Science Instruction for Students With Emotional and Behavioral Disorders

Abstract

This review examined classroom science instruction for students with emotional and behavioral disorders (EBD). A total of 11 group and single-subject studies were analyzed. Across all group studies, a conservatively calculated mean effect size of .471 was obtained indicating the interventions as a whole had at least a small to moderate impact on students’ with EBD achievement. Findings were further analyzed by student characteristics, intervention type, dependent measures utilized, and study variables. A significant result of these subanalyses indicates that while additional research is needed, students with EBD may benefit from inquiry approaches provided the method implemented includes enough structure to ensure student engagement. Results also suggest that mnemonic instruction is highly effective at increasing students’ with EBD knowledge and retention of science facts.

Keywords

academic achievement general education curriculum evidence-based practice emotional and behavioral disorders exceptionalities science instruction

Students with emotional and behavioral disorders (EBD) are characterized by a range of non-academic and academic behavioral problems (Lane, Wehby, & Barton-Arwood, 2005). Studies have shown that students with EBD make less academic progress, earn lower grades, and have more disciplinary referrals than students in any other disability categories (Bradley, Henderson, & Monfore, 2004; Wagner et al., 2003). Until recently, interventions for students with EBD have primarily addressed non-academic difficulties (Dunlap & Childs, 1996; Vaughn, Levy, Coleman, & Bos, 2002). These non-academic difficulties include externalizing and internalizing behaviors that impact students’ abilities to relate to teachers and peers, utilize problem-solving skills, and follow classroom and school rules.

One reason why interventions for students with EBD target non-academic behavior is to improve student behavior to the point that they can be educated in general education settings alongside their non-disabled peers (Furlong, Morrison, & Jimerson, 2004). With the increased emphasis of placing students in the least restrictive environment (LRE) that came with the amendments to the Individuals With Disabilities Act in 1997 (IDEA, 1997), approximately 80% of students with EBD are currently receiving most of their core instruction in the general education setting (Gaylord, Quinn, McComas, & Lehr, 2005). These changing factors combined with the large gaps in students’ with EBD academic achievement in core content areas (Lane, 2004) has resulted in academic instruction for students with EBD, particularly in regular education settings, coming to the forefront.

The recent research that has focused on academic instruction for students with EBD has found that targeting academic outcomes often leads to non-academic behavioral improvements. For instance, Lane (2004) conducted a review of research that looked at academic instruction for students with EBD. She examined a total of 26 studies including 14 studies in the area of reading, 11 in math, and 1 in writing. In general, Lane (2004) found that academic focused intervention studies reported both positive effects on students with EBD academic achievement and on non-academic behaviors.

Despite the increased focus on academic instruction for students with EBD, one area that has yet to receive much attention is classroom science instruction. The lack of attention to science instruction is unfortunate and may put students with EBD at a distinct disadvantage to acquire critical knowledge and skills. The United States Department of Labor predicts there will be tremendous job growth in STEM (Science, Technology, Engineering, and Math) related fields over the next 20 years (Terrell, 2007). Furthermore, they note that even the lowest paying vocations in STEM fields pay a livable wage (Terrell, 2007). Therefore, professionals that provide educational program to students with EBD must ensure their students master basic competencies in science if they hope to equip them with the skills needed in the 21st century.

Approaches to Teaching Science

Traditionally, science instruction in general education classrooms used textbook or lecture style presentations as the main instructional method. Both these instructional types can be problematic for students with EBD as they often have limited content knowledge, difficulty connecting prior knowledge to new information, and poor academic motivation (Scruggs & Mastropieri, 2000a). In recent years, the field of science education has promoted the use of inquiry-based science instruction as the preferred instructional method as it replicates typical science practices (American Association for the Advancement of Science, 1997; National Science Teachers Association Board of Directors, 2004).

The most recent review of literature regarding students with EBD and science instruction was conducted by Scruggs and Mastropieri (1995). In their review, they categorized interventions used into two broad groups: activities-oriented and content-oriented approaches. Hands-on and discovery learning strategies are the primary instructional modes included under activities-oriented approaches (i.e., inquiry) to teaching science (Scruggs & Mastropieri, 1995). Unfortunately, the term inquiry-based instruction has not been clearly defined and its instructional components vary widely across studies (Klahr & Li, 2005). While the specific features of inquiry instruction may differ across approaches, the National Research Council (2012) in the recent framework on K–12 science education clearly defines common components found within inquiry-based instruction. These components include practices with an emphasis on data, evidence as a foundation for claims, and the use of argumentation and analysis of evidence to develop ideas about science. It is also reasonable to conclude that all inquiry approaches include the use of hands-on learning in combination with these other instructional components. All inquiry approaches can generally be located along a continuum from pure discovery approaches that are completely student directed to teacher directed approaches (Scruggs & Mastropieri, 2007).

Unfortunately, the impact of inquiry-based instruction for students specifically with EBD is unclear. It may be possible that the hands-on aspect of inquiry-based instruction increases the motivation for EBD students to stay engaged. If this were true, students with EBD would benefit academically. On the other side though, it may be possible that the reduced structure associated with inquiry approaches could decrease time-on-task behavior. This lack of structure could in turn result in lower academic achievement for students with EBD.

Although few researchers have examined the efficacy of science interventions for students with disabilities, Therrien, Taylor, Hosp, Kaldenberg, and Gorsh (2011) conducted a meta-analysis of science instruction for students with learning disabilities (LD). They found that (a) teacher directed inquiry-based instruction and (b) mnemonics were effective at improving science-related outcomes for students with LD. Therefore, to expand on this research, the purpose of this review was to synthesize the efficacy of classroom science instruction for students with EBD. Utilizing the analysis framework developed by Therrien and colleagues (2011), we specifically sought to answer the following questions: (a) What classroom based instructional methods are effective at increasing the science achievement of students with EBD? (b) What additional science intervention study characteristics produce higher effect size (ES) for students with EBD?

Method

Literature Search

Prior to looking for articles, we formed search criteria for the studies that would be considered for review. Articles must have been (a) published in a peer review journal in 1980 or later (b) focused on classroom-based interventions in the area of science. Classroom based interventions were defined as core or supplemental instruction that enhanced students’ knowledge of science concepts and was not geared solely toward improving reading skills, (c) was experimental or quasi-experimental (including both group and single-subject designs) in nature, and (d) included school-aged (i.e., Grades K–12) participants with EBD as subjects (results did not have to be disaggregated for subjects with EBD).

All studies from three previous literature reviews on science education and students with disabilities were initially collected (Mastropieri & Scruggs, 1992, 1998; Scruggs & Mastropieri, 1995). Next, a search of ERIC and PsycINFO was completed dating from 1995 to 2010 using keywords: science and emotional disturbance; science and emotional disorder*; science and behavior* disorder*; science and exceptional children; science and special education. A hand-search of the following journals from 2005 to 2010 was also conducted: Behavioral Disorders, Exceptional Children, International Journal of Science Education, Journal of Emotional and Behavioral Disorders, Journal of Research in Science Teaching, Journal of Special Education, and Science Education. These search methods produced 34 articles that met our initial criteria.

The 34 articles were then reviewed in more depth to ensure they met the inclusion criteria and to determine whether they were amenable to meta-analysis techniques. Through the review process, we excluded 23 articles. Eight of these articles were excluded because they did not evaluate a science intervention (Cawley, Hayden, Cade, & Baker-Kroczynski, 2002; Ervin, DuPaul, Kern, & Friman, 1998; McFarland & Shepard, 1995; Pomplun, 1997, 1998; Rogevich & Perin, 2008; Scruggs & Marsing, 1987; Scruggs, Mastropieri, Veit, & Osguthorpe, 1986). Another 6 articles were excluded because they did not include students with EBD (Brigham, Scruggs, & Mastropieri, 1992; Mastropieri, Scruggs, & Butcher, 1997; Mastropieri, Scruggs, Whittaker, & Bakken, 1994; Nolet & Tindal, 1994; Scruggs, Mastropieri, Bakken, & Brigham, 1993; Scruggs, Mastropieri, Sullivan, & Hesser, 1993). Eight more articles were excluded because they were conceptual or qualitative articles (Okilwa & Shelby, 2010; Parmar, Deluca, & Janczak, 1994; Ryan, Reid, & Epstein, 2004; Scruggs & Mastropieri, 1993, 1994a, 1994b, 2000a, 2000b). The MacDougall, Schnur, Berger, and Vernon (1981) article was excluded because the presentation of data was not conducive for calculating an ES.

After completing this review process, 11 articles were identified as meeting all inclusion criteria (Bay, Staver, Bryan, & Hale, 1992; Bowman-Perrott, 2009; Bowman-Perrott, Greenwood, & Tapia, 2007; Cavanaugh, Heward, & Donelson, 1996; Kern, Bambara, & Fogt, 2002; King-Sears, Mercer, & Sindelar, 1992; Mastropieri, Emerick, & Scruggs, 1988; Mastropieri et al., 1998; Mastropieri et al., 2006; McCarthy, 2005; Scruggs, Mastropieri, & Sullivan, 1994). However, the Bowman-Perrott (2009) article was excluded because it reported on the same data set as the Bowman-Perrott et al. (2007) article. Therefore, a total of 10 articles that included 11 studies (group, n = 8; single subject, n = 3) were analyzed.

Coding

For the purpose of analysis, the articles were then coded for seven criteria. Random assignment indicated whether the subjects were randomly assigned to experimental groups (0 = non-random, 1 = random, 2 = single subject). EBD broken down indicated whether the results of students with EBD were analyzed separately (0 = no, 1 = yes). Grade level of subjects was identified and coded for each study. Intervention types were coded as either inquiry instruction, supplemental-mnemonic instruction, or supplemental-non-mnemonic instruction (0 = inquiry instruction, 1 = mnemonic instruction, 3 = supplemental other than mnemonics). Intervention duration was recorded in number sessions. Intervention location was coded as regular education or special education (0 = regular education, 1 = special education). Dependent measures were coded as immediate (i.e., measure taken right after instruction), delayed (i.e., measure taken after a delay between instruction and assessment), generalization (i.e., measure that did not test the same content as that covered in the instructional sessions), or behavior (i.e., non-academic measure of students engagement or disengagement during instructional sessions; 0 = immediate, 1 = delayed, 2 = generalization, 3 = behavior).

The studies (n = 11) in the final sample were concurrently coded by four raters. Any disagreements were reconciled during the coding sessions with the reconciled codes used in the subsequent analyses.

ES Calculations

Group studies

For the group studies, the standardized mean difference between treatment and control group was calculated. Depending on the statistical information provided in the article, the standardized mean difference was calculated in one of the following four ways (Therrien et al., 2011): (a) Using treatment and control groups’ means and standard deviations, (b) using t-values from independent t test results, (c) using F-values from analysis of variance (ANOVA) results, (d) using the mean square errors (MSE) and pretest–posttest correlations for calculating the pooled SD and then using the standard formula listed in (a) above to determine the ES for analysis of covariance (ANCOVA) results. Missing correlations were estimated using a conservative estimate of .7 as higher correlations lead to lower ESs and all known correlations were in the .4 to .5 range. Across group studies, a total of 21 ESs were calculated.

After the standardized mean differences were calculated, each was scaled to Hedges’s g to account for the overestimation that occurs when calculating ES using studies with small sample sizes (Hedges, 1981). Standard errors (SE) and weights (w) were determined so that an accurate model of the effects in the analysis could be produced (Lipsey & Wilson, 2001).

Finally, if a study had more than one ES per dependent measure category (e.g., two immediate dependent measures), the smaller ES value was used. Although we report the average overall, immediate, delayed, generalization, and behavioral ESs, the most conservative immediate measure was used to analyze the specific study components (e.g., intervention type).

Single-subject studies

To evaluate the efficacy of the single-subject research studies, we calculated two indices: the percentage of non-overlapping data (PND) and the percentage of all non-overlapping data (PAND). PND represents the percent of the intervention data points that surpass the highest baseline data point (Scruggs, Mastropieri, & Casto, 1987). To calculate PND, it is necessary to determine the highest (or lowest depending on the intervention) data point in the baseline phase (Phase A) and how many data points in the intervention phase (Phase B) exceed this point. The PND for the subject is the percent of data points in the intervention phase that are higher than the highest data point in the baseline phase. When evaluating PNDs, Scruggs et al. (1987) suggest that the most effective interventions have a PND greater than 70, mildly effective interventions have a PND between 50 and 70, and interventions with no observable differences have a PND less than 50.

The percent of all non-overlapping data (PAND) explains the percent of all data remaining after removing the minimum number of data points which would eliminate all data overlap between Phase A and Phase B (Parker, Hagan-Burke, & Vannest, 2007). Both PND and PAND examine the non-overlapping data points between phases, but as Parker and colleagues (2007) explain, “PAND uses all data from both phases, avoiding criticism leveled at PND for wastefulness and overemphasis on one unreliable data point” (p. 196). To calculate PAND, the total number of points that overlap are divided by the total number of points in the phases. If these points were removed, the lines connecting the data points in each phase would be completely separate. For the most effective interventions, Parker et al. (2007) found PND and PAND to be very similar with numbers greater than 50 demonstrating moderate effects and numbers closer to 100 showing the most effective interventions. Finally, to calculate the study indices, we calculated PND and PAND for each subject within each study. Then we took the average of the subject PNDs and PANDs to give us the ESs for the study.

Analysis

Usually in meta-analyses, once ESs are calculated, single and group study results are scaled to a comparable statistic and statistical techniques such as modified weighted regression (Hedges & Olkin, 1985) or Hedges (1982) analog to ANOVA are conducted. In this situation, however, techniques such as these were inappropriate due to the limited number of studies included in the analysis. Instead, we utilized an approach similar to Slavin’s (1986) Best Evidence Synthesis by investigating the research questions by examining the weighted means and confidence intervals for group studies and the mean PAND and mean PND indices (both indices separately) from each single-subject study. Moreover, to provide a deeper contextual perspective of the studies reviewed, we describe the studies descriptively by instructional type (inquiry, mnemonic, supplemental instruction non-mnemonic).

Results

Overall Effects

The purpose of this article was to examine classroom-based science instruction for students with EBD. A total of 11 studies published between 1980 and 2010 were reviewed. Across all group studies using the most conservative immediate, delayed, generalization, and behavior measure for each study, a mean ES of .471 (SE = .077) was calculated for the group studies (n = 8) indicating the interventions as a whole had at least a small to moderate positive effect on students’ with EBD achievement in science (small effect, ES = below .50; medium effect, ES = .50-.80; and large effect, ES = above .80; Cohen, 1988). The single-subject studies were not aggregated due to the small number of studies (n = 3), but instead were analyzed individually at the study level. For these studies, a PND and a PAND range of 50 to 100 was obtained indicating a minimal to a high effect.

In addition to the overall ES, ESs were also calculated for immediate, delayed, and generalization dependent measures. This procedure produced a total of 22 ESs. Immediate measures accounted for 64% of the dependent measures implemented across studies. The mean ES across immediate measures was .567. Delayed measures accounted for 32% of the dependent measures utilized. The mean ES for the delayed measures was .820. Two generalization measures were implemented in the studies reviewed. The mean ES for the generalization measures were .254. Only one group study (McCarthy, 2005) implemented a behavioral measure, which obtained an ES of .080. Descriptions of the dependent variables are reported in Table 1.

Table 1.

Characteristics of Qualifying Studies.

Article	Sample			Dependent variables	Effect size
	Participants	Grades	Assignment		ES (SE) / PND
Inquiry (n = 4)
Bay, Staver, Bryan, and Hale (1992)	107 total 6 with EBD	4–6	Randomly assigned to conditions	Immediate post-test	2.519 (1.004)
Data provided to look only at students with EBD	4 in discovery	Students were instructed in small groups outside the general education setting	20 questions
	2 in direct instruction		True and false
			Multiple choice
			Matching
				Delayed post-test	1.712 (.851)
				Same as immediate
				Administered 2 weeks later
				Generalization test	−0.330 (.699)
				2 weeks later
Kern, Bambara, and Fogt (2002)	6 total	Ages 13–14	ABAB design	Percentage of time engaged observers rotated between 10-s intervals for 1 min per student for an average of 28 min per lesson	PND 87.5%
	All students with EBD	Middle school alternative setting	Classes acted as their own control		PAND 75%
				Percentage of time destructive behavior did not occur	PND 100%
					PAND 40%
Mastropieri et al. (1998)	75 total	4 General education classroom	Non-randomly by teacher to conditions	Immediate post-tests
Data provided to look at all students with disabilities	1 student with EBD			20-point multiple choice test	.531 (.246)
				10-item conceptual understanding test	2.004 (.292)
McCarthy (2005)	18 total	Jr. High Self-contained; partial hospitalization	Students were not randomly assigned; teachers were randomly assigned to by coin flip	Immediate post-tests
	All students with EBD			16-point multiple choice test	1.905 (.550)
				20-point short answer test	3.866 (.785)
				Two performance assessments	2.028 (.562)
				Behavior	.080 (.449)
				10-point scale
				Awarded the presence of behaviors during each class for each student
Mnemonic (n = 2)
King-Sears (1992)	37 total	6–8	Randomly assigned to conditions by teacher	Immediate post-test	1.736 (.497)
No data provided to look only at students with EBD	7 with EBD	Special education classroom		Free recall of 48 definitions and matching during Week 4	In favor of imposed keyword condition
	1 student in the imposed keyword mnemonic
	5 in the induced keyword mnemonic group
Mastropieri, Emerick, and Scruggs (1988)	7 total	1–4	Counter-balanced design	Immediate post-test
	All students with EBD	Self-contained classroom		Vocabulary recall	0.930 (.412)
				Delayed post-test (same as immediate measure)	1.091 (0.439)
Supplemental non-mnemonic (n = 4)
Bowman-Perrott, Greenwood, and Tapia (2007)	11 total	9–12	Alternating Treatment Design	Immediate post-test (descriptions of the weekly tests were not included)	PND 50%
	All students with EBD	Alternative school setting			PAND 72.2%
				Number of intervals on-task Classrooms 1 and 2 combined 30 s time sampling procedure for roughly 26 min per student	PND 100%
					PAND 100%
Cavanaugh, Heward, and Donelson (1996)	23 total	9	Alternating Treatment Designs	Next day tests using passive response as baseline and active response as intervention
No data provided to look only at students with EBD	8 students with EBD	Supplemental at-risk program
				1 × 12 review	PND 57.1%
					PAND 81.3%
				2 × 12 review	PND 75%
					PAND 85.7%
				2 × 6 review	PND 57.1%
					PAND 71.4%
Mastropieri et al. (2006)	213 total	8	Classes matched and then assigned randomly	Immediate
Data provided to look at all students with disabilities	7 students with EBD	General education classrooms	Each lead teacher taught at least 1 of each condition	Science content post-test	.330 (.137)
				Generalization
				High-stakes science test at end of the year	.276 (.137)
Scruggs, Mastropieri, and Sullivan (1994)	36 total	4–5	Students assigned randomly to 1 of 3 conditions	Study A (experimenter provided vs. control)
Data provided to look at all students with disabilities	1 student with EBD	Special education		Immediate
				Factual recall	.475 (.400)
				Causal explanation	4.423 (.750)
				Delayed (1 week after instruction)
				Factual recall	.289 (.521)
				Causal explanation	1.424 (.588)
				Study B (student generated vs. control)
				Immediate
				Factual recall	.649 (.405)
				Causal explanation	4.703 (.785)
				Delayed (1 week after instruction)
				Factual recall	.617 (.548)
				Causal explanation	1.499 (.615)

Note. Some ESs are different than ESs reported in Therrien et al. (2011) due to the following: When able, we calculated ESs using students with EBD only; and unequal sample sizes were taken into account in this analysis. EBD = emotional/behavior disorder; ES = effect size; PND = percent of non-overlapping data points; PAND = percent of all non-overlapping data.

Comparison of Instructional Strategies

Inquiry instruction

A total of four studies compared traditional instruction (e.g., textbook based with lecture) to inquiry instruction. Each inquiry treatment condition included (a) hands-on instructional activities and (b) teacher directed experiments conducted by the student. The group studies (n = 3) that fell into this category produced an immediate mean ES of .844, and the one single subject produced a mean PND of 93.75% and a PAND of 57.5%. The ESs for each individual inquiry study are reported in Table 1.

Two of the four studies in this category compared inquiry instruction to a control condition that relied heavily on the science textbook for content knowledge acquisition. In addition to using the text, Mastropieri and colleagues (1998) also incorporated teacher presentation, guided practice, class discussion, and videotapes to instruct students in the textbook condition. While in the inquiry instruction condition, Mastropieri and colleagues modified the curriculum and materials to meet the needs of the students and students were given time for group work, observation, and reflection. The intervention took place 3 days a week for 7 weeks, and resulted in students in the activities-based classroom scoring significantly higher than students in the textbook condition on the academic multiple choice, performance, and elaboration post-tests.

Traditional textbook instruction was compared with inquiry-based science instruction in McCarthy (2005). Similar to Mastropieri et al. (1998), the textbook condition included teacher demonstration, textbook discussion, and independent practice. The inquiry condition included a time for review and small group experiments. Large group discussion over experiment concepts and results was also a key feature of the treatment condition. Each of the intervention sessions lasted 45 min and took place 3 days a week for 8 weeks. Results indicated that students receiving the hands-on, inquiry-based, instruction performed significantly higher than students in the textbook condition on a hands-on and short answer science content assessment. There were no significant differences between groups on the multiple choice science measure or in regard to student behavior.

A single-subject design was used in Kern and colleagues (2002) to compare a lecture/written assignment instructional style (baseline phase) to a hands-on inquiry approach (intervention phase). Along with the hands-on materials used in the intervention, the treatment condition also focused on providing students multiple opportunities to choose instructional activities. A classroom behavior system was used throughout the study. The intervention took place for 11 days, and each session was 40 min. Findings showed that when these curricular modifications were in place, students had increased levels of engagement and exhibited fewer problem behaviors.

The last inquiry study (Bay et al., 1992) included a direct instruction condition, which used graphic organizers, teacher modeling, and guided/independent practice worksheets. This was compared to a discovery teaching condition (i.e., inquiry) where students were taught using hands-on materials, the relationships between concepts were highlighted, experiments were conducted, and students made and tested predictions. Each condition consisted of five sessions lasting 40 to 60 min each. Students in both conditions performed equally well on the immediate performance-based post-test, but students in the discovery teaching condition outperformed students in the comparison condition on similar performance-based delayed and generalization measures.

Supplemental-mnemonic instruction

Mnemonic studies (n = 2) obtained an ES of 1.258 (see Table 1 for the ES from the individual mnemonic studies). Both these group studies used keyword mnemonics: A keyword and illustration combination are taught and in turn used by students as a prompt to remember the definition of the targeted vocabulary word (Atkinson, 1975).

In the first mnemonic strategy study (King-Sears et al., 1992), 48 vocabulary words were taught in one of three conditions: systematic teaching (control), induced keyword instruction, or imposed keyword instruction. In the systematic condition, students were presented with vocabulary words, asked to say the word out loud, and then were presented with the picture prompt to help them remember the word. In the induced keyword condition, the student developed the mnemonic after they were instructed on how to create mnemonics, whereas in the imposed keyword condition, the teacher provided the mnemonic. The intervention was conducted during a 4-week period, which resulted in students in both the induced and imposed keyword instructional groups making significant progress on weekly and cumulative vocabulary post-tests, over their counterparts in the systematic teaching group.

The second study examining the use of keyword mnemonics was conducted by Mastropieri, Emerick, and Scruggs (1988). During the first week of the study, half of the students received traditional vocabulary instruction using index cards with definitions on one side and the vocabulary word on the other side. The other half of the students were taught using keyword mnemonics. In the second week, new science content was introduced and the groups were taught using the alternate method. The science vocabulary instruction took place 3 days a week for 2 weeks. Results from the study indicated that students in the keyword mnemonics condition significantly outperformed students receiving traditional vocabulary instruction on daily and cumulative vocabulary tests. Students in the treatment condition also scored significantly higher than the control group on a delayed vocabulary measure indicating high levels of vocabulary retention.

Supplemental non-mnemonic instruction

The five supplemental studies that make up this category include peer-assisted learning (n = 2) studies and non-mnemonic strategies geared to ensure students retain science facts (n = 3). The supplemental group studies (n = 3) produced a mean ES of .374. The mean PND from the single-subject studies was 69% and the mean PAND from the single-subject studies was 83% (see Table 1).

Cavanaugh et al. (1996) used a single-subject approach to examine active vocabulary review using response cards to a passive vocabulary review method where the teacher read the definition out loud to students. In this study, 12 science facts were randomly assigned to two conditions. Each condition was compared using the following: 1 by 12 format (e.g., 12 lesson points were reviewed once each), 2 by 12 format, and 2 by 6 format (e.g., 6 points were reviewed twice). The passive review was used during the baseline period and active review was used as the intervention phase. The review process took place during a 30-min science lesson. The results of the study indicated that students in the active/response card review session scored higher on both the next-day and weekly vocabulary assessments.

Scruggs et al. (1994) also measured students’ retention of science facts. This study was unique in that it produced two separate ESs: one for the condition where an explanation of a given science concept was provided by the experimenter and one for a condition where the student generated his or her own explanation of the fact. Each of the treatment conditions was compared to a control condition where the students were asked to remember 14 facts about animals that were stated by the teacher. One 15-min session was taught using one of the three conditions. “Students in the student-generated explanation condition scored descriptively highest on all three measures” (Scruggs et al., 1994, p. 454). Results from the two treatment conditions (i.e., student generated and experimenter-provided explanation) were not significantly different on the immediate recall and explanation measures. However, both treatment conditions significantly outperformed the control condition on the delayed measure.

The last two studies that met our criteria examined the impact of peer tutoring/assistance on science achievement. In the first study (Bowman-Perrott et al., 2007), students were paired with a partner. Each student was given the opportunity to be both a tutor and a tutee. The intervention sessions consisted of vocabulary review and either comprehension practice or study guide completion. During the intervention, a behavior monitoring system was also used in conjunction with the classroom token economy. The tutoring strategy lasted 4 weeks and was implemented 3 days a week for 30-min sessions. When looking at the two classrooms that implemented this intervention, only one classroom saw a slight increase in science test scores among students. But when looking at the behavior measure, students in both classrooms showed an increase in the amount of their time spent on-task.

The impact peers can play on science achievement was also studied in Mastropieri et al. (2006). In this study, a traditional lecture-based instructional method was compared with a peer-assisted instructional method. The first group used lecture, notes provided by the teacher, experiments, and worksheets. In the peer-assisted condition, the same teacher presentation as in the comparison condition was used, but differentiated activities replaced worksheet time. This study took place during a 12-week science unit. The researchers concluded that differentiated learning using peer partners had a positive effect on both the science content post-tests as well as on the state high-stakes tests.

Comparison of Study Characteristics

In addition to aggregating studies based on the instructional focus of the intervention, our coding procedure allowed us to look at the following moderating variables: location of the intervention (e.g., general education setting), duration of the intervention, grade level of students, the assignment of participants to treatments (e.g., randomly assigned), and if the results were reported separately for students with EBD. The immediate ES means, SE, and confidence intervals for each of these variables are reported in Table 2.

Table 2.

Mean Effect Sizes, Standard Errors and Confidence Intervals for Categories Analyzed.

		95% Confidence interval
Category designation	Mean ES (SE)	Low	High
Overall ES—Dependent measure
Immediate	.567 (.102)	.367	.767
Delayed	.820 (.271)	.288	1.352
Generalization	.254 (.135)	−.010	.517
Behavior	.080 (.449)	−.800	.960
Intervention—Instructional focus
Inquiry	.844 (.219)	.415	1.274
Supplemental-mnemonic	1.258 (.317)	.636	1.880
Supplemental non-mnemonic	.374 (.124)	.131	.616
Study characteristic—Location
Regular education	.378 (.120)	.142	.613
Special education	1.062 (.194)	.682	1.442
Study characteristic—Duration
Short	.691 (.174)	.349	1.032
Long	.503 (.126)	.256	.750
Study Characteristic—Grade Level
Kindergarten through 5th grade	.768 (.160)	.454	1.081
6th through 12th grade	.430 (.133)	.170	.690
Study characteristic—Random assignment
Randomly assigned	.942 (.207)	.536	1.348
Not randomly assigned	.447 (.117)	.217	0.677
Study characteristic—Results for students with EBD reported separately
Reported separately	1.401 (.313)	.787	2.015
Not reported separately	.468 (.108)	.257	.680

Note. This table contains the effects for group studies only. Effect sizes were calculated using the most conservative immediate dependent measures. The overall effect size was calculated using all dependent measures. ES = effect size; SE = standard error.

Discussion

Similar to the Therrien and colleagues (2011) review that examined classroom science programs for students with LD, the vast majority of dependent measures in this review were immediate or delayed assessments of science knowledge or skill acquisition. Overall, there was a large impact on these measures (immediate ES = .567; delayed ES = .820). Measures aggregated within this category typically involved assessments of science factual knowledge including multiple choice and short answer tests and oral recall. The dependent measures for studies examining inquiry instruction consisted of multiple choice, matching, true/false, and short-answer questions. The mnemonic studies contained dependent measures examining the percent of immediate definition recall and some matching items. The supplemental non-mnemonic studies included dependent measures that assessed explanations and active responses from students. Although the sizable impact on delayed measures is encouraging, it should be noted that the reported length of time between instruction and delayed post-test was extremely short (i.e., 2 weeks or less between instruction and assessment). Furthermore, we must take into consideration that not all studies reported delayed measures thus potentially inflating the average ES of the delayed measures. Additional research is needed to explore the efficacy of science interventions on students’ knowledge and skill maintenance.

Since No Child Left Behind (NCLB) mandates science to be a required assessment area, evaluating the impact of science instruction on generalized achievement measures is critical. However, only two studies reported generalization measures, and the mean ES increase in achievement was small (ES = .254). Results from one study (Mastropieri et al., 2006) were promising as it indicated that peer-assisted science instruction utilizing tiered material has the potential to impact student achievement on an end of the year science assessment.

Although the studies included in this review were specifically selected because of their inclusion of students with EBD, there were a surprising limited number of articles (i.e., 1 group and 2 single subject) that included behavioral measures. Results on behavioral measures were mixed. Bowman and colleagues (2007) reported strong results (PND = 100; PAND = 100) for class-wide peer tutoring (CWPT) on students’ on-task behavior. Kern and colleagues (2002) obtained positive results on increasing student engagement (PND = 87.5; PAND = 75) and reducing destructive behavior (PND = 100; PAND = 40) via hands-on instruction and providing students choices within instructional activities. McCarthy (2005), however, reported a non-significant difference (ES = .080) on a behavioral scale between students in inquiry science instruction and those in traditional lecture/textbook instruction.

Furthermore, all of the studies that collected behavioral data were conducted within special education placements (e.g., alternative schools) and not within general education science classroom. More research is needed to ascertain if various approaches to science instruction can positively impact the behavior of students with EBD within traditional classrooms. This question is particularly pertinent to the investigation of inquiry instruction as proponents contend that the applied nature of inquiry instruction increases student engagement while others contend that the lack of structure sometimes associated with inquiry instruction can be problematic for students with EBD.

Comparison of Instructional Strategies

All content area instruction, but particularly science, varies greatly in instructional practices, the amount of content covered, and often in the diverse nature of student learners included in the classrooms. As the comparisons of instructional strategies are examined, it is critical to keep the nature of the instructional setting, content, and the student population at the forefront of the discussion.

Inquiry instruction

Across all inquiry studies, a large ES (group ES = .844; mean PND of 93.75% and a mean PAND of 57.5%) was obtained. While showing promising results, it is too early to draw conclusions about the potential impact of inquiry-based approaches on students’ with EBD achievement. Although we examined a 30-year period, only 4 experimental studies that examined inquiry approaches for students with EBD were found. Furthermore, across all these studies, only 35 students with EBD were included. Additional studies need to be conducted to ascertain if and under what conditions inquiry instruction positively impacts students’ with EBD achievement.

In addition, definitional issues of the type of instructional components and supports contained within effective inquiry for students with EBD needs to be examined. Studies on inquiry approaches have been criticized for often lacking clear operational definitions of the intervention that can be replicated or applied by others outside of the respective research teams (Klahr & Li, 2005). Inquiry interventions evaluated here are best described as structured inquiry approaches as the experiments conducted were directed by the teacher. Additional research is warranted to determine what instructional components are essential for students with EBD within a science inquiry approach.

Mnemonics instruction

Apart from the studies that examined inquiry-based instruction, the remaining articles investigated supplemental supports aimed at improving the science achievement of students with EBD. Similar to previous literature reviews (e.g., Mastropieri et al., 1998; Therrien et al., 2011), supplemental instruction with mnemonics resulted in a significant improvement (ES = 1.258) in the acquisition and retention of science factual knowledge for students with EBD. The consistently positive and strong results reported for mnemonic instruction within this review and across numerous other studies (e.g., Mastropieri & Scruggs, 1994; Scruggs, Mastropieri, Berkeley, & Marshak, 2010) and reviews (e.g., Gajria, Jitendra, Sood, & Sacks, 2007) provide the solid and extensive evidence needed to recommend the use of mnemonic instruction to increase students’ factual knowledge.

Supplemental non-mnemonics instruction

Along with mnemonic instruction, the efficacy of four other supplemental interventions was examined. Instructional approaches aggregated within this category included peer tutoring, response card instruction, and the inclusion of casual explanations to improve acquisition and retention of science facts. Across these studies, an overall group mean ES of .374 and a mean PND of 69% (PAND of 83%) was obtained. Only peer tutoring was examined in more than one study. Results from these studies indicate emerging evidence of the potential effectiveness of peer tutoring with students with EBD in science. Bowman-Perrott and colleagues (2007) reported the improvement of student behavior and achievement within a peer tutoring modeling while Mastropieri and colleagues (2006) reported improvement in student achievement on a high-stakes end-of-the-year achievement test after being involved in a peer tutoring program that used scaffolded material. Further research is warranted to examine the use of peer tutoring within science classroom as well as to examine the other instructional approaches aggregated within this category.

Comparison of Study Characteristics

Location

Along with type of instruction, the instructional location and intervention duration were examined. Location of instruction significantly impacted ES magnitude with interventions implemented within special education settings obtaining an overall group mean ES of 1.062 and those implemented in general education classrooms obtaining an overall mean ES of .378. All single-subject studies were conducted in special education settings. The high level of interobserver agreement of treatment implementation (100%) and observed behaviors (83–99.3) found in the single-subject designs may suggest that fidelity of implementing the intervention is higher in special education settings. Although speculative, the higher score obtained by students with EBD in special education settings as compared to general education settings may indicate the need for the implementation of systematic and intensive behavioral supports for inquiry approaches to be effective. These supports are often found in specialized placements for students with EBD and typically are missing within the general education classroom.

Duration

The duration of the interventions across group and single-subject studies averaged 15.5 sessions with a range of 1 to approximately 60 sessions. Overall group ESs suggest short-duration (i.e., 15 min or less) interventions (ES = .691) were more effective than long-duration (i.e., 30 min or more) interventions (ES = .503). This result is likely due to the type and timing of the assessment measures utilized. Short duration studies implemented immediate experimenter designed measures that were very sensitive to student growth. While longer duration studies were more likely to implement non-experimenter generated delayed and/or generalization measures that were less sensitive to student growth. Therefore, instead of being an indication of the superiority of short duration intervention, the higher ES for short duration interventions is likely due to the differences between the measures used in the short- and long-duration studies.

Student characteristics

An in-depth examination of student characteristics was not possible due to the inconsistent reporting of demographic information across the studies reviewed. An examination of results based on student grade level was possible. Only students in Grades 4 through 9 were well represented in the studies reviewed. An ES comparison based on grade level indicated that younger students (i.e., Grades 1–5) obtained a higher mean ES (.768) than the ES (.430) obtained by older students (i.e., Grades 6–12). This result aligns with the findings reported for students with LD (Therrien et al., 2011) and indicates that it may be easier to positively impact science achievement of students with EBD at younger grades.

Study characteristics

Two variables associated with group study design were investigated: assignment of students to conditions and whether students with EBD results were reported separately. Studies that assigned students randomly to conditions had a higher mean ES (.942) than studies that did not assign students randomly (ES = .447). Only three of the eight group studies reported results separately for students with EBD. A comparison of mean ES based on whether the results for students with EBD were reported separately indicates a higher ES for studies that reported results only for students with EBD (ES = 1.401) compared with studies that did not report the results for students with EBD separately (ES = .468). This finding provides preliminary justification that students with EBD may be as or more responsive to effective science instruction than their non-disabled peers.

Limitations

There are four limitations to the conclusions of this analysis. First, although a 30-year time period was examined, only 11 studies that involved 72 students with EBD were found and included in this analysis. Significant additional research is therefore needed to fully ascertain the impact of classroom-based science instruction on students with EBD achievement. Second, the measures implemented within the studies reviewed were, in general, experimenter-generated proximal measures of student achievement. Most of the included dependent measures did not assess for understanding but rather factual recall, and only a few studies examined distal measures of student achievement. Future studies must examine the impact of classroom science instruction on distal measures (e.g., high-stakes tests) of science achievement. Third, only a few studies examined the impact of classroom science instruction on non-academic measures such as student engagement. Future studies must assess the impact of science instruction on students’ with EBD behavior and task engagement. Conducting these studies is particularly important for instructional approaches such as science inquiry that tend to be less structured than traditional lecture and textbook approaches to science instruction. Finally, the majority of the studies included in this review examined the effects of the interventions on small populations of students with researchers as the primary interventionists. To generalize these findings, it is important that future research replicates these findings with larger sample sizes and examines the feasibility and fidelity of teachers implementing these practices within their own classrooms.

Conclusion

The overall results indicate that classroom-based science instruction has the potential to increase students with EBD achievement on proximal measures of science content and process knowledge. Evidence is strongest for the use of mnemonics to increase the acquisition and retention of students’ with EBD science factual knowledge. Preliminary evidence indicates that students with EBD may benefit from inquiry methods that include the implementation of hands-on experiments conducted under the direction of the teacher. Additional research is needed to verify this preliminary finding and to determine what instructional components are needed within a science inquiry approach for students with EBD to be successful both behaviorally and academically.

Footnotes

Authors’ Note

The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grants R305A090094 and R305B10005 to The University of Iowa.

References

American Association for the Advancement of Science. (1997). Resources for science literacy: Professional development. New York, NY: Oxford University Press.

Atkinson

R. C.

(1975). Mnemotechnics in second-language learning. American Psychologist, 30, 821–828. doi:10.1037/h0077029

Bay

Staver

J. R.

Bryan

Hale

J. B.

(1992). Science instruction for the mildly handicapped: Direct instruction versus discovery teaching. Journal of Research in Science Teaching, 29, 555–570. doi:10.1002/tea.3660290605

Bowman-Perrott

(2009). Classwide peer tutoring: An effective strategy for students with emotional and behavioral disorders. Intervention in School and Clinic, 44, 259–267. doi:10.1177/1053451208330898

Bowman-Perrott

Greenwood

C. R.

Tapia

(2007). The efficacy of CWPT used in secondary alternative school classrooms with small teacher/pupil ratios and students with emotional and behavioral disorders. Education and Treatment of Children, 30, 65–87. doi:10.1353/etc.2007.0014

Bradley

Henderson

Monfore

D. A.

(2004). A national perspective on children with emotional disorders. Behavior Disorders, 29, 211–223.

Brigham

F. J.

Scruggs

T. E.

Mastropieri

M. A.

(1992). Teacher enthusiasm in learning disabilities classrooms: Effects on learning and behavior. Learning Disabilities Research & Practice, 7, 68–73.

Cavanaugh

R. A.

Heward

W. L.

Donelson

(1996). Effects of response cards during lesson closure on the academic performance of secondary. Journal of Applied Behavior Analysis, 29, 403–406. doi:10.1901/jaba.1996.29-403

Cawley

Hayden

Cade

Baker-Kroczynski

(2002). Including students with disabilities into the general education science classroom. Exceptional Children, 68, 423–435.

10.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

11.

Dunlap

Childs

(1996). Intervention research in emotional and behavioral disorders: An analysis of studies from 1980 to 1993. Behavioral Disorders, 21, 125–136.

12.

Ervin

R. A.

DuPaul

G. J.

Kern

Friman

P. C.

(1998). Classroom-based functional and adjunctive assessments: Proactive approaches to intervention selection for adolescents with attention deficit hyperactivity disorder. Journal of Applied Behavior Analysis, 31, 65–78. doi:10.1901/jaba.1998.31-65

13.

Furlong

M. J.

Morrison

G. M.

Jimerson

S. R.

(2004). Externalizing behaviors of aggression and violence and the school context. In Rutherford

R. B.

Jr. Quinn

M. M.

Mathur

S. R.

(Eds.), Handbook of research in emotional and behavioral disorders (pp. 243–261). New York, NY: Guilford.

14.

Gajria

Jitendra

A. K.

Sood

Sacks

(2007). Improving comprehension of expository text in students with LD: A research synthesis. Journal of Learning Disabilities, 40, 210–225.

15.

Gaylord

Quinn

McComas

Lehr

(Eds.). (2005). Impact: Feature issue on fostering success in school and beyond for students with emotional/behavioral disorders Vol. 18(2). Minneapolis, MN: University of Minnesota, Institute on Community Integration. Retrieved from http://ici.umn.edu/products/impact/182/default.html

16.

Hedges

L. V.

(1981). Distribution theory for Glass’s estimator of effect size and related estimators. Journal of Educational Statistics, 6, 107–128. doi:10.3102/10769986006002107

17.

Hedges

L. V.

(1982). Fitting categorical models to effect sizes from a series of experiments. Journal of Educational Statistics, 7, 199–137. doi:10.3102/10769986007002119

18.

Hedges

L. V.

Olkin

(1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press.

19.

Individuals with Disability Education Act Amendments of 1997 [IDEA]. (1997). Retrieved from http://thomas.loc.gov/home/thomas.php

20.

Kern

Bambara

Fogt

(2002). Class-wide curricular modification to improve the behavior of students with emotional or behavioral disorders. Behavioral Disorders, 27, 317–326.

21.

King-Sears

Mercer

C. D.

Sindelar

P. T.

(1992). Toward independence with keyword mnemonics: A strategy for science vocabulary instruction. Remedial and Special Education, 13, 22–33. doi:10.1177/074193259201300505

22.

Klahr

(2005). Cognitive research and elementary science instruction: From the laboratory, to the classroom, and back. Journal of Science Education and Technology, 14, 217–238. doi:10.1007/s10956-005-4423-5

23.

Lane

Wehby

Barton-Arwood

(2005). Students with and at-risk for emotional and behavioral disorders: Meeting their social and academic needs. Preventing School Failure, 49, 6–9.

24.

Lane

K. L.

(2004). Academic instruction and tutoring interventions for students with emotional/behavioral disorders: 1990 to present. In Rutherford

R. B.

Quinn

M. M.

Mathur

S. R.

(Eds.), Handbook of research in emotional and behavioral disorders (pp. 462–486). New York, NY: Guilford.

25.

Lipsey

M. W.

Wilson

D. B.

(2001). Practical meta-analysis. Thousand Oaks, CA: SAGE.

26.

MacDougall

Schnur

Berger

Vernon

(1981). The use of activity centered science activities to facilitate the mainstreaming of elementary school children with special needs. Science Education, 65, 467–475. doi:10.1002/sce.3730650502

27.

Mastropieri

M. A.

Emerick

Scruggs

T. E.

(1988). Mnemonic instruction of science concepts. Behavioral Disorders, 14, 48–56.

28.

Mastropieri

M. A.

Scruggs

T. E.

(1992). Science for students with disabilities. Review of Educational Research, 62, 377–411. doi:10.3102/00346543062004377

29.

Mastropieri

M. A.

Scruggs

T. E.

(1994). Applications of mnemonic strategies with students with mild mental disabilities. Remedial & Special Education, 15, 34–43.

30.

Mastropieri

M. A.

Scruggs

T. E.

(1998). Constructing more meaningful relationships in the classroom: Mnemonic research into practice. Learning Disabilities Research and Practice, 13, 138–145.

31.

Mastropieri

M. A.

Scruggs

T. E.

Butcher

(1997). How effective is inquiry learning for students with mild disabilities? Journal of Special Education, 31, 199–211. doi:10.1177/002246699703100203

32.

Mastropieri

M. A.

Scruggs

T. E.

Mantzicopoulos

Sturgeon

Goodwin

Chung

(1998). “A place where living things affect and depend on each other”: Qualitative and quantitative outcomes associated with inclusive science teaching. Science Education, 82, 163–179. doi:10.1002/(SICI)1098-237X(199804)82:2<163::AID-SCE3>3.0.CO;2-C

33.

Mastropieri

M. A.

Scruggs

T. E.

Norland

J. J.

Berkeley

McDuffie

Tornquist

E. H.

Connors

(2006). Differentiated curriculum enhancement in inclusive middle school science: Effects on classroom and high-stakes tests. Journal of Special Education, 40, 130–137. doi:10.1177/00224669060400030101

34.

Mastropieri

M. A.

Scruggs

T. E.

Whittaker

M. E. S.

Bakken

J. P.

(1994). Applications of mnemonic strategies with students with mild mental disabilities. Remedial and Special Education, 15, 34–43. doi:10.1177/074193259401500106

35.

McCarthy

C. B.

(2005). Effects of thematic-based, hands-on science teaching versus a textbook approach for students with disabilities. Journal of Research in Science Teaching, 42, 245–263. doi:10.1002/tea.20057

36.

McFarland

Shepard

(1995). Oral and written compositions of students with learning disabilities in the content area of science. Learning Disabilities: A Multidisciplinary Journal, 6, 13–18.

37.

National Research Council. (2012). A framework for K-12 science education: Practices, crosscutting concepts, and core ideas. Washington, DC: The National Academies Press.

38.

National Science Teachers Association Board of Directors. (2004). National science teachers’ association position statement: Science inquiry. Retrieved from http://www.nsta.org/about/positions/inquiry.aspx

39.

Nolet

Tindal

(1994). Instruction and learning in middle school science classes: Implications for students with disabilities. Journal of Special Education, 28, 166–187. doi:10.1177/002246699402800204

40.

Okilwa

N. S. A.

Shelby

(2010). The effects of peer tutoring on academic performance of students with disabilities in grades 6 through 12: A synthesis of the literature. Remedial and Special Education, 31, 450–463. doi:10.1177/0741932509355991

41.

Parker

R. I.

Hagan-Burke

Vannest

(2007). Percent of all non-overlapping data (PAND): An alternative to PND. Journal of Special Education, 40, 194–204.

42.

Parmar

R. S.

Deluca

C. B.

Janczak

T. M.

(1994). Investigations into the relationship between science and language abilities of students with mild disabilities. Remedial and Special Education, 15, 117–126. doi:10.1177/074193259401500207

43.

Pomplun

(1997). When students with disabilities participate in cooperative groups. Exceptional Children, 64, 49–58.

44.

Pomplun

(1998). Cooperative groups: Alternative assessment for students with disabilities? The Journal of Special Education, 30, 1–17. doi:10.1177/002246699603000101

45.

Rogevich

M. E.

Perin

(2008). Effects on science summarization of a reading comprehension intervention for adolescents with behavior and attention disorders. Exceptional Children, 74, 135–154.

46.

Ryan

J. B.

Reid

Epstein

M. H.

(2004). Peer-mediated intervention studies on academic achievement for students with EBD: A review. Remedial and Special Education, 25, 330–341. doi:10.1177/07419325040250060101

47.

Scruggs

T. E.

Marsing

(1987). Teaching test-taking skills to behaviorally disordered students. Behavioral Disorders, 13, 240–244.

48.

Scruggs

T. E.

Mastropieri

M. A.

(1993). Current approaches to science education: Implications for mainstream instruction of students with disabilities. Remedial and Special Education, 14, 15–24. doi:10.1177/074193259301400104

49.

Scruggs

T. E.

Mastropieri

M. A.

(1994a). The construction of scientific knowledge by students with mild disabilities. Journal of Special Education, 28, 307–321. doi:10.1177/002246699402800306

50.

Scruggs

T. E.

Mastropieri

M. A.

(1994b). Successful mainstreaming in elementary science classes: A qualitative study of three reputational cases. American Educational Research Journal, 31, 785–811. doi:10.3102/00028312031004785

51.

Scruggs

T. E.

Mastropieri

M. A.

(1995). Science education for students with behavioral disorders. Education and Treatment of Children, 18, 322–334.

52.

Scruggs

T. E.

Mastropieri

M. A.

(2000a). The effectiveness of mnemonic instruction for students with learning and behavior problems: An update and research synthesis. Journal of Behavioral Education, 10, 163–173. doi:10.1023/A:1016640214368

53.

Scruggs

T. E.

Mastropieri

M. A.

(2000b). Mnemonic strategies improve classroom learning and social behavior. Beyond Behavior, 10, 13–17.

54.

Scruggs

T. E.

Mastropieri

M. A.

(2007). Science learning in special education: The case for constructed versus instructed learning. Exceptionality, 15, 57–74. doi:10.1080/09362830701294144

55.

Scruggs

T. E.

Mastropieri

M. A.

Bakken

J. P.

Brigham

F. J.

(1993). Reading versus doing: The relative effects of textbook-based and inquiry-oriented approaches to science learning in special education classrooms. The Journal of Special Education, 27, 1–15. doi:10.1177/002246699302700101

56.

Scruggs

T. E.

Mastropieri

M. A.

Berkeley

S. L.

Marshak

(2010). Mnemonic strategies: Evidence-based practice and practice-based evidence. Intervention In School & Clinic, 46, 79–86. doi:10.1177/1053451210374985

57.

Scruggs

T. E.

Mastropieri

M. A.

Casto

(1987). The quantitative synthesis of single-subject research: Methodology and validation. Remedial and Special Education, 8, 24–33. doi:10.1177/07419325870080020

58.

Scruggs

T. E.

Mastropieri

M. A.

Sullivan

G. S.

(1994). Promoting relational thinking: Elaborative interrogation for students with mild disabilities. Exceptional Children, 60, 450–457.

59.

Scruggs

T. E.

Mastropieri

M. A.

Sullivan

G. S.

Hesser

L. S.

(1993). Improving reasoning and recall: The differential effects of elaborative interrogation and mnemonic elaboration. Learning Disability Quarterly, 16, 233–240.

60.

Scruggs

T. E.

Mastropieri

M. A.

Veit

D. T.

Osguthorpe

R. T.

(1986). Behaviorally disordered students as tutors: Effects on social behavior. Behavioral Disorders, 12, 36–44.

61.

Slavin

R. E.

(1986). Best-evidence synthesis: An Alternative to meta-analytic and traditional reviews. Educational Researcher, 15, 5–11. doi:10.3102/0013189X015009005

62.

Terrell

(2007). STEM occupations: High-tech jobs for a high-tech economy. Occupational Outlook Quarterly, 51(1), 26–33.

63.

Therrien

W. J.

Taylor

J. C.

Hosp

J. L.

Kaldenberg

E. R.

Gorsh

(2011). Science instruction for students with learning disabilities: A meta-analysis. Learning Disabilities Research & Practice, 26, 188–203. doi:10.1111/j.1540-5826.2011.00340.x

64.

Vaughn

Levy

Coleman

Bos

C. S.

(2002). Reading instruction for students with LD and EBD: A synthesis of observation studies. Journal of Special Education, 36, 2–13.

65.

Wagner

Marder

Blackorby

Cameto

Newman

Levine

. . . Sumi

(2003). The achievements of youth with disabilities during secondary school. Menlo Park, CA: SRI International.