The Effects of Computer Instruction on College Students' Reading Skills.
Subject:
Computer-assisted instruction (Usage)
Reading (Study and teaching)
College students (Education)
Author:
Kuehner, Alison V.
Pub Date:
03/22/1999
Publication:
Name: Journal of College Reading and Learning Publisher: College Reading and Learning Association Audience: Academic Format: Magazine/Journal Subject: Education Copyright: COPYRIGHT 1999 College Reading and Learning Association ISSN: 1079-0195
Issue:
Date: Spring, 1999 Source Volume: 29 Source Issue: 2

Accession Number:
54805824
Full Text:
This paper reviews the research concerning computer-based reading instruction for college students. Most studies suggest that computers can provide motivating and efficient learning and help students improve their reading skills. However, it is not clear whether the computer or the instructing via the computer best accounts for student gains. Analysis also reveals many methodological flaws in the studies. The conclusion is that computer-based instruction can be effectively used with college students to improve reading skills, but that attention to the method rather than the means of instruction is most important.

The computer revolution has hit higher education as college students increasingly receive their instruction via computers. But along with the widespread use of computers in classrooms has come concerns about their effectiveness. As Tanner (1984) says, "When a new technology is touted as having so much potential for education, its glamour cannot be allowed to obscure the need to validate its usefulness" (p. 37).

Research on computer-based instruction at the college level has been published since the late 1960s and, in 1980, the first meta-analysis was conducted (a meta-analysis uses statistical analysis on the results of many different studies to generalize from the findings). From the perspective of Kulik, Kulik, and Cohen's 1980 meta-analysis of 59 studies, results for computer-based instruction in colleges look promising. Kulik et al. reported that "the computer has made a small but significant contribution to the effectiveness of college teaching" (p. 538) particularly in terms of student achievement and students' attitudes toward their instruction. An updated meta-analysis (Kulik & Kulik, 1986) reached similar conclusions about achievement and student attitudes. Interestingly enough, both the 1980 and 1986 meta-analyses reported almost identical effect sizes for student achievement, 0.25 and 0.26 respectively, meaning that the typical control student performed at the 50th percentile in comparison to the typical computer using student who performed at the 60th percentile. But the most striking finding of both meta-analyses was a "substantial savings in instructional time" (Kulik et al., 1980, p. 537) suggesting that computers could cut learning time to "two thirds the time required by conventional teaching methods" (Kulik & Kulik, 1986, p. 100). Still, Kulik et al., caution that the impact of computer-based instruction on student achievement at the college level is not as dramatic as gains at the elementary and secondary levels, and that other teaching methods might be just as effective as computers. These meta-analyses provide only a broad overview of the efficacy of computer-based instruction in college classrooms across many disciplines, and, as will be discussed later, may be flawed in their conclusions.

Literature reviews focusing on computers to teach reading are more cautious in their findings. Balajthy (1987), for instance, believes that the "results of research on computer-based instruction in reading are at best equivocal" (p. 63). He explains that while using computer-based instruction to supplement traditional instruction is effective, so is almost any type of supplemental instruction, whether or not it uses computers. In a recent review examining reading achievement in adult education, Rachal (1995) found "no significant differences between computer-assisted instruction (CAI) and traditional reading instruction" (p. 249) for adults reading at the ninth grade level or above.

To date, no meta-analysis or research review has looked specifically at the research on the use of computers to teach college-level reading skills. With this in mind, the goal of this paper is to examine this body of research. First, the Educational Resources Information Center (ERIC) database was searched, using various combinations of the following keywords: CIA or computer-based instruction (CBI) and reading and college or college student(s). Bibliographies of articles and reports found through the ERIC search were also used to locate other relevant studies. Studies were chosen to review on the basis that they described experimental or quasi-experimental studies involving college students, at two or four-year institutions, using computers for reading instruction--either reading instruction per se or reading for study in a content area class. Only studies published since 1980 were included since software and hardware prior to 1980 is largely outdated. Also, repeat studies using the same software and conducted by the same researchers were not included, just the most recent version of the research was examined.

Expectations for Computer-Based Instruction

Much of the excitement over computer instruction is fueled by the belief that computers can help college students, sometimes in ways human teachers or tutors cannot. One of the most powerful arguments advanced for computer-based instruction is that computers can individualize instruction (Askov & Clark, 1991; Kamil, 1987; Reinking, 1987; Seigel & Davis, 1987; Turner, 1988; Watkins, 1991). Computers should be capable of adjusting the content of lessons or the rate of instruction according to the learners' needs. For example, Watkins reports on one student who needed 8 hours to complete an assignment, a pace that would exhaust the patience of most human tutors, but not an electronic one. Moreover, Taraban (1996) describes a program that could advise students as to which reading strategies work best for Them. The computer monitors the student's reading behaviors and performances on exercises or quizzes, accumulates sufficient information to make correlations between the student's reading activities and performance scores, then recommends strategies that have helped the student in the past.

Computers could place in the readers' hands more control over learning, which Askov and Clark (1991) and Turner (1988) contend is empowering for low-level readers. Now with long-distance capabilities for delivering information, computers can be flexible to meet students' needs; computer instruction could conceivably be administered at various sites at any time of the day or night (Turner). Such accessibility has been recognized as advantageous for students who must juggle multiple responsibilities (Askov & Clark; Turner).

Other hypothetical advantages to computer instruction have less to do with the nature of instruction than with students' attitudes toward their learning. Particularly for the struggling student, observers report computer instruction can provide an important degree of privacy; only the computer program knows the student's skill level (Askov & Clark, 1991; Turner, 1988; Watkins, 1991). In fact, Askov and Clark contend that computer instruction provides a certain cachet. Students who might be embarrassed to admit they are taking a remedial reading class can avoid stigma by saying they are attending a computer class. Finally, researchers expect to capitalize on the excitement surrounding new technologies to motivate students to learn. In the age of television and video games, perhaps computers can engage students with graphics, animation, and game-like features in ways that will make learning fun (Kamil, 1987).

In short, optimistic educators believe computers will be patient, responsive, personalized tutors providing extra help with assignments in ways that engage and encourage learners. However, where there is hope, there is also fear, the fear that computer-based instruction will prove to be an expensive, ineffective attempt to improve learning. Askov and Clark (1991), among others, point to the high costs of installing hardware and purchasing software, of maintaining and upgrading equipment, and of providing computer training for teachers, as well as expert, technical support in the classroom or lab. Computers can crash (Kamil, 1987) and software, once installed, can be inflexible and may not exactly suit the needs of the course or the students (Kamil; Watkins, 1991). Moreover, integrating computers into the curriculum takes teacher time and energy, extra work which might cause resistance or create resentment among faculty (Askov & Clark).

Reinking (1988-89) offers some more troubling criticisms of reading software, arguing that most are neither pedagogically sound nor based on current research about the teaching of reading. Rather, he claims, assumptions underlying reading software development are fundamentally flawed. For instance, programmers rely on the misconception that reading is best taught by focusing on isolated skills, rather than on integrating these skills into the act of reading. Often these programs ignore the process of good readers to emphasize products, such as correct responses to multiple choice questions, and therefore do not teach reading comprehension so much as measure it. In short, bad software equates with poor learning. Computers, some observers warn, may not be the panacea for education.

Research-Based Answers: Computer vs. Traditional Instruction

To test hypothetical or observed advantages and drawbacks of computer-based instruction, we can turn to research. Typical research studies pit computer-based instruction against traditional teaching methods. Students in a control group might complete assignments by filling in worksheets or reading printed texts, while students in an experimental group might complete the same assignments using a computer program or by reading texts on a computer screen. An early concern of researchers was whether reading the same text in print or on a computer has any impact on reading ability. Other important research questions consider whether computer-based instruction has any effect on attitudes, learning time, or student achievement.

Mode of Delivery

Studies that examine mode of delivery investigate whether the experience of reading material on a printed page differs from reading on a computer screen. This is important because computer-based instruction, almost by definition, requires students to read significant amounts of information on a computer screen. Despite some concerns that computers might impede reading ability because of eye strain, or slow readers down, there seems to be little difference between these two reading methods. Fish and Feldmann (1987) found no significant comprehension differences among sophisticated readers (graduate students) when reading the same text on page or screen. Similarly, Askwall (1985) reported the same text presented on a computer or on paper had no effect on undergraduates' reading speed or comprehension. Therefore no detectable differences seem to exist between reading information from print or from a computer screen.

Attitudes

Most researchers agree that students have a positive attitude toward learning on computers (Balajthy, 1988; Kulik & Kulik, 1986; Kulik, et al., 1980; Lang & Brackett, 1985; McCreary & Maginnis, 1989; Mikulecky, Clark, & Adams, 1989; Wepner, Feeley, & Wilde, 1989). Only one study (Wepner, Feeley, & Minery, 1990) reported negative student reactions toward computer-based instruction, which the researchers attributed to "poor lab conditions" (overcrowded lab with outdated, unreliable hardware and software) and an "unfortunate change in instructors midway through the course" (p. 353). More typical are Mikulecky, et al.'s, findings that students' attitudes toward computer-assisted instruction were strongly positive. Students reported on questionnaires that they enjoyed using the computer lessons and learned from them. In this case, the researchers maintain that students recognized the computer taught them useful reading strategies.

This positive student attitude, however, can be problematic if students confuse interest with effectiveness, cautions Balajthy (1988). That is, students in Balajthy's (1988) study rated the 2 computer-based instructional methods as being more effective than traditional workbook exercises, when, in fact, the group using workbooks showed greater achievement gains. Balajthy conjectures that students equate their interest in computer-based instruction with learning effectiveness and therefore may not, if left to their own devices, be capable of choosing the mode of instruction that would be most helpful to them. So while computers seem to motivate learning, this same motivation may misdirect students' attention toward unproductive activities and therefore not pay off in achievement gains.

Learning Time

Computer-based instruction reduces learning time (Kulik & Kulik, 1986; Kulik, et al., 1980; Wepner et al. 1990; Wepner et al., 1989). Wepner et al. (1990) found students using computers could complete an entire program in the same time it took the control group to get through two thirds of the same material. The researchers noted that this 32% reduction in instructional time "correspond [s] precisely" (p. 352) with Kulik and colleagues (1986, 1991, 1980) findings in their meta-analyses. Wepner et al. (1989) report a similar result in an earlier study, noting that the computer's ability to efficiently manage instruction (in this case, to calculate words per minute read and comprehension scores, and to supply reading materials) saved time since the computer users "consistently finished before the allotted time while the control group sometimes had to do their paperwork after class" (p. 8). They hypothesized that this time-saving feature of the computer may account for students' positive reactions toward computer-based instruction.

The one study to contradict these findings is Balajthy's (1988) comparison of students who used traditional workbook exercises vs. 2 groups who used 2 different computer programs to study vocabulary. As noted previously, the workbook users outperformed the computer users on vocabulary quizzes, yet these students also spent significantly less time on the text exercises. These findings suggest the workbook exercises were the most efficient use of student time (students learned the most in the least amount of time), even though students rated this method of instruction as least interesting and least effective. The fact that students spent less time on the workbook could be explained by their low-interest ratings, yet it is not clear why these exercises also proved to be more effective learning tools.

Achievement

An important concern for researchers has been whether computer-based instruction improves student achievement, achievement most often measured in quantifiable terms such as the differences between scores on pre- and post-tests. Many individual studies focusing on reading skills have found computer-based instruction to be effective for improving reading comprehension (Dixon, 1993; Grabe, Petros, & Sawler, 1989; Kester, 1982; Lang & Brackett, 1985; Mikulecky, et al., 1989; Price & Murvin, 1992; Skinner, 1990; Wepner et al., 1990). Vocabulary (Culver, 1991; Lang & Brackett, 1985) and reading rate (Culver; Wepner et al., 1990) also appear to benefit from computer-based instruction. A closer look at these studies, though, raises the question of whether the computer or the instruction via the computer made the difference in student learning.

For example, studies often compare one group that receives computer-based instruction to a control group that receives no special instruction to show that computer-based instruction is effective. Price and Murvin (1992) reported that a computer program supplementing the textbook in an accounting class boosted student success rates in the course when compared to students in previous classes who had no access to the computer-assisted instruction. Similarly, Grabe et al. (1989) found that students in an educational psychology class scored better on exams when they used computer-assisted instruction to study the assigned textbook reading as opposed to studying the textbook on their own. Moreover, Mikulecky et al. (1989) looked at undergraduates in a biology class who used a computer program .to help them understand the reading material as compared to a control group who studied the textbook on their own. The computer group scored significantly higher on exams, and even on subsequent exams, suggesting the computer had modeled and taught students effective reading strategies. But while both treatment and control groups worked with the same textbook for the same amount of time, only the treatment group received instruction (through the computer program) about how to identify, compare, contrast, and connect key concepts in the reading, skills that were necessary to do well on the test.

These studies seem to indicate that computer-based instruction can provide effective supplemental instruction. Another example is Kester's (1982) study in which students in basic skills classes who used computer-assisted instruction at least 2 hours a week to supplement their regular classwork made significantly greater gains in reading skills than students who did not engage in supplemental instruction. Dixon (1993) found that students completing a required remedial reading course averaged 4 years of growth in reading comprehension in the first study and 3 years of growth in a repeat study, leading Dixon to conclude that computer-assisted instruction is effective for remediating under-prepared freshmen.

It might also be that computer-based instruction, when compared to traditional instruction, provides a different type of instructional experience. Skinner (1990) compared 2 groups of students using 2 different versions of the same computer program to a control group who used text-only materials to study for a classroom management class. The computer groups performed consistently better on quizzes than the text-only group. Skinner hypothesized that the computer programs were effective study tools because they gave the students immediate feedback and were motivating. But also, students working under computer instruction were required to complete tutorial units while the text group had no such requirement.

Like the studies that indicate computer-assisted instruction helps college students' reading comprehension, studies suggest computer-based instruction can also improve students' vocabulary. In Lang and Brackett's (1985) research, college freshmen using computers to learn vocabulary and comprehension skills showed gains of one to two years in grade level reading ability over the course of the semester. This study, however, lacked a control group. Culver (1991) reports that computer instruction can improve English as a Second Language (ESL) students' vocabulary. Over the course of a semester, the researcher noted an overall increase of 3.9 grade levels in vocabulary development for students using a computerized, levelized reading program. But, like Lang and Brackett's study, Culver's study lacked a control group against which to compare these gains.

Computer-based instruction also seems to improve students' reading rates. Wepner et al. (1990) concluded that reading rate for developmental reading students improved using computer-assisted instruction. Growth for the computer users compared to the central group was statistically significant. In this study, however, students in the computer group were able to finish all the assigned units while the control group completed only two thirds of the similar text materials. Culver (1991), too, found reading rate improvements in ESL students using computer-based instruction in a developmental reading class. The majority of students improved their reading speed, with an average 3.4 grade level increase for the semester. But, as mentioned previously, this study lacked a control group.

In most of the above cited studies, computer-based instruction did improve students' reading skills. However, attributing achievement gains to the computer alone may be misleading, since the computer often provided additional or different instruction that the control groups did not receive.

On the other hand, some studies have found computer-based instruction has little or no effect on reading skills (Burke & others, 1992; Jobst & McNinch, 1994; Kleinmann, 1987), comprehension and vocabulary (McCreary & Maginnis, 1989; Taylor & Rosecrans, 1986), or efficiency (Wepner et al., 1989). For example, in a study much like many of those cited above, Burke and others (1992) placed students into practice labs to study, either with a computer-based approach or with a text-based approach. The researchers found no significant difference in the achievement of the 2 groups. However, when compared to a group who did not use a practice lab, the 2 groups who participated in practice labs, whether computer or text-based, scored significantly higher on a standardized reading test. This led Burke and others to conclude that the amount of practice time, not the mode of presentation, best accounts for differences in student achievement. Conversely, Kleinmann was careful to set up a study that used identical text and computer-assisted instructional materials and equal instructional time. He found that both groups made significant gains in reading achievement, and no significant difference in gains existed between the groups. Kleinmann concluded that while supplemental instruction appears to be effective for ESL students in a developmental reading program, supplemental computer-based instruction does not seem to be any more effective than supplemental traditional instruction. In fact, Burke and others and Kleinmann's studies, which directly addressed the question of whether achievement gains are due to more practice time or to computer-based instruction, suggest the answer lies in additional instructional time.

Taking a different approach, Jobst and McNinch (1994) set up a computer-based and text-based reading assignment for students in a technical writing class. Rather than create identical study materials (a case study), they deliberately constructed materials that would take advantage of each method: The printed version was cheap and easy to use; the computer version allowed for graphics and student choice about moving around in the text. Despite the researchers' expectations of increased achievement among the computer users, no significant differences were found in retention of the material or in students' exam scores. This study raised concern that the time involved in developing computer-assisted tutorials did not pay off in student results.

Analyzing student achievement when students use computers or texts is not a simple process. Some studies of computer-based instruction reveal that student aptitude might influence achievement. Two studies (Price & Murvin, 1992; Skinner, 1990) suggest that poor readers can benefit more than capable readers from computer use. Price and Murvin, who added supplemental computer-based instruction to an accounting class, reported the results of their colleague's research that students with reading skills below college level stayed in their accounting class and succeeded at higher rates than previous students. Students with college level reading skills also benefited from the computer instruction, but not as dramatically. Similarly, Skinner concluded that computer-based instruction is effective for college students, but particularly for those with a record of poor past performance. Low-achieving students using computer-based instruction scored 15% higher on quizzes than low-achieving students using text-based study materials. Skinner hypothesizes this improvement is due to "the structure and frequent opportunities to respond provided by CBI" (p. 358). Indeed, as noted earlier, students working under computer instruction were required to complete tutorial units while the text group had access to a human tutor with no required tutorials. In these two studies, the use of computer-based instruction to require supplemental work may be one reason less able students improved under the computer treatment when compared to students who studied on their own.

In contrast, the study by Grabe et al. (1989) illustrates a more problematic interchange between student aptitude and instructional effectiveness. In their experiment, when students were given free access to computer-assisted tutorials for study, the better students tended to use the computers. These computer users outperformed their classmates on most exams (even taking into account the fact they were better readers). Despite these advantages to computer users, the number of students using computer-assisted instruction over the course of the semester declined drastically. Researchers were not sure why this was so or why less capable students made less use of computer-assisted tutorials even though such instruction might benefit them.

Based on the above cited studies of college students, it appears that computer-based instruction can improve students' reading abilities. The majority of studies indicate that computer-assisted instruction increases student achievement. This finding, though, might be because computer-based instruction supplements or adds new instruction not provided to those students using "traditional methods." In these cases, more instruction, different instruction, or more time on task may account for the gains by computer users. Other factors, such as student ability, may further influence achievement gains of computer users. The structured approach of some computer-based instruction may help less able students who are unable to study effectively on their own.

Research-Based Answers: Computer vs. Computer Studies

Although most studies compare computer-based to text-based instruction, a few researchers (Balajthy, 1988; Blohm, 1987; Gay, 1986; Kulik & Kulik, 1986; Skinner, 1990; Taylor & Rosencrans, 1986) have examined differences among various computerized instructional methods. A key question is whether the method of computer instruction affects student achievement.

Unfortunately, no consistent terminology describes features of computer-based instruction. It is not clear, for instance, what makes a program "interactive." Nevertheless, Kulik and Kulik (1986) established three main categories of computer instruction: (a) computer-assisted, providing drill and practice or tutorial instruction; (b) computer-manage & providing evaluation, feedback, guidance and record keeping for the student; and (c) computer-enriched, serving as a tool to solve problems or as a model to illustrate ideas or relationships. Kulik and Kulik's (1986, 1991) meta-analyses found no difference in achievement among these instructional methods. All types of computer-based instruction had small but positive effects on student learning. They concluded that college students can readily adapt to a variety of computer-based instructional methods.

Another area of considerable interest is learner control vs. program control in computer-based instruction. In fact, learner control has become a field of study in itself and will only be touched on in this paper in the context of reading instruction. In learner control situations, subjects typically make decisions about how the computer program operates; for instance, they may decide whether or not to preview or review material, complete practice exercises, or do extra work if they receive low scores. In program control, the computer guides the learner's course through the program and usually "makes decisions" about whether to review material or to assign extra exercises for the learner.

Two studies (Balajthy, 1988; Gay, 1986) caution against giving poor readers significant control over their learning, whether using computer-assisted or traditional methods. In these studies, learner control hurt low-aptitude students who lacked effective learning and reading strategies. As both researchers explain, these students are unable to accurately monitor the success or failure of their own learning. In Gay's study, when students were given control over the computer program to study modules on DNA structure, subjects avoided difficult or unfamiliar material and tended to overstudy familiar topics. On the other hand, subjects with high prior knowledge of the topic under learner control conditions were significantly more efficient in their use of time than subjects with high prior knowledge under program control or than low prior knowledge subjects under learner control or program control. Blohm (1987) also found that providing proficient readers with learner control (in this case, computerized access to lookup aids, such as clarification of technical language) improved their reading comprehension. That is, competent readers successfully monitored their own comprehension and took advantage of computerized tools when compared to competent readers reading the same material via computer with no lookup aids. Both Gay and Blohm studies suggest that students can be given more learner control if their prior understanding of a topic is relatively high.

In one of the few studies comparing 2 different methods of computer-assisted instruction to traditional instruction, Skinner (1990) allowed students to use the same computer program under guided (GUIDED) or unguided (SOLO) conditions. Under the SOLO method, students were able to choose which computerized tutorials to complete, while under the GUIDED method students were required to complete entire units of tutorials. The control group used text-only tutorials. As mentioned earlier, Skinner found that low-achieving students benefited significantly from computer-based instruction. The study also revealed that students seemed to prefer the guided method of computer instruction. That is, even when students in the SOLO group could operate the program as they wished, most treated it like a GUIDED program. This finding might explain the higher levels of achievement for both the computer users and for the low-ability students, since other research suggests program control benefits less capable students.

Skinner's (1990) results, though, are contradicted by other studies. Taylor and Rosecrans (1986) also examined a control group (non-computer users) and 2 different computer-assisted treatments, being students receiving computer-assisted instruction in a structured manner and students using computer-assisted instruction during their free time (unstructured). In this study, the control group outperformed the 2 experimental groups. In another three-way study, Balajthy (1988) compared students using traditional workbook exercises to students using 2 different computer programs; a fast-paced video game and a slow-moving text exercise. In this case, the workbook users outscored the 2 groups of computer users. As in Skinner's study, students seemed motivated by the computer-based instruction (rating it as highly effective) and spent more time using the computers. But, unlike Skinner's subjects, the students in Balajthy's (1988) study did not benefit as much from the computer-based instruction as did their counterparts who studied with text workbooks.

A tentative conclusion might be that interactive yet guided practice (as advocated by Burke and others, 1992) is a beneficial approach for computerized remedial reading instruction, and, for these students, better than unaided homework. Students with more prior knowledge about a topic (Gay, 1986), or with good reading skills (Blohm, 1987) may benefit from more control over their own learning (Grabe et al., 1989). Still more research is needed to sort out the various influences of the type of computer program or instructional method and of the characteristics of the learner on achievement.

Criticisms of the Research

Although research studies should provide a more reliable, objective assessment of computer-based instruction than anecdotal or hypothetical observations, research has its limits and problems.

As noted earlier, the type and quality of computer programs vary greatly. Balajthy (1987) contends that "a variety of observers have indicated that the computer is not being well-used in the field of education" (p. 56). He also suggests there is a "lack of quality software" (p. 57), a point Reinking (1988-89) supports when he argues that most reading software is neither pedagogically sound nor based on current research about the teaching of reading. Balajthy (1987) points out that "almost all computer-based research is based on the programmed instruction model, which ... is presently out of favor among reading researchers and teachers" (p. 56). In these situations, the software or hardware limitations may also limit the research findings. If computers are not being effectively used to teach, then researchers will not see the results of good computer-based instruction.

More troublesome, though, are claims of flawed research studies and meta-analyses. In an examination of the meta-analyses done on computer-based instruction by Kulik and colleagues, Clark (1986) claims 75% of the studies used were poorly designed (based on a random sampling of 30% of the studies included in the meta-analyses). He also notes that over 50% of the studies he examined failed to control the amount of instruction each group received, so that more instructional time might account for the increased learning of the computer users. Moreover, Clark points to studies in which the method of instruction differed between experimental and control groups. In these studies, the type of instruction, rather than the computer, may account for any measured effect. Reinking and Bridwell-Bowles (1991) also contend that many computer-based studies fail to properly control variables, such as time on task. Again, if the computer group is spending more time studying than the control group, this extra time, rather than the computer, might account for differences between the 2 groups.

A recurring criticism of research design is failure to control for the Hawthorne effect that tends to operate on the experimental group (Balajthy, 1987; Clark, 1986; Reinking et al., 1991). The novelty of using computers, explains Balajthy (1987), might result in increased student effort to learn. Evidence for this effect is bolstered by the finding in the most recent meta-analysis (Kulik & Kulik, 1986) that computer-based instruction is more effective over short periods of time (less than 4 weeks) and effectiveness decreases over longer periods. Perhaps, after 3 or 4 weeks, the novelty of using a computer wears off. On the other hand, shorter studies might be more tightly controlled and therefore better able to measure significant differences between the 2 groups (Kulik & Kulik, 1986). Another concern is lack of control over the "same teacher" effect. That is, if different instructors design the curriculum and/or teach the control group and the experimental group, differences in achievement might be attributed to the instructor rather than to the method of instruction. As evidence of this problem, Clark notes that when the same teacher designs both the computer and traditional instruction, computer-based effect sizes for college students reduce to insignificant levels.

In fact, when Clark (1986) re-analyzed the studies, controlling for such variables as the "same teacher" effect or instructional methods, his revised effect sizes were much lower. He concludes that meta-analyses overestimate the effect of computer-based instruction on achievement. It would appear that many of Clark's criticisms of the meta-analyses apply to the studies examined in this paper. Differences in instructional methods or time on task between control and experimental groups may account for the differences between groups.

The Correct Question?

In light of the criticisms over the research and meta-analyses, Clark (1986) contends that it is basically misleading or unproductive to pit computer-based instruction against traditional teaching methods. When studies are correctly designed, Clark asserts, no discernible differences in student achievement exist that can be attributed to computers, and there is no reason to believe that there should be. The computer, he argues, is just a delivery system for instruction. The type of instruction, rather than the means by which it is sent to the student, is paramount. Therefore, the correct question researchers could investigate productively would be how computers might deliver good instruction most effectively and cost-efficiently.

Balajthy (1987) also believes research into computer-based instruction could be more productive by focusing on the question, "In what ways can the computer improve on conventional classroom effectiveness and efficiency?" (p. 55). Unlike Clark (1986) though, Balajthy (1987) insists that research and research reviews show "there is no doubt that computer instruction is effective" (p. 55). Like Clark, Balajthy (1987) emphasizes identifying effective teaching methods, then considering how computers can effectively deliver that instruction. In this process, Balajthy (1987) advocates examining the various student-based, computer-based, or instructional factors that influence computer effectiveness.

Conclusions

As Tanner (1984) urges, we as educators should not allow the excitement of computers arriving in our classrooms and labs to blind us to the need to examine carefully how we use computers with our students. It would seem, based on the examination of research included in this review, that both Clark's (1986) and Balajthy's (1987) emphasis on good instruction via computers, rather than on computer instruction itself is important. As educational researchers, we might do well to heed Rachel's (1995) criticisms of computer-based research, noting how frequently studies lacked control over treatment time, did not randomly assign subjects to groups, or used a small number of subjects for study. His suggestions to future researchers are excellent: pre-testing and randomly assigning an adequate number of students to control and experimental groups taught by equally competent instructors; carefully documenting time on task and post-testing students after equal number of hours for each treatment; using appropriate software; and reporting methodology and findings as clearly as possible.

Despite problems with the research, in light of the studies cited here, there are some good reasons to use computers for reading instruction with college students. Computer-based instruction can provide motivating and efficient learning since two of the most significant advantages to using computers are that students have positive attitudes toward learning with computers and that computers, in most situations, can reduce instructional time.

Moreover, computer-based instruction as a supplement to traditional teaching methods appears to increase student achievement, though it is not clear whether computer-based instruction itself or the instruction given students via the computer best accounts for student gains. This uncertainty suggests that teachers need to consider carefully the computer program itself. Instruction should be based on sound pedagogy. In fact, supplemental materials or additional instruction may be provided to students to improve reading skills without computer aid. Here, some evidence indicates that remedial students can benefit more from computer instruction, but only if they are not given a significant degree of control over their own learning. For these students, program control or explicitly taught learning strategies might be more advantageous. Certainly there is promise that computers can help teach large numbers of college students, including remedial readers, but only if they are used wisely.

References

Askov, E. N., & Clark, C. J. (1991). Using computers in adult literacy instruction. Journal of Reading, 34, 434-437.

Askwall, S. (1985). Computer supported reading vs. reading text on paper: A comparison of two reading situations. International Journal of Man-Machine Studies, 425-439.

Balajthy, E. (1987). What does research on computer-based instruction have to say to the reading teacher? Reading Research and Instruction, 27 (1) 55-65.

Balajthy, E. (1988). An investigation of learner-control variables in vocabulary learning using traditional instruction and two forms of computer-based instruction. Reading Research and Instruction, 27 (4), 15-24.

Blohm, P. J. (1987). Effect of lookup aids on mature readers' recall of technical text. Reading Research and Instruction, 26 (2), 77-88.

Burke, M., & others. (1992, March). Computer-assisted vs. text-based practice: Which method is more effective? A version of this paper was presented at the Annual Midwest Reading and Study Skills Conference, Kansas City, MO. (ERIC Document Reproduction Service No. Ed 350 046)

Clark, R. E. (1986). Instruction studies: Analyzing the meta-analyses. Educational Communication and Technology, 33, 249-262.

Culver, L. (1991). Improving reading speed and comprehension for ESL students with the computer (Practicum Report No. FLO19486). Florida, Nova University. (ERIC Document Reproduction Service No. Ed 335 960).

Dixon, R. A. (1993, March). Improved reading comprehension: A key to university retention? Paper presented at the Annual Midwest Regional Reading

and Study Skills Conference, Kansas City, MO.

Fish, M. C., & Feldmann, S. C. (1987). A comparison of reading comprehension using print and microcomputer presentation. Journal of Computer-Based Instruction, 14 (2), 57-61.

Gay, G. (1986). Interaction of learner control and prior understanding in computer-assisted video instruction. Journal of Educational Psychology, 78, 225-227.

Grabe, M., Petros, T., & Sawler, B. (1989). An evaluation of computer assisted study in controlled and free access settings. Journal of Computer-Based Instruction, 16 (3), 110-116.

Jobst, J.W., & McNinch, T.L. (1994). The effectiveness of two case study versions: Printed versus computer-assisted instruction. Journal of Technical Writing and Communication, 24, 421-433.

Kamil, M. L. (1987). Computers and reading research. In D. Reinking (Ed.), Reading and computers: Issues for theory and practice (pp. 57-75). New York: Teachers College Press.

Kester, D. L. (1982, August). Is micro-computer assisted basic skills instruction good for black, disadvantaged community college students from Watts and similar communities? Paper presented at the International School Psychology Colloquium, Stockholm, Sweden.

Kleinmann, H.H. (1987). The effect of computer-assisted instruction on ESL reading achievement. The Modern Language Journal, 71, 267-273.

Kulik, C.C. & Kulik, J.A. (1986). Effectiveness of computer-based education in colleges. AEDS Journal, 19 (2-3), 81-108.

Kulik, C.C. & Kulik, J.A. (1991). Effectiveness of computer-based instruction: An updated analysis. Computers in Human Behavior, 7, 75-94.

Kulik, J. A., Kulik, C. C., & Cohen, P. A. (1980). Effectiveness of computer-based college teaching: A meta-analysis of findings. Review of Educational Research, 50, 525-544.

Lang, W. S., & Brackett, E. J. (1985, February). Effects on reading achievement in developmental education: Computer-assisted instruction and the college student. Paper presented at the Annual Meeting of the South Carolina Reading Association, Columbia, SC.

McCreary, R., & Maginnis, G. (1989, May). The effects of computer-assisted instruction on reading achievement for college freshman. Paper presented at the Annual Meeting of the International Reading Association, New Orleans, LA.

Mikulecky, L., Clark, E.S., & Adams, S. M. (1989). Teaching concept mapping and university level study strategies using computers. Journal of Reading, 32, 694-702.

Price, R. L., & Murvin, H. J. (1992). Computers can help student retention in introductory college accounting. Business Education Forum, 47, 25-27.

Rachal, J.R. (1995). Adult reading achievement comparing computer-assisted and traditional approaches: A comprehensive review of the experimental literature. Reading Research and Instruction, 34 (3), 239-258.

Reinking, D. (1987). Computers, reading, and a new technology of print. In D. Reinking (Ed.), Reading and computers: Issues for theory and practice (pp. 3-23). New York: Teachers College Press.

Reinking, D. (1988-89). Misconceptions about reading and software development. Computing Teacher, 16 (4), 27-29.

Reinking, D., & Bridwell-Bowles, L. (1991). Computers in reading and writing. In Barr, R., Kamil, M. L., Mosenthal, E B., & Pearson, ED. (Eds.) Handbook of reading research (pp. 310-340), New York: Longman.

Seigel, M. A., & Davis, D. M. (1987). Redefining a basic CAI technique to teach reading comprehension. In D. Reinking (Ed.), Reading and computers: Issues for theory and practice (pp. 111-126). New York: Teachers College Press.

Skinner, M. E. (1990). The effects of computer-based instruction on the achievement of college students as a function of achievement status and mode of presentation. Computers in Human Behavior, 6, 351-60.

Tanner, D. E. (1984). Horses, carts, and computers in reading: A review of research. Computers, Reading, and Language Arts, 2, 35-38:

Taraban, R. (1996) A computer-based paradigm for developmental research and instruction. Journal of Developmental Education, 20 (11), 12-14, 16, 18, 20.

Taylor, V. B., & Rosecrans, D. (1986). An investigation of vocabulary development via computer-assisted instruction (CAI). (ERIC Document Reproduction Service No. ED 281 168).

Turner, T. C. (1988). Using the computer for adult literacy instruction. Journal of Reading, 31, 643-647.

Watkins, B. T. (1991). Using computers to teach basic skills. Chronicle of Higher Education, 38 (6) A23-26.

Wepner, S.B., Feeley, J. T., & Minery, B. (1990). Do computers have a place in college reading courses? Journal of Reading, 33, 348-354.

Wepner, S. B., Feeley, J. T., & Wilde, S. (1989). Using computers in college reading courses? Journal of Developmental Education, 13, 6-8, 24.

Alison Kuehner has been teaching English at Ohlone College in Fremont, California for ten years. Last year, while on sabbatical leave earning a Master's Degree in Reading Instruction, she had time to research this article.
Gale Copyright:
Copyright 1999 Gale, Cengage Learning. All rights reserved.