982). Control group children and higher-scoring intervention children may have had more opportunities at home to develop important readiness skills, such as letter identification and phonological awareness, leading them to know more words on the word recognition test at pretest, and to have developed strategies for learning to read words. A well-known effect in reading, called the Matthew effect, supports this explanation. The Matthew effect illustrates a pattern whereby good readers (in this case, children who begin school prepared for reading) read more, especially outside of school, and acquire reading skills at a faster rate than children who begin instruction lacking certain skills (Stanovich, 1986).

Experiments 1 and 2 suggest that PA may impact the effectiveness of the word-families intervention. We designed Experiment 3 to better control for PA.

Experiment 3

Participants

Fifty-six kindergarteners from two schools (two classrooms in one school and three classrooms in the other) were matched on CTOPP pretest scores within each classroom, and matched pairs were randomly assigned to groups. Classrooms and schools were combined to yield one intervention group and one control group. Two students did not provide assent and were dropped from the sample. Nine students were excluded from analyses because they were absent for posttesting (2 intervention group; 1 control group), or because they were assigned to the intervention group but completed none of the intervention (N = 6). The final sample comprised 45 children, 21 in the intervention group and 24 in the control group.

Procedure

The same flashcard sets as in Experiment I were used, but the order of sets was standardized for all children: -at, -ag, -ap, -ill, -et, -ot, -ug, -ail. We encountered many disruptions in the planned intervention procedure (practicing a flashcard set 3 times per day for 3 days), such as one school starting later than the other and lengthy school closures due to two severe weather events. We decided to allow intervention children to practice as much as they needed and then move onto the next word-family set. Therefore, the number of word-family sets completed and the amount of practice with each set varied, as in Experiment 1. The control children listened to an RA read a book from the classroom for an equivalent amount of time as the RA worked with children in the intervention group. At the December posttest, RAs switched classrooms or schools to test different children than the ones with whom they intervened.

Results and Discussion

Table 2 (bottom section) indicates that word recognition performance at posttest was similar for both groups. One reason may be lack of sufficient practice due to the many interruptions in the procedure. As support, the number of completed word-family sets moderately and significantly correlated with word recognition posttest accuracy scores (r = .591, p < .01) after controlling for pretest performance. Additionally, number of completed sets moderately and significantly correlated with word recognition posttest response time scores after controlling for pretest scores (r = -.503, p < .05). Children who completed more sets were faster on the posttest.

To further examine the effect of practice, we divided the intervention group into a high-practice (6-8 sets) and low-practice group (5 or fewer sets) based on a median split of the number of completed word-family sets. Table 3 shows word recognition performance for these groups compared to control. Because pretest distributions were skewed, two separate Mann-Whitney U tests were conducted on pretest and posttest accuracy scores with alpha at .025 one-tailed controlling for Type 1 error. Group differences were not significant at pretest (U = 41.5, p > .05), but they were at posttest, U = 24.5, p < .02, one-tailed, r = .47. Also, no students in the high-practice group received reading remediation compared to 45.5% for the low-practice group, [X.sub.2] (1, N = 21) = 5.97, p < .02, [phi]= .53. Taken together, these results suggest that word-families flashcards can facilitate word recognition skills if children have adequate practice, and that the amount of practice needed may depend on the reading readiness of the child.

General Discussion

In this novel set of experiments, we used word-families flashcards with colored letters to investigate the efficacy of reading-by-analogy training in a field setting. Our goal was to create a method that could easily be used by parents and paraprofessionals with little formal training. To that end, our brief experiments in home and school settings have provided evidence supporting the use of rime analogy training for word learning. Experiment 1 results indicated that a modest amount of weekly practice with word-families flashcards over eight weeks enabled children in the intervention group to outperform control children on reading test words without the aid of the color-coding from the flashcards. This contrasts with the school-based research by Walton et al. (2001, Experiment 2) which found weak treatment effects for analogy training with kindergarteners over an 11-week span, and no benefit of analogy training when words were read without clue words.

Experiments 2 and 3 attempted to replicate the findings of Experiment 1 in school settings. We used control groups that received an equivalent amount of contact with RAs, which improved upon both previous school-based studies (Peterson & Haines, 1992; Walton et al., 2001, Experiment 2). However, our interventions in both experiments did not show a clear benefit of the word-families flashcards training. Several background characteristics of children, such as their SES, PA, and whether they received reading remediation, affected the efficacy of the intervention. In Experiment 2, intervention children who had higher PA and higher SES making greater gains than those with lower PA or SES. In Experiment 3, students receiving reading remediation made fewer word recognition gains than those who did not receive remediation. Importantly, children who practiced more flashcard sets clearly benefitted compared to those who practiced fewer sets. In sum, the results from all three experiments suggest that children can learn to read words using color-coded word-families flashcards with sufficient practice, and that to benefit from this approach they need foundational PA skills. Indeed, Goswami (1994) emphasized the importance of PA as a foundation for using orthographic rime analogies.

Despite our improvements over past research, several design limitations warrant discussion. First, our small samples required pooling of data from classrooms and schools. Although we demonstrated comparability of classrooms and schools, pooling may have resulted in subtle differences between experiments that prevented replication. Additionally, the small sample sizes reduce the generalizability of our findings. Second, the constraints of the schools' schedules limited how long our RAs could work with the children. Finally, as we have already discussed, conducting research in schools creates many factors outside of researchers' control such as variation among children in characteristics related to the study outcomes (e.g., PA and SES), and events that alter the original research design.

That we found evidence for the benefit of a word-families intervention amongst all the 'noise' in our data collection suggests it is possible to use such an approach in real-life settings like the kindergarten classroom. Typically, large effect sizes are obtained in controlled intervention research conducted in laboratories, but when interventions are moved to the field, they lose some degree of efficacy due to added variability and lack of controls in naturalistic settings (Hulleman & Cordray, 2009; Kupper-Tetzel, 2014). Across our three experiments, we obtained moderately strong effect sizes despite the challenges of data collection in real-world settings, which suggests that our word-families flashcard intervention can be efficacious in many kindergarten classrooms like those in our own studies. Additionally, implementing the intervention with children in real school classrooms gave us the opportunity to examine how children's backgrounds and characteristics would impact results. This information itself is important for early childhood teachers. Like any intervention, the word-families intervention is not a one-size-fits-all approach. Teachers should be aware that practice will need to vary depending on children's skills, and that prereading skills such as letter knowledge and PA need to be sufficiently developed to effectively use this approach. We have already discussed the importance of PA, particularly onset-rime awareness, for orthographic rime analogy training. Letter knowledge is also important because identifying the new word requires children to know the letter combinations that represent the onsets in the existing word and target word (Roberts & McDougall, 2003) little is known about the skills and strategies children actually use and how these compare with their everyday reading of single words. This study, with 4- and 5-year-olds (N=125. Therefore, a word-families intervention is no different than any other approach to word learning in requiring these two important skills.

Our set of experiments suggests that using color-coded word-families flashcards to highlight the onset-rime patterns in words can be implemented in authentic settings to facilitate development of word recognition in kindergarteners. Future research with larger samples and more systematic controls for extraneous factors can be conducted to further explore the benefits of using word-families to teach kindergarteners to read words. As stated earlier, the aim of our research was not to engage in a debate over the Common Core State Standards regarding how literacy instruction should proceed, or to demonstrate that teaching onset-rime units is the better, or only approach. Rather, like Anthony and Lonigan (2004), we believe that it is more important to consider the role of individual differences in phonological sensitivity and to choose the approach that is most developmentally appropriate for individual children. If kindergarteners find word learning using grapheme-phoneme correspondences difficult, word-families flashcards that use colors to highlight onset-rime patterns may be a reasonable alternative.

Author Note

Parts of Experiments 1 and 3 were presented at the annual meeting of the American Educational Research Association. Parts of Experiment 2 were presented at the annual meeting of the Eastern Psychological Association.

References

Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press.

Anthony, J. L., & Lonigan, C. J. (2004). The nature of phonological awareness: Converging evidence from four studies of preschool and early grade school children. Journal of Educational Psychology, 96(1), 43-55. http://doi.org/10.1037/0022-0663.96.1.43

Anthony, J. L., Lonigan, C. J., Driscoll, K., Phillips, B. M., & Burgess, S. R. (2003). Phonological sensitivity: A quasi-parallel progression of word structure units and cognitive operations. Reading Research Quarterly, 38, 470-487.

Beck, I. L., & McKeown, M. G. (2004). Elements of reading: Teacher edition level K. Orlando, FL: Steck-Vaughn.

Blaustein, S. H. (2017). Review of the Comprehensive Test of Phonological Processing-Second Fdition. In J. F. Carlson, K. F. Geisinger, & J. L. Jonson (Eds.), The twentieth menial measurements yearbook. Lincoln, NE: Buros Institute.

Bowey, J. A. (1995). Socioeconomic status differences in preschool phonological sensitivity and first-grade reading achievement. Journal of Educational Psychology, 87(3), 476-487.

Brack, M., & Treiman, R. (1992). Learning to pronounce words: The limitations of analogies. Reading Research Quarterly, 27(4), 375-388. http://doi.org/Doi10.2307/747676

Cisero, C. A., & Royer, J. M. (1995). The development and cross-language transfer of phonological awareness. Contemporary Educational Psychology, 20, 275-303.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.

Ehri, L. C. (1992). Reconceptualizing the development of sight word reading. In P. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 107-143). Hillsdale, NJ: Erlbaum.

Fry, E. B., Kress, J. E., & Fountoukidis, D. L. (1993). The reading teacher's book of lists (3rd ed.). Englewood Cliffs, NJ: Prentice flail.

Goswami, U. (1995). Phonological development and reading by analogy: What is analogy and what is it not? Journal of Research in Reading, 18(2), 139-145.

Goswami, U., & Bryant, P. E. (1990). Phonological skills and learning to read. Hillsdale, NJ: Erlbaum.

Hulleman, C. S., & Cordray, D. S. (2009). Moving from the lab to the field: The role of fidelity and achieved relative intervention strength. Journal of Research on Educational Effectiveness, 2, 88-110. doi: 10.1080/19345740802539325

Hurford, D. P. (2003). Review of the Comprehensive Test of Phonological Processing. In B. S. Plake, J. C. Impara, & R. A. Spies (Eds.), The fifteenth mental measurements yearbook. Lincoln, NE: Buros Institute.

Kupper-Tetzel, C. E. (2014). Strong effects on weak theoretical grounds: Understanding the distributed practice effect. Zeitschrift fur Psychologie, 222, 71-81. doi: 10.1027/2151-2604/a000168

Liberman, I., Shankweiler, D., Fischer, F., & Carter, B. (1974). Explicit syllable and phonemes segmentation in the young child. Journal of Experimental Child Psychology, 18, 201-212.

McLoyd, V. C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53, 185-204.

Moustafa, M. (2014). Improving Common Core's foundational skills in reading, K-2. The California Reader, 48(1), 11-19.

National Institute of Child Health and Human Development (N1CHD). (2000). Report of the National Reading Panel. Teaching children to read: Reports of the subgroups (NIH Publication No. 00-4754). Washington, DC: U.S. Government Printing Office.

Perfetti, C. A. (1992). The representation problem in reading acquisition. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 145-174). Hillsdale, NJ: Erlbaum.

Perfetti, C.A. & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven (Ed.), Precursors of functional literacy (pp. 189-213). Philadelphia, PA: John Benjamins.

Peterson, M. E., & Haines, L. P. (1992). Orthographic analogy training with kindergarten children: Effects on analogy use, phonemic segmentation, and letter-sound knowledge. Journal of Reading Behaviour, 24, 109-127.

Roberts, L., & McDougall, S. (2003). What do children do in the rime-analogy task? An examination of the skills and strategies used by early readers. Journal of Experimental Child Psychology, 84(4), 310-337. http://d0i.0rg/10.1016/S0022-0965(03)00029-8

Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55(2), 151-218.

Sirin, S. R. (2005). Socioeconomic status and academic achievement: A meta-analytic review of research. Review of Educational Research, 75(3), 417-453.

Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21(4), 360-406.

Torgesen, J. K., & Mathes, P. G. (2000). A basic guide to understanding and teaching phonological awareness. Austin, TX: Pro-Ed.

Treiman, R. (1985). Onsets and rimes as units of spoken syllables: Evidence from children. Journal of Experimental Child Psychology, 39, 161-181.

Treiman, R. & Zukowski, A. (1990). Levels of phonological awareness. In S. Brady & D. Shankweiler (Eds.), Phonological processes in literacy (pp. 67-83). Hillsdale, NJ: Krlbaum.

Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). the development of reading-related phonological processing abilities: New evidence of bi-directional causality from a latent variable longitudinal study. Developmental Psychology, 30, 73-87.

Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1999). The Comprehensive Test of Phonological Processing. Austin, TX: Pro-Ed.

Walton, P. D., Walton, L. M., & Felton, K. (2001). Teaching rime analogy or letter recoding reading strategies to prereaders: Effects on prereading skills and word reading. Journal of Educational Psychology, 93(1), 160-180. http://doi.org/10.1037/0022-0663.93.1.160

White, K. R. (1982). The relation between socioeconomic status and academic achievement. Psychological Bulletin, 91(3), 461-481.

Wright, C. R. (2003). Review of the Comprehensive Test of Phonological Processing. In R. S. Plake, J. C. Impara, & R. A. Spies (Eds.), The fifteenth mental measurements yearbook. Lincoln, NE: Buros Institute.

Ziegler, J. C., & Goswami, U. C. (2005). Reading acquisition, developmental dyslexia, and skilled reading across languages: A psycholinguistic grain size theory. Psychological Bulletin, 131, 3-29.

Southern Connecticut State University

Southern Connecticut State University

(1) CTOPP was used rather than CTOPP-2 because the subtests in both versions are identical (only items are re-ordered) and much of the validity evidence in the CTOPP-2 manual combines CTOPP and CTOPP-2 (Blaustein, 2017), making it difficult to determine whether CTOPP-2 had improved validity compared to the original test. This justified our decision to continue using CTOPP because of limited funds to purchase the new test.

(2) Separate reliability coefficients by test administration and experiment can be provided to readers. For economy of wording, we chose to present the range of coefficients.
Table 1. Demographic Variables by School for Each Experiment

                                     Experiment 1
                                        School
Variable                 School A (N = 15)  School B (N = 11)

Mean Age (mos.)          63.00 (3.25) (a)      64.18 (3.31)
Mean CTOPP Score (b)     10.47 (1.77)           9.36 (2.11)
Gender (% Female)        60.0                  45.5
% English Home Language  93.3                  90.9

                                     Experiment 2
                                        School
Variable                 School A (N = 16)  School B (N = 11)

Mean Age (mos.)          63.81 (4.15) (a)      64.55 (4.13)
Mean CTOPP Score (b)     10.63 (2.22)          10.18 (2.71)
Gender (% Female)        62.5                  18.2
% English Home Language  87.5                  81.8
% Free/Reduced Lunch     12.5                  36.4
% Remedial Reading       18.8                   9.1

                                     Experiment 3
                                        School
Variable                 School A (N= 19)   School C (N = 26)

Mean Age (mos.)          64.53 (3.84) (a)      65.54 (4.93)
Mean CTOPP Score (b)     10.84 (2.54)          10.77 (2.14)
Gender (% Female)        52.6                  46.2
% English Home Language  94.7                  69.2
% Free/Reduced Lunch     10.5                   7.7
% Remedial Reading       21.1                  38.5

                                     Experiment 1
                                        School
Variable                            Test               P

Mean Age (mos.)                         t(24) = .91   .373
Mean CTOPP Score (b)                  t(24) = -1.45   .160
Gender (% Female)         [X.sup.2] (1, n=26) = .54   .462
% English Home Language   [X.sup.2] (1, n=26) = .053  .819

                                     Experiment 2
                                        School
Variable                            Test               P

Mean Age (mos.)                         t(25) = .452  .655
Mean CTOPP Score (b)                   t(25) = -.466  .645
Gender (% Female)        [X.sup.2] (1, N=27) = 5.19   .023
% English Home Language   [X.sup.2] (2, N=27) = .167  .920
% Free/Reduced Lunch     [X.sup.2] (1, N=27) = 2.15   .143
% Remedial Reading        [X.sup.2] (1, N=27) = .482  .488

                                     Experiment 3
                                        School
Variable                            Test              P

Mean Age (mos.)                         t(43) = .744  .461
Mean CTOPP Score (b)                   t(43) = -.104  .918
Gender (% Female)         [X.sup.2] (1, N=45) = .184  .668
% English Home Language  [X.sup.2] (1, N=45) = 4.46   .035
% Free/Reduced Lunch      [X.sup.2] (1, N=45) = .109  .741
% Remedial Reading        [X.sup.2] (1, N=45)= 1.55   .213

(a) Standard deviations (SD) in parentheses.
(b) Standard score (M= 10, SD = 3).
Note: CTOPP = Comprehensive Test of Phonological Processing Sound
Matching subtest

Table 2. Pretest and Posttest Word Recognition Performance of
Intervention and Control Groups for Each Experiment

                            Pretest                       Posttest
Group         N        M    Mdn   Mode   SD   Range    M    Mdn    Mode

                                  Experiment 1
Intervention  14      1.79  0     0     2.91   0-9    5.64   5.5   8
Control       12 (a)  1.67  0     0     4.14   0-14   3.27   0     0
                                  Experiment 2
Intervention  14 (b)  1.21  0.50  0.00  1.76   0-5    8.00   8.00  3; 10
Control       13 (c)  2.46  0.00  0.00  4.33   0-15  10.50  11.00  8.00
                                  Experiment 3
Intervention  21      3.81  1.00  0.00  4.83   0-15   5.24   3.00  0.00
Control       24      3.13  0.00  0.00  4.88   0-16   5.00   2.50  0.00

               Posttest
Group          SD   Range

              Experiment 1
Intervention  4.18   0-13
Control       5.26   0-16
              Experiment 2
Intervention  3.85   3-14
Control       4.23   2-16
              Experiment 3
Intervention  5.15   0-16
Control       5.30   0-16

(a) Posttest data are based on 11 participants because of a student
absence at time of testing. (b) Posttest data are based on 13
participants because of a student absence at time of testing. (c)
Posttest data are based on 12 participants because of a student absence
at time of testing.

Table 3. Experiment 3 Word Recognition Performance of High-Practice and
Low-Practice Intervention Groups Compared to Control

                               Pretest                   Posttest
Group          n    M    Mdn   Mode   SD   Range   M    Mdn   Mode   SD

High Practice  10  4.80  3.00   0    5.53  0-15   8.00  9.00   9    5.79
Low Practice   11  2.91  0.00   0    4.16  0-10   2.73  2.00   0    2.87
Control        24  3.13  0.00   0    4.88  0-16   5.00  2.50   0    5.30

               Posttest
Group           Range

High Practice   0-16
Low Practice    0-8
Control         0-16
COPYRIGHT 2019 Project Innovation (Alabama)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Durwin, Cheryl C.; Moore, Dina
Publication:Reading Improvement
Date:Dec 22, 2019
Words:7175
Previous Article:IN-DEPTH STUDY.
Next Article:"CULTURAL CAPITAL": VERSUS "GETTING IT DONE:" AN ADOLESCENT READER'S EXPERIENCE WITH GOTHIC TEXTS.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters