Silent 2

Psychology in the Schools, Vol. 41(2), 2004 Published online in Wiley InterScience (www.interscience.wiley.com).

© 2004 Wiley Periodicals, Inc. DOI: 10.1002/pits.10152

SILENT VERSUS ORAL READING COMPREHENSION AND EFFICIENCY R. STEVE MCCALLUM, SHANNON SHARP, SHERRY MEE BELL, AND THOMAS GEORGE

University of Tennessee, Knoxville Seventy-four students read passages from an individually administered test of reading comprehension (a subtest from the Test of Dyslexia, a test of reading and related abilities currently in development; McCallum & Bell, 2001), and then answered literal and inferential questions. Students were randomly assigned to one of two conditions; 39 students read the passages silently and 35 read orally, with time recorded for each passage read. Comprehension and time were dependent measures for a Multivariate Analysis of Covariance (MANCOVA) and two follow-up Analyses of Covariance (ANCOVA). After controlling for reading ability, results from the MANCOVA showed a significant combined effect ( p ! .05); however, a comparison of mean reading comprehension scores showed no significant difference between silent readers and oral readers ( p " .05). On the other hand, with reading ability controlled, silent readers took significantly less time to complete passages compared to those who read orally ( p ! .02). In fact, students took 30% longer to read orally than silently, on average. When test directions do not specify either oral or silent reading and error analysis is not a goal, testing will be more efficient via silent responding with no loss of comprehension. © 2004 Wiley Periodicals, Inc.

Efficacy of silent versus oral reading has been the focus of a number of investigations with inconclusive results, due perhaps to different sample characteristics, test conditions, and dependent measures across studies. This study was designed to determine whether differences occur in performance and efficiency for a sample of elementary students as a function of reading individually administered test passages silently versus orally. Several researchers in the 1970s found oral reading to produce superior comprehension under some conditions. For example, Swalm (1972) and Elgart (1978) found that young readers (second and third graders, respectively) comprehended better after reading orally. More recently, Fletcher and Pumfrey (1988) found that oral reading led to greater comprehension than silent reading for 7- and 8-year-olds. Passages and comprehension questions were taken from an individually administered standardized reading instrument, typically requiring oral reading of text and oral responding to comprehension questions; two passages were completed for each of the three reading modes. The passages were not presented as a part of a complete test administration. Test format was not specified, i.e., the authors did not indicate whether passages were presented in a one-to-one versus group administration. In contrast, Rowell (1976) and Miller and Smith (1985) found mixed results. Specifically, Rowell found comprehension to be higher under an oral reading condition for urban/suburban fifth graders and for males but not for rural fifth graders and females. In a study of 94 second through fifth graders, Miller and Smith (1985) asked students to read one passage silently and one passage orally. After each passage, the students answered both literal and inferential questions. Low ability readers had significantly higher oral reading scores, while medium ability readers had significantly higher silent reading scores. High ability readers did not differ. Consistent with findings by Miller and Smith (1985) and Swalm (1972), several others have found that low ability or younger readers tend to comprehend better when reading orally (Burge, 1983; Mullikin, Henk, & Fotner, 1992). However, in a study using a design similar to their 1985 study, Miller and Smith (1990) found that poor readers comprehended better following silent reading; as in their earlier study, high ability readers showed no significant differences.

Correspondence to: R. Steve McCallum, Educational Psychology and Counseling, University of Tennessee, Knoxville, Claxton Complex A 525, Knoxville, TN 37996–3452. E-mail: [email protected]

241

242

McCallum, Sharp, Bell, and George

Additional evidence in support of oral reading superiority for some low ability readers was reported by Fuchs and Maxwell (1988). In a study of third through eighth grade boys with learning disabilities, Fuchs and Maxwell examined relationships between silent versus oral reading, oral versus written recall, and structural importance of the text read. When stories were read orally and as the level of text importance increased, comprehension did as well; however, this relationship was not demonstrated under the silent reading condition. According to the authors, oral reading is superior for facilitating reading comprehension for males with learning disabilities, although recall format and level of text importance were identified as potentially mediating variables. A more recent study focuses on the differential effects of silent or oral reading for different age groups, with ability level controlled. According to Prior and Welling (2001), oral reading produced higher comprehension scores in third and fourth graders; however second graders performed equally well under both conditions, contrary to some findings that young readers perform better under oral conditions. In contrast to findings supporting clear superiority of either oral or silent reading, a study by Holmes and Allison (1985) indicated no differences in reading silently versus reading orally (to oneself ). Four modes of reading were compared: oral reading to an audience of one person, oral reading to one’s self, silent reading to one’s self, and reading along silently while listening to content being read on tape. Urban fifth-grade students of average IQ were divided according to reading ability. Each worked independently, following directions provided in a materials packet. Following each passage read, students wrote answers to comprehension questions (literal and higher order) on paper. For the overall group, as well as the good readers, the lowest level of comprehension was associated with reading while listening, but there was no significant difference in comprehension when reading to one’s self orally versus silently. For the overall group, as well as the good and average reader subgroups, reading orally to an audience led to a decrease in higher-level comprehension. For the subgroup of poor readers, no one reading mode produced greater comprehension in general, although more literal comprehension questions were answered when reading orally to an audience than silently to one’s self. Theories explaining the relative merits of oral versus silent reading vary. Research finding support for oral reading has suggested that students comprehend better when reading orally because of the assumed need to concentrate and focus on the words more, to articulate each one (Elgart, 1978). Swalm (1972) suggested that oral reading may be superior because the student is forced to read each word to satisfy a listener. When reading silently, the student can skip over difficult words, which may lead to poorer comprehension. It has also been suggested that reading orally is superior because it allows the student to use two senses, sight and sound, while silent reading uses only the sense of sight (Elgart, 1978; Swalm, 1972). On the other hand, silent reading may be superior because “during oral reading, the child is also faced with the problems of pronunciation, interpretation, and intonation, and many children find the mechanics of reading aloud so absorbing or so difficult that they are unable to attend to the content” (Rowell, 1976, p. 368). Because of the inconsistency of research findings, and because none of the previous findings were clearly obtained from traditionally administered individualized tests, developers and educators who need to evaluate comprehension efficiently cannot recommend unequivocally either an oral or silent format. If one reading format yields consistently better comprehension, test administrators and authors should choose it unless other criteria are important (e.g., error analyses that can be obtained only from oral performance). This study was designed to answer two questions: (a) Which reading technique, silent or oral, leads to better reading comprehension during administration of an individually-administered test?; and (b) which reading technique is more efficient?

Silent Versus Oral Reading Comprehension

243

Method Participants Originally 108 students from an elementary and middle school in a rural county in East Tennessee were administered the Test of Dyslexia (TOD; McCallum & Bell, 2001). Students from kindergarten through sixth grade were randomly selected to participate after receipt of parental permission. The silent versus oral procedure was balanced across participants. That is, participants were assigned to either the oral or the silent condition in an alternating fashion. In addition to the TOD, the 74 participants of this study were administered the Comprehensive Test of Basic Skills (CTBS McMillan/McGraw Hill, 1989). Ultimately, 36 males and 38 females participated in the study. Ages ranged from 6 years, 3 months through 13 years, 3 months with a mean age of 123.81 months, standard deviation was 18.18. Approximately equal numbers of girls and boys were assigned per grade level; there were no fewer than 12 participants per grade and no more than 16 per grade. Four children in the study were receiving special education services. Instrument The TOD is an individually administered test, comprised of 13 subtests and currently in development. The Reading Comprehension subtest is constructed in a format similar to an informal reading inventory and consists of reading sentences and/or passages of increasing difficulty, followed by two to five questions based on the readings. Some questions are literal, such that the answers are explicitly stated in the passage. Other questions are inferential, requiring the answers to be implied from what was read. The TOD is designed to be an ecologically valid measure of reading and related abilities; it measures reading comprehension in a manner similar to the various Brigance Inventories (e.g., Brigance Comprehensive Inventory of Basic Skills-Revised; Curriculum Associates, 1999) and the Qualitative Reading Inventory-3 (Leslie & Caldwell, 2001). Assessing reading comprehension in this manner is like what students actually do when asked to read in school. However, this method tends to be more time consuming than vocabulary or cloze procedures such as those used in the Woodcock-Johnson III (Woodcock, McGrew, & Mather, 2001). Pilot test data from the TOD (from the sample of 108 children) indicated adequate to strong reliabilities (Cronbach’s alphas range from .80 to .97 with the exception of one subtest, Word Memory, which was .68). Internal consistency for the Reading Comprehension subtest was strong (Cronbach’s alpha # .95). (Further information about the TOD can be obtained from the authors). Evidence for construct validity of the TOD is indicated by significant correlations between the TOD subtests and Reading subtest and Composite scores from the nationally normed CTBS-4. The CTBS is group administered; directions are read by the teacher and students read silently and independently to complete the reading tasks. The CTBS scores were available for children in grades 2 and higher. Correlations between the TOD academic achievement scores and CTBS Reading scores are moderate to strong, ranging from .36 to .64. Of particular relevance for this study, Reading Comprehension correlates .50 with CTBS Reading (measures basic understanding, ability to analyze text, evaluate and extend meaning, and identify reading strategies), .36 with CTBS Vocabulary (measures knowledge of word meanings, multimeaning words, and words in context), and .46 with the CTBS Reading Composite (composite of Reading and Reading Vocabulary subtests). Procedures Students were randomly assigned to treatment groups. All examiners were school psychologists or graduate students in a school psychology training program; they tested each student individually. Administration of the entire TOD took approximately one and one half hours per student;

244

McCallum, Sharp, Bell, and George

the Reading Comprehension subtest took only approximately 10 minutes. The test was administered during school hours at a time deemed most appropriate by the student’s teacher. Testing was conducted on school grounds in classrooms and/or offices according to privacy and availability. During the Reading Comprehension subtest, students were asked to read sentences and/or passages of increasing difficulty; about half of the students in each grade level read silently, while the other half read aloud. Six- and seven-year-olds were asked to begin by reading sentences, then answering questions based on the sentences. They continued by reading passages and answering questions based on what they read. Testing for the 8- through 13-year-olds began with passages. All subjects continued until they answered five questions in a row incorrectly. The number of correctly answered questions comprised the total comprehension score. Examinees who began with the passages and answered the first five correctly were automatically given credit for all items below that point. Those who did not answer the first five correctly were redirected to read the sentences to determine if they could answer those questions correctly. Students were also timed for each passage read, and a total time score was calculated by adding all times. Reading ability was operationalized by the participants’ Total Reading normal curve equivalent scores (NCEs) on the group administered CTBS. The CTB NCEs for this sample ranged from 1 to 99 with a mean of 53.05 and standard deviation of 21.27. Data Analyses Data were analyzed initially using a Multivariate Analysis of Covariance (MANCOVA), followed by two Analyses of Covariance (ANCOVA). Reading comprehension scores (TOD Reading Comprehension scores) and reading time (seconds required to complete passages) were the dependent measures for the MANCOVA. Because previous findings describing the influence of ability on silent versus oral reading comprehension are mixed and because of the relatively small sample size, reading ability was covaried. The significant MANCOVA F was followed by two ANCOVAs to evaluate independently the dependent measures, comprehension, and efficiency. Results Descriptive statistics are presented in Table 1. Specifically, means and standard deviations for comprehension and time under silent and oral conditions are reported. Results from the MANCOVA yielded a significant F value (Hotelling’s Trace). That is, the combined effect (i.e., comprehension and speed) differed under the two conditions of silent versus oral reading [F(2,70) # 3.20, p ! .05]. The covariate (reading ability) was also significant [F(2,70) # 15.41, p ! .01]. To determine whether the two dependent measures differed under the two conditions independently, two follow-up ANCOVAs were conducted with reading ability entered as the covariate. There was no statistical difference for comprehension [F(1,73) # .13, p " .05]. The

Table 1 Means and Standard Deviations for Reading Comprehension and Reading Time Under Two Reading Conditions: Silent Versus Oral Number of comprehension questions answered correctly

Oral Silent

Time in seconds required to read passages

N

M

SD

M

SD

35 39

26.31 25

6.85 7.22

336.34 257.25

156.29 125.67

Silent Versus Oral Reading Comprehension

245

mean number of comprehension questions answered correctly by children reading orally was 26.31; the mean for those reading silently was 25. However, the mean reading time for silent readers (M # 257.25) was significantly lower than for the mean reading time for oral readers (M # 336.34) [F(1,730 # 5.94,p ! .02]. In fact,when reading the same passages,students reading aloud took approximately 30% longer than those reading silently,on average. Discussion This study was designed to determine the superiority of oral versus silent reading,as operationalized by comprehension mastery and efficiency (speed of performance), under a unique condition, one-to-one test administration using an ecologically sensitive measure. Results support research indicating no significant differences between reading comprehension scores after reading orally versus reading silently (e.g., Holmes & Allison, 1985), although participants in the silent condition read more efficiently (i.e., faster). As previously noted, data obtained during instruction and/or from group administration have yielded mixed results. Some studies found higher reading comprehension scores following oral reading; specifically, the superiority of oral reading for younger and/or low ability readers was demonstrated by Fletcher and Pumfrey (1988); Fuchs and Maxwell (1988); Miller and Smith (1985); and Swalm (1972), although others failed to confirm these findings (e.g., Prior & Welling, 2001). In fact, some research supports the superiority of silent reading for poor readers (Miller & Smith, 1990). Differences in the literature may be related to different conditions under which data were obtained (e.g., individual vs. group administration, assessment vs. instruction) and/or characteristics of the sample (e.g., reading ability, age). Our results add to the literature by providing clear support for the efficiency of silent reading in the context of an individual assessment of reading comprehension. That is, even though silent reading did not lead to greater reading comprehension in the individual test administration format, it was more efficient than oral reading because it was faster. Because our results were obtained during one-to-one test administration and because both reading effectiveness (comprehension) and efficiency (time) were assessed with ability controlled, these results have critical implications, specifically for users and developers of individually administered tests containing a reading comprehension component. Busy school psychologists, educational diagnosticians, and teachers will likely choose the most efficient test administration format under certain conditions. That is, if there is no need to obtain data from the oral condition (e.g., to determine error patterns) and if the test manual does not specify, examinees should be instructed to read silently to save time. Individually administered achievement tests such as the Woodcock-Johnson III (Woodcock et al., 2001) are often used because of brevity. However, they have been criticized because of their simplicity and lack of ecological validity (Tomey, 2002). On the other hand, traditional informal reading inventories are more like reading tasks assigned in the classroom but they are lengthy to administer. Our results (also see Bell, McCallum, & Cox, 2003) show that a short but ecologically valid test can be efficiently administered by requiring examinees to read silently sentences and passages, then evaluating their comprehension. Conclusions and generalizations must be considered in light of the testing format (individual administration), the measure used to assess reading comprehension (informal reading inventory format), sample characteristics controlled (ability), and the age of our sample. Comprehension and efficiency may vary as a function of these variables. Results can be generalized to elementary grade students evaluated in a one-to-one assessment of reading comprehension using sentence and passage reading. Additional research is needed. In particular, the influence of ability on comprehension and efficiency should be manipulated systematically using larger samples and under varying conditions (e.g., group vs. individual administration).

246

McCallum, Sharp, Bell, and George References

Bell, S.M., McCallum, R.S., & Cox, E.A. (2003). Toward a research-based assessment of dyslexia: Using cognitive measures to identify reading disabilities. Journal of Learning Disabilities, 36(6), 505–516. Burge, P.D. (1983). Comprehension and rate: Oral vs. silent reading for low achievers. Reading Horizons, 23, 201–206. CTB MacMillan/McGraw-Hill. (1989). Comprehensive Tests of Basic Skills (4th ed.). Monterey, CA: Author. Curriculum Associates. (1999). Brigance Comprehensive Inventory of Basic Skills Revised. North Billerica, MA: Author. Elgart, D.B. (1978). Oral reading, silent reading, and listening comprehension: A comparative study. Journal of Reading Behavior, 10, 203–207. Fletcher, J., & Pumfrey, P.D. (1988). Differences in text comprehension amongst 7–8-year-old children. School Psychology International, 9, 133–145. Fuchs, L.S., & Maxwell, L. (1988). Interactive effects of reading mode, production format, and structural importance of text among LD pupils. Learning Disability Quarterly, 11, 97–105. Holmes, B.C., & Allison, R.W. (1985). The effect of four modes of reading on children’s comprehension. Reading Research and Instruction, 25, 9–20. Leslie, L., & Caldwell, J. (2001). Qualitative Reading Inventory-3. New York: Addison,Wesley Longman. McCallum, R.S., & Bell, S.M. (2001). The Test of Dyslexia. Unpublished test instrument, University of Tennessee, Knoxville. Miller, S.D., & Smith, D.E.P. (1985). Differences in literal and inferential comprehension after reading orally and silently. Journal of Educational Psychology, 77, 341–348. Miller, S.D., & Smith, D.E.P. (1990). Relations among oral reading, silent reading and listening comprehension of students at differing competency levels. Reading Research and Instruction, 29, 73–84. Mullikin, C.N., Henk, W.H., & Fortner, B.H. (1992). Effects of story versus play genres on the comprehension of high, average, and low-achieving junior high readers. Reading Psychology, 13, 273–290. Prior, S.M., & Welling, K.A. (2001). Read in your head: A Vygotskian analysis of the transition from oral to silent reading. Reading Psychology, 22, 1–15. Rowell, E.H. (1976). Do elementary students read better orally or silently? Reading Teacher, 29, 367–370. Swalm, J.E. (1972). A comparison of oral reading, silent reading and listening comprehension. Education, 92, 111–115. Tomey, H.A. (2002, November). Critical issues in the reauthorization of IDEA: IDA’s position. Paper presented to the 53rd annual convention of the International Dyslexia Association. Atlanta, GA. Woodcock, R.W., McGrew, K.S., & Mather, N. (2001). Woodcock-Johnson, Third Edition, Tests of Cognitive Abilities and Tests of Achievement. Chicago, IL: Riverside Publishing.