Printer Friendly

Teacher interactions in mainstream social studies and science classes.

* In response to continuing critiques regarding the efficacy of separate instruction for students with mild academic problems (Dunn, 1968; Reynolds, Wang, & Walberg, 1987), there has been increasing interest in assessing whether the general educational environment is, or can become, sufficiently responsive to the instructional needs of mainstreamed students (Madden & Slavin, 1983). Though debating mainstreaming as a concept is still popular among many educators, there remains a relative paucity of research on the effectiveness of teachers in mainstream classes. Drawing on the work of researchers during the 1970s and early 1980s, some educators report that regular classroom teachers are unresponsive to students who do not learn easily. For example, Eder (1988) noted that teachers assigned low-achieving student to groups and tasks where academic expectations are minimized. Thus, students with the greatest academic needs obtained the least academic resources. Bryan (1978) reported similar findings. Her work has shown that students with learning disabilities (a) receive twice as much negative feedback and (b) are twice as likely to be ignored by regular class teachers as students without learning disabilities. Similarly, a recent study has shown that although teachers in regular classes allocate more instructional time than do teachers in separate special education classes, student with mild disabilities may be found on task more often in resource room settings (Rich & Ross, 1989).

Any evidence of instructional inequity within a class or across classes is disturbing; but these findings have been somewhat mitigated by research that identifies instructional similarities. Teacher feedback, for example, particularly instructional praise, is often inadequate or missing in both regular (Strain, Lambert, Kerr, Stagg, & Lenkner, 1983) and special education classes (Algozzine, Morsink, & Algozzine, 1988; Gable, Hendrickson, Young, Shores, & Stowitschek, 1983). Other studies have shown that substantial time allocated for instruction and high-impact teaching interactions may be lacking for students with learning disabilities in both special and regular classes (Baker & Zigmond, 1990; Zigmond & Baker, 1990). Whereas students appear to adjust well to returning to the group instruction formats in regular classes, they may not make any more (or less) educational progress there (Zigmond & Baker, 1990). This is noteworthy because many practicing teachers do not use instructional models or questions adequately until they are trained to do so (Hendrickson, Roberts, & Shores, 1978; Hendrickson & Stowitschek, 1980). Finally, various researchers have found that neither mainstreamed students nor students without disabilities receive adequate time for academic instruction (Ivarie, Hogue, & Brulle, 1984; Thompson, White & Morgan, 1982).

If these instructional deficits affect children similarly, it is important to note that high-quality instruction has a similar impact (Stevens & Rosenshine, 1981). Larrivee (1986), for example, has noted that mainstreamed students benefit from the same "effective teaching" behaviors that are delivered to regular education students. Affleck, Madge, Adams, and Lowenbraun (1988) have found that enhanced, integrated classes are of at least equivalent value to (a) youngsters with mild disabilities in resource rooms and (b) students without disabilities in regular classes. Similarly, Wang, Peverly, and Randolph (1984) have noted improvements in regular education students' performance when effective teaching practices were implemented for mainstreamed students.

If, indeed, improvements are needed in the general educational environment, then it is important to assess the impact of those improvements on all learners. In our previous mainstreaming research, we found that following a relatively modest inservice package of effective teaching practices, participating middle school social studies teachers improved their instructional interactions with all students (Brady, Swank, Taylor, & Freiberg, 1988; Swank, Taylor, Brady, & Freiberg, 1989). Of particular interest was (a) the experimental teachers' increased use of academic questions and reinforcement and (b) increases in guidance and information received by mainstream students. However, results in that first study demonstrated that classroom type (i.e., whether the classes were homogeneous or heterogeneous) and student type (i.e., whether the students were mainstream, low achievers, or mixed achievers) had affects as important as the inservice.

The current study was conducted as an extension and replication of our earlier investigation on instructional interactions in mainstream classes. As Good and Brophy (1987) have pointed out, instructional adaptations must be related to subject matter and students, as well as other contextual variables. The differential approaches found for science and social studies instruction, for example (see Joyce & Weil, 1980; Orlich et al., 1985; Ruggiero, 1988), suggest that differential interactions between teachers and students may be evident. For example, teachers use lecture, guided and unguided discovery, texts, experimentation, and field trips quite differently across these two content areas. Carnine and his colleagues (Carnine, 1989; Darch & Carnine, 1986), however, demonstrated that interactive technology, combined with instructional design principles, can increase instructional efficiency despite content area differences. Thus, we examined the effects of a teaching effectiveness package on both (a) social studies and (b) science teachers' interactions with (a) mainstreamed and (b) regular education middle school students. We also examined the extent of impact on noninteractive student behavior. As in our previous study, the effects of class and student type were examined, but these were further considered in terms of subject matter (social studies vs. science).



Middle school social studies and science teachers from three districts adjacent to metropolitan Houston participated in the study. The districts were selected based on approximate equivalence in terms of (a) size of school enrollment, (b) socioeconomic status of the community, and (c) availability of mainstreamed students in social studies and science classes. In addition, the districts were in close geographic proximity to one another. Teachers were selected based on the following criteria: * Currently teaching social studies or science in

the 6th, 7th, or 8th grade. * Enrollment of at least one, but preferably two,

student(s) eligible for special education

services mainstreamed into their classes. * Willingness to participate in the study.

Ultimately, 35 teachers participated.

Teachers were grouped according to three factors: (a) content (social studies vs. science), (b) experimental condition (control vs. experimental) and (c) classroom type (heterogeneous vs. homogeneous class). The configuration of teachers among these three factors is presented in Table 1. Homogeneous classes were defined operationally as those having a limited, generally low range of student achievement, whereas heterogeneous classes had a wide range of student achievement. Classroom type was determined in two steps. First, heterogeneous and homogeneous classes were selected by principals. This selection was then confirmed by teachers. A class was selected only if agreement on class type was achieved. Random assignment of teachers as control and experimental was impossible because some teachers were willing to participate in the study but unwilling to participate in the experimental intervention. Thus, the distribution of teachers among subject matter content, homogeneous and heterogeneous classes, and experimental and control conditions was unequal.
 Distribution of Teachers Across All Factors
 Heterogeneous Heterogeneous
Content Teachers Classes Classes
 Experimental (n = 13)
 studies 5 3 2
Science 8 2 6
 Control (n = 22)
 studies 13 6 7
Science 9 2 7

Observation Procedures

Dependent Measures. An observation code describing teacher-student interactions and student noninteractive behaviors and based on the code used by Brady, Swank, Taylor, and Freiberg (1988) was used in the study. Five categories of teacher-student interactions were labeled as either (a) academic or (b) nonacademic interactions. A separate category was identified for any type of negative interaction. Noninteractive behaviors likewise were labeled as (a) academic and (b) nonacademic. The observation system employed an interval-collection procedure (see "Data Collection") with a two-stage hierarchy, where interactive behavior was the primary target and noninteractive performance was the secondary target; that is, noninteractive behavior was coded only if there had been no occurrence of any of the interactive codes. Thus, the data for each teacher included the percentage of intervals in which there were (a) interactions between the teacher and individual students and (b) specific noninteractive student behaviors. The observation code is presented in Table 2.
 Teacher-Student Observation Code
 Academic Variables
 Teacher-Student Interactions
 Teacher asks direct, evaluative, open-ended
 academic questions. Example: "Can you describe
 how magnets work?"
 Teacher directs, demonstrates, or physically
 guides student toward on academic response.
 Example: "Draw your latitude lines like this. . ."
 Teacher delivers academic information,
 instructions, or explanations. Includes lectures. Example:
 "Houston is the home of the nation's major space
 Teacher corrects academic error. Example: "Not
 quite; remember that Texas is west of Louisiana."
 Teacher provides positive feedback (praise,
 encouraging gestures, affirmations) designed to
 reinforce academic behavior. Example: "Nice job!
 You know every major city in the state."
 Nonacademic Variables
 Teacher asks organizational, or social, or other
 nonacademic question. Example: "How should
 you respond when Susan is angry with you?"
 Teacher directs, demonstrates or physically guides
 student through social, conduct, organizational, or
 other nonacademic response. Example: "Move
 away from Susan when she's angry."
 Teacher provides social, organizing, or other
 nonacademic information or instructions. Example:
 "If you argue with her, you both will be wrong and
 might end up in trouble."
 Teacher corrects conduct, social, or other
 nonacademic behavior. Example: "Walk away from her
 if she makes you angry. Don't make it worse by
 Teacher provides positive feedback (praise,
 encouraging gestures, affirmations) designed to
 reinforce social or other nonacademic behavior.
 Example: "Good, Marty. You just ignored her.
 That was the right thing to do."
 Teacher delivers negative academic or nonacademic
 feedback (e.g., comments, gestures) or punishes
 academic, social, conduct, or other nonacademic behavior.
 Examples: "Of course Alaska is part of the United
 States. What a stupid question!"; "Only silly people
 do that, Mary. You're awful!"
 Noninteractive Student Behavior
On Task:
 Student performs assigned academic task.
 Example: Student "attends" during lecture, completes
 class assignments, etc.
Learning Material:
 Student actively uses learning materials during
 academic task. Example: Student uses ruler, map,
 and pencil to mark latitude lines.
Nonacademic Involvement:
 Student uninvolved or engaged in nonacademic
 task. Example: Student reads a comic book or
 plays tic-tac-toe during class assignments.
 Student engaged in conduct or discipline problem.
 Example: Student tantrums or argues with teacher.
Academic or Social Interactions with Peers:
 Student involved in academic, tutorial, or social interaction
 with another student. Examples: Student works
 with peer to identify the names of states. Student chats with
 peer about weekend volleyball game.

Data Collection. All teachers were observed once per day during their regular 50-minute social studies or science classes. Dependent measures were coded by observers using a partial interval system (Kazdin, 1982), which employed a portable cassette, cueing tape, and lightweight headphones to note the beginning and end of an interval. Cued observations of 14 seconds (s) each were followed by a 6-s recording period. Observations followed a six-student rotation, that is, were rotated sequentially among the teacher and six preselected students, from Student 1 to Student 2 and so on, until all had been observed. At this point the rotation was repeated. In classes with only one mainstreamed student, the observation cycle was adapted to a three-person rotation. That is, following Student 3, observers returned to Student 1 and repeated the sequence. The observation sequence always started by observing a mainstreamed student first.

Selection of individual students for observation was based on the following criteria: 1. At least one mainstreamed student in each

classroom was selected. 2. Two regular education students adjacent

to or in close proximity to the observed

mainstreamed student(s) were then selected. 3. In heterogeneous classes, both low-achieving

and average-achieving regular education

students were selected. 4. If more than two mainstreamed students were

in the class, the students to be observed were

selected randomly.

Mainstreamed students were identified primarily as having learning disabilities. Students with other disabilities, however, were included because (a) the target population in this study was teachers, not students, and (b) each district mainstreamed students on an individual rather than categorical basis. Students identified as having disabilities related to health, physical, hearing, or vision impairments, however, were excluded from the study. The same set of selected students was observed across all observation sessions.

Observer Reliability. To establish the reliability of observations, a second observer simultaneously but independently coded the behavior of the teacher and the students. Two observers, separated by approximately two meters, listened to the same interval tape with separate headphones. Before the session, the primary observer identified the students to be observed and established the rotation order. Thus, when the observation started, the cueing tape prompted both observers to code the same students in the same sequence. Reliability was calculated using a generalizability coefficient on all behavior codes (see Frick & Semmel, 1978). This yielded a median reliability index of .93 across all observations. Experimental Procedures

The study employed (a) preintervention observations and feedback, (b) the experimental intervention, (c) postintervention observations and feedback, and (d) follow-up observations and feedback. During all observations, observers were blind to the composition of the experimental and control groups.

Preintervention Observations and Feedback. Three days of preintervention observations per teacher (one class per day) were conducted on consecutive days in October and November. All teachers also were observed by a second set of observers as part of an adjunct study (see Brady, Taylor, Freiberg, & Swank, 1988), using the Stallings Observation Instrument (Stallings & Needels, 1985). Following the observations, all teachers received profiles and verbal feedback of their Stallings observations.

Intervention. The experimental group received intervention over a 6-week period during January and February. Workshop content was approved by the state education agency and resulted in career ladder credit for participants. The content of the intervention was designed to increase teachers' academic interactions and decrease the amount of class time devoted to discipline, management, off-task student behavior, and organizational activities. Specifically, the intervention included material on (a) effective uses of instructional time; (b) classroom organization and grouping; (c) classroom management; (d) instruction interactions and questioning; (e) self-assessment, monitoring, and instruction; and (f) planning and delivering instruction with mainstreamed and low-achieving students.

The format of the intervention included (a) review of previous content, (b) integration of new topics with previous activities, (c) discussion and exchange of ideas, (d) summary of activities, (e) evaluation of the seminar, and (f) assignment of specific practice activities for the following week. These practice activities included audio taping students, observing other classes, using seating charts, and using self-assessment of teaching time. (A more detailed description of the experimental intervention is available in Brady, Taylor, Freiberg, & Swank, 1988.)

Postintervention Observations and Feedback. Postintervention observations were conducted in February using the same 3-day observation schedule from the preintervention observations. As in the earlier observations, teachers again were observed for the adjunct study. All teachers were given profiles and verbal feedback summarizing their teaching behavior on this second observation system.

Follow-up Observations and Feedback. Follow-up observations were conducted during April using the same schedule regimen as in previous observations. Feedback of results of the adjunct observation system was given to both control and experimental groups. Experimental Design

This study employed a control-group factorial design, with the presence or absence of the intervention as the experimental variable (condition) and with content (social studies or science), classroom type (heterogeneous or homogeneous) and student type (mainstreamed or nonmainstreamed) as moderator variables. The 16 dependent variables were assessed at three points in time: preintervention, postintervention, and follow-up. Due to the hierarchical nature of the observation system, the interactive codes were considered separately from the noninteractive codes. Negative interactions were not analyzed because of their extremely low occurrence and lack of variance.

A factorial multivariate analysis of variance was performed separately for each set of variables. Alpha was set a priori at .05 for each multivariate analysis. Post hoc univariate tests were conducted when the multivariate tests reached the .05 level of significance. Type 1 error was controlled on the univariate tests by dividing the overall alpha by the number of tests conducted yielding the .005 significance level for the interactive variables and .01 for the noninteractive variables. Because univariate tests do not always identify which variables contributed to the multivariate differences, a discriminant analysis for each significant effect also was conducted.

Because we hypothesized that (a) the experimental group would have significantly more effective teaching patterns following the intervention than the control group and (b) the experimental group would maintain these changes through the follow-up period, the three repeated measures (pre, post, and follow-up) were transformed to provide two contrasts to test the hypotheses. The first contrast consisted of the difference between all preintervention observations and the average of the postintervention and follow-up observations. Because this contrast compares the preintervention observations with all other observations, it was referred to as the pre-post contrast. The second contrast compared the postintervention means to the follow-up means and reflected whether any potential changes that occurred would be sustained over time. This was referred to as the follow-up contrast. The third transformed variable, which is not independent of the other two, was the average of pre, post, and follow-up observations. Because this analysis ignored effects of time and treatment, it provided an analysis of the between-subjects effects.


To address the major hypotheses, we have separated our presentation of results by interactive and noninteractive variables. For each set of variables, the two orthogonal contrasts, pre-post and follow-up, are considered first. Analysis of the between-subjects effects follows. Student-Teacher Interactions

There were two significant multivariate tests for the teacher-student interaction variables. The Condition x Content effect reached significance, F(20,163) = 2.11; p < .05, as did Class type F(20, 163) = 2.38; p < .05.

Condition by Content. In considering the Condition x Content result, one univariate test, academic information, F(1, 182) = 8.34; p < .005, reached the more stringent alpha level in the pre/post contrast. Considering the value of the contrast, the control group actually increased their delivery of academic information slightly more than the experimental group; the results were not consistent by content, however. The greatest increase in academic information was made by the science teachers in the experimental group, whereas the experimental social studies teachers decreased the time spent giving academic information following training. In the control group, the social studies teachers had the greatest gain in academic information, with a much smaller gain by the science teachers in the control group. Two additional univariate tests approached but did not reach the more stringent alpha level. These should be considered as possible trends rather than actual effects. The experimental group spent less time giving academic reinforcement following the intervention, but this was primarily due to the social studies teachers, F(1, 82) = 6.62; p < .02; the experimental science teachers increased their reinforcement after the intervention. The control group science teachers, on the other hand, decreased their academic reinforcement; and the control social studies teachers increased slightly.

One follow-up contrast showed a possible trend for nonacademic information, F(1, 182) = 4.08; p < .05. The experimental group increased their nonacademic information slightly from post to follow-up, but this was due primarily to the science teachers; the experimental social studies teachers decreased nonacademic information. The control science teachers also decreased their use of nonacademic information from post to follow-up, whereas the control social studies teachers showed little change. These results were somewhat unusual because there were no significant changes in nonacademic information from preintervention to postintervention.

Table 3 gives the standardized discriminant function coefficients for the 10 measures used in the analysis for each contrast. The variables contributing to the discriminant function that best differentiated the Condition x Content for the pre-post contrast were (a) academic reinforcement, (b) nonacademic information (c) nonacademic questions, and (d) academic information. Of these four variables, the one with the lowest coefficient is the variable that was significant in the univariate tests. It must be remembered that these coefficients represent the contribution of each variable in the presence of the others. Thus, due to the correlation of academic information with academic reinforcement and nonacademic questions, the coefficient of academic information is reduced. Because the coefficient for nonacademic information in the pre-post contrast is positive, and because this variable did not differentiate in the univariate case, it appears that there was a tendency for the experimental science teachers and the control social studies teachers to increase their use of academic information, academic reinforcement, and nonacademic questions compared to their cohorts following the treatment. This was particularly so in the absence of a concomitant change in the frequency of nonacademic information.
 Discriminant Function Coefficients for the
 Condition x Class Contrast
Variable Pre-Post Follow-up
 Changes Changes
Academic variables
 Questions .25 -.25
 Guidance .38 -.06
 Information -.42(**) .03
 Corrections .01 -.22
 Reinforcement -.65(*) -.04
Nonacademic variables
 Questions -.48 -.39
 Guidance .10 -.05
 Information .51 -.40(*)
 Corrections -.14 .29
 Reinforcement -.03 .40
(*)p < .05, (**)p < .005, univariate tests.

The differences are not as striking for the follow-up changes, but there was a slight tendency for the experimental science and control social studies teachers to decrease the frequency of their nonacademic information and questions from posttest to follow-up when compared to control science and experimental social studies teachers, particularly when there was no corresponding change in the frequency of nonacademic reinforcement. Because there was no corresponding change in the pre-post contrast for nonacademic information, it may be more helpful to look at the means for the three assessment periods rather than just at the post versus follow-up. All groups except the experimental science teachers increased their nonacademic information from pre to post and then decreased during the follow-up assessment; the experimental science teachers did just the opposite. It should be noted that the experimental science teachers started with the highest level of nonacademic information at pre-test and also finished at the highest level, whereas the experimental social studies teachers ended with the lowest level of nonacademic information (less than one third of the time spent by the next lowest group).

Class (Homogeneous vs. Heterogeneous) Effects. The second multivariate significant effect occurred for the class type variable, F(20,163) = 2.38; p < .05. This is a main effect and, because the Class x Condition interaction was not significant, was not dependent on the experimental intervention. Univariate results indicated that teachers with heterogeneous classes increased their use of academic corrections from the first observations to the last two, whereas teachers with homogeneous classes did the opposite, F(1, 182) = 8.54; p < .005. Consideration of the discriminant function coefficients indicated that this difference was particularly evident when there was no corresponding change in the frequency of academic reinforcement. There was, however, a decrease in nonacademic questions from post to follow-up for the homogeneous classes, with the opposite effect for the heterogeneous classes, F(1, 182) = 8.07; p < .005.

There were several directional findings here, as well. Teachers with heterogeneous classes tended to decrease academic guidance from pre to post follow-up more than the homogeneous classes. F(1, 182) = 5.16; p < .03. In addition, teachers with heterogeneous classes declined in the time spent giving nonacademic information, F(1, 182) = 5.83; p < .02, and making nonacademic corrections, F(1, 182) = 5.65; p < .02, from the post to follow-up assessments, whereas teachers of homogeneous classes tended to increase these behaviors. Again, even though these effects occur overtime, they cannot be attributed to the intervention.

Between-Subjects Effects. Only two significant multivariate results occurred for the between-subjects effects. Because these effects ignored changes over time any interaction with treatment condition were ignored. Thus, we did not interpret the Condition x Class x Content effect, even though the multivariate test reached significance. However, there were differences in overall interactive effects related to the student type, Wilks's lambda = .87; F(10,173) = 2.53; p < .05. Follow-up univariate tests indicated that mainstreamed students received more academic questions, F = 12.44; p < .005; academic information, F =8.98; p < .005; and academic reinforcement, F = 10.12; p < .005 than did regular education students (all degrees of freedom equal to 1, 182). In addition, there was a tendency for mainstreamed students to obtain more academic guidance, F = 5.31; p < .03, and academic corrections, F = 5.29; p < .03. In considering the standardized discriminant functions coefficients, there was some indication that these differences were greatest in the absence of concomitant differences in the frequency of nonacademic corrections. Noninteractive Student Behavior

When we examined the noninteractive student behaviors alone, there were several significant multivariate tests for experimental condition F(10, 173) = 2.26; p < .05; class type, F(10, 173) = 4.03; p < .05; course content, F(10,173) = 6.26; p < .05; and the experimental Condition x Class type interaction, F(10, 173) = 3.13; p < .05. In addition, the experimental Condition x Content interaction approached significance, F(10, 173) = 1.79; p < .065. However, these effects are all subsumed by the significant experimental Condition x Class type x Content interaction, F(10, 173) = 4.45; p < .05. Univariate tests indicated significant effects on the pre-post contrast for on-task behavior, F(1, 182) = 9.52; p < .01, and nonacademic involvement, F(1, 182) = 9.09; p < .01, and on the follow-up contrast for peer involvement, F(1, 182) = 7,86; p < .01. In the standardized discriminant function analysis it appeared that, in the absence of change in the use of learning materials, student time on task provided the most differentiation among the noninteractive variables. Likewise, the follow-up changes for peer involvement differentiated best in the absence of variations in nonacademic involvement.

Pre-Post Contrasts. In considering the means for the pre-post contrast, it is evident that the time on task and nonacademic involvement are inversely related (as they should be). Students in all groups, except the homogeneous classes of experimental science teachers, decreased their time on task and increased their nonacademic involvement when comparing their preintervention to postintervention means. The group that did change in the predicted direction had the lowest on task mean (76.5) and the highest nonacademic involvement mean (21.3) of all groups during the preintervention observations. (Note: All means in this section are reported as percentage of total intervals of occurrence.)

Unfortunately, those students with the largest on-task declines and the largest nonacademic involvement increases were in homogeneous classes of the experimental social studies teachers. The students in the heterogeneous classes of control group social studies teachers also showed some decline in on-task behavior, with concomitant increases in time spent nonacademically involved. This decline, however, was only about half that of the previously described group. The remaining groups were relatively stable across the three assessments.

Follow-up Contrast. With respect to the follow-up changes, the major differences occurred when students in (a) heterogeneous classes of experimental science teachers and (b) homogeneous classes of control science teachers declined in the time spent interacting with peers from postintervention to follow-up. The second group, however, had only half the change of the former. Students in homogeneous classes of experimental social studies teachers increased their time spent in peer interactions from post to follow-up. Again, because there was no concomitant difference in the pre-post contrast, it might be more illuminating to consider the means at all three assessment points. Heterogeneous science classes in the experimental condition had the highest level of peer interaction of all groups during preintervention observations (16.8), increased even more during the postobservations (24.0), but then fell to the lowest level of all groups during the follow-up (6.3). Homogeneous science classes in the control condition also increased in peer interactions from pre (10.7) to post (17.8), but then declined below their preintervention levels during follow-up (9.3). Homogeneous social studies classes in the experimental condition continued to increase in their peer interactions across the post (8.0) and follow-up observations (14.6). All other groups remained relatively stable. There were no significant effects on the average of the pre, post, and follow-up observations for any effect that was not subsumed by the previous interaction.


The results of the present study extend our earlier findings to a previously underinvestigated group of teachers, science middle school teachers with mainstreamed students. Unfortunately these results are, in part, inconsistent with certain previous findings. In our previous study (Brady, Swank, Taylor, & Freiberg, 1988) using a similar intervention in social studies classrooms only, we found that experimental teachers increased their use of academic questions and academic reinforcement over the control group. Also noted was an increase in nonacademic information by all teachers, although this was moderated by class type (i.e., the experimental heterogeneous teachers tended to increase more than the others).

Those results were only partially replicated in this study. First, there were no training effects in this study on academic questions. We did see a tendency (.005 < p < .05) toward an increase in academic reinforcement following the intervention, but only among science teachers. The experimental social studies classes, which showed an increase in our previous study, decreased their use of academic reinforcement following the intervention. It is also important to note that, in our earlier work, the intervention did not significantly increase teachers' use of academic information. Such an effect was noted in this study, though it was moderated by the instructional content. Science classes showed an increase in academic information following the intervention, but the social studies classes showed a slight decline, contrary to our hypothesis. Both the control science and social studies classes showed small increases, though not as large as those of the experimental science classes.

The main question here involved the inverse effect on science versus social studies classes in response to the intervention. It is also unclear why other positive effects of the intervention occurred only with the science teachers. For example, in addition to their increases in academic reinforcement and information, the students in these teachers' classes (a) increased their on-task time and (b) decreased the time they spent involved in nonacademic activities, although these effects were only noted for homogeneous classes. Our explanation for this remains speculative, at best. Since the experimental social studies teachers failed to substantially increase their delivery of academic information in either study, one might conclude that teachers already deliver this content information in a relatively stable manner over time. Delivery of the middle school science curriculum, on the other hand, may be more susceptible to improvement if teachers learn to make better use of labs, experiments, or simulations. For example, teachers who deliver more instructional information, along with opportunities for students to participate actively in learning activities, seem likely to increase academic reinforcement for participation and learning. The linkage of reinforcement to information, particularly when these behaviors are academic in nature, has a great deal of face validity and, indeed, is quite encouraging. This explanation, however, remains highly speculative until future investigations yield direct and additional data regarding the differences between teachers of different content areas.

As in our previous study, we again found important differences and similarities based on student type. In both studies, mainstreamed students received more academic guidance and information than did regular education students; previously, they also received more nonacademic guidance and information and spent less time on task when they were not interacting. The current study found that mainstreamed students received more academic questions, corrections, and reinforcement, as well; in an important departure from the previous study, there were no differences between mainstreamed and general education students on nonacademic interactions. These results are particularly encouraging given reports by some researchers that low achieving children may obtain (a) poorer quality and fewer teaching interactions and (b) a preponderance of noninstructional interactions (Brophy & Good, 1970, 1986; Bryan, 1978). Our data support the findings of researchers who note that instructional interactions among groups of students is a function of variables other than category or achievement level (Gable et al., 1983; Ysseldyke, Algozzine, Shinn, & McGue, 1982). Where student type was associated with instructional differences, mainstreamed students obtained higher, not lower, quality instruction.

In conclusion, important changes in the teaching interactions of one group of teachers, science teachers, were observed as a result of the teacher effectiveness training. As important were the similarities and differences in teaching interactions based on student type. Given the paucity of research documenting (a) the quality of instruction delivered to mainstreamed students, (b) the extent to which teaching effectiveness training can improve instruction to these student (Ivarie et al., 1984; Madden & Slavin, 1983) and (c) the sensitivity of teacher appraisal systems for these teachers (Swank et al., 1989), our findings are optimistic indeed. In addition, the differences in teaching based on the course content supports preliminary findings by Zigmond and Baker (1990) that the probability of mainstreaming success can be enhanced by considering the type of class or instructional content into which students with disabilities are initially placed. Our results then, when considered with the emerging data base on effective teaching, suggests that the general education environment is, or can become, sufficiently supportive for students with a range of academic skills. It is clear, however, that if the general education environment is to become that accepted educational setting for all students, numerous instructional variables will need to be studied and improved (Reynolds et al., 1987; Reith, Polsgrove & Semmel, 1981; Thompson, 1983).


Affleck, J., Madge, S., Adams, A., & Lowenbraun, S. (1988). Integrated classroom versus resource model: Academic viability and effectiveness. Exceptional Children, 54, 339-348. Algozzine, B., Morsink, C., & Algozzine, K. (1988). What's happening in self-contained special education classrooms? Exceptional Children, 55, 259-265. Baker, J., & Zigmond, N. (1990). Are regular education classes equipped to accommodate students with learning disabilities? Exceptional Children, 56, 515-526. Brady, M. P., Swank, P. R. Taylor, R. D., & Freiberg, J. (1988). Teacher-student interactions in middle school mainstreamed classes: Differences with special and regular education students. Journal of Educational Research, 81, 332-340. Brady, M. P., Taylor, R. D., Freiberg, J., & Swank, P. (1988). Middle school study of mainstreamed students dissemination report: A summary of year one findings and year two research activities. Houston: University of Houston, College of Education. (ERIC Document Reproduction Service No. ED 295 364) Brophy, J., & Good, T. (1970). Teachers' communication of differential expectations for children's classroom performance: Some behavioral data. Journal of Educational Psychology, 61, 365-374. Brophy, J., & Good, T. (1986). Teacher behavior and student achievement. In M. Wittrock (Ed.), Handbook of research on teaching (3rd ed., pp. 328-376). New York: Macmillan. Bryan, T. (1978). Social relationships and verbal interactions of learning disabled children. Journal of Learning Disabilities, 11(2), 58-66. Carnine, D. (1989). Teaching complex content to learning disabled student: The role of technology. Exceptional Children, 55, 524-533. Darch, C., & Carnine, D. (1986). Teaching content area material to learning disabled students. Exceptional Children, 53, 240-246. Dunn, L. (1968). Special education for the mildly retarded--Is much of it justifiable? Exceptional Children, 35, 5-22. Eder, D. (1981). Ability grouping as a self-fulfilling prophecy: A micro-analysis of teacher-student interaction. Sociology of Education, 54, 151-162. Frick, T., & Semmel, M. (1978). Observer agreement and reliabilities of classroom observational measures. Review of Educational Research, 48, 157-184. Gable, R., Hendrickson, J., Young, C., Shores, R., & Stowitschek, J. (1983). A Comparison of teacher approval statements across categories of exceptionality. Journal of Special Education Technology, 6, 15-22. Good, T. L., & Brophy, J. E., (1987). Looking in classrooms. New York; Harper & Row. Hendrickson, J., Roberts, M., & Shores, R. (1978). Antecedent and contingent modeling procedures to teach reading to learning disabled children. Journal of Learning Disabilities, 11, 524-528. Hendrickson, J., & Stowitschek, C. (1980). Teacher use of diagnostic questioning and modeling in language development. Journal of Special Education Technology, 4, 17-27. Ivarie, J., Hogue, D., & Brulle, A. (1984). An investigation of mainstream teacher time spent with students labeled learning disabled. Exceptional Children, 51, 142-149. Joyce, B., & Weil, M. (1980). Models of teaching (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall. Kazdin, A. (1982). Single-case research designs: Methods for clinical and applied settings. New York: Oxford. Larrivee, B. (1986). Effective teaching for mainstreamed students is effective teaching for all students. Teacher Education and Special Education, 9, 173-179. Madden, N., & Slavin, R. (1983). Mainstreaming students with mild handicaps: Academic and social outcomes. Review of Educational Research, 53, 519-569. Orlich, D. C., Harder, R. J., Callahan, R. C., Kravas, C. H., Kauchak, D. P., Pendergrass, R. A., & Keogh, A. J. (1985). Teaching strategies: A guide to better instruction (2nd ed.). Lexington, MA: Health. Reith, H., Polsgrove, L., & Semmel, M. (1981). Instructional variables that make a difference: Attention to task and beyond. Exceptional Education Quarterly, 2, 61-71. Reynolds, M., Wang, M., & Walberg, H. (1987). The necessary restructuring of special and regular education. Exceptional Children, 53, 391-398. Rich, H. L., & Ross, S. (1989). Student's time on learning tasks in special education. Exceptional Children, 55, 508-515. Ruggiero, V. R. (1988). Teaching thinking across the curriculum. New York: Harper & Row. Stallings, J., & Needels, M. (1985). Stallings Observation Instrument (rev. ed.). Stanford: SRI International. Stevens, R., & Rosenshine, B. (1981). Advances in research on teaching. Exceptional Education Quarterly, 2, 1-10. Strains, P. S., Lambert, D., Kerr, M. M., Stagg, V., & Lenkner, D. (1983). Naturalistic assessment of children's compliance to teacher's requests and consequences for compliance. Journal of Applied Behavioral Analysis, 16, 243-249. Swank, P., Taylor, R., Brady, M., & Freiberg, J. (1989). Sensitivity of classroom observation systems: Measuring teacher effectiveness. Journal of Experimental Education, 57, 171-186. Thompson, R. (1983). Teacher-student interaction patterns within the learning environment of mainstreamed classrooms (OSERS Report No. 023CH10250). Washington, DC: U.S. Department of Education, Office of Special Education and Rehabilitative Services. Thompson, R., White, K., & Morgan, D. (1982). Teacher-student interaction patterns in classrooms with mildly handicapped students. American Educational Research Journal, 19, 220-236. Wang, M., Peverly, S., & Randolph, R. (1984). An investigation of the implementation and effects of a full-time mainstreaming program. Remedial and Special Education, 5, 21-32. Ysseldyke, J., Algozzine, R., Shinn, M., & McGue, M. (1982). Similarities and differences between low achievers and students classified as learning disabled. Journal of Special Education, 16, 73-85. Zigmond, N., & Baker, J.(1990). Mainstream experiences for learning disabled students (Project MELD): Preliminary report. Exceptional Children, 57, 176-185. MICHAEL P. BRADY (CEC # 153) is Associate Professor of Special Education Programs and PAUL R. SWANK is Associate Professor of Measurement and Statistics in the Department of Educational Psychology at the University of Houston, Texas. RONALD D. TAYLOR is Assistant Professor in the Department of Psychology at Columbia College in Columbia, Missouri. JEROME FREIBERG is a Professor in the Department of Curriculum Instruction at the University of Houston, Texas.
COPYRIGHT 1992 Council for Exceptional Children
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Brady, Michael P.; Swank, Paul R.; Taylor, Ronald D.; Freiberg, Jerome
Publication:Exceptional Children
Date:May 1, 1992
Previous Article:Relationship between measures of adaptive functioning and community adjustment for adults with mental retardation.
Next Article:CEC's standards for professional practice in advocacy: members' attitudes and activities.

Related Articles
Peer interactions in mainstreamed and specialized classrooms: a comparative analysis.
Mathematics achievement of hearing impaired adolescents in different placements.
Effects of social integration on preschool children with handicaps.
Academic behavior and grades of mainstreamed students with mild disabilities.
Factors influencing the social status of children with mild academic disabilities in regular classrooms.
Learning disabilities and social competence: a social ecological perspective.
Observations of students with learning disabilities in general education classrooms.
Teacher-student classroom interactions: the influence of gender, academic dominance, and teacher communication style.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters