Twenty-First Century Adaptive Teaching And Individualized Learning Operationalized As Specific Blends Of Student-Centered Instructional Events: A Systematic Review And Meta-Analysis

Adaptive teaching and individualization for K-12 students improve academic achievement What is the aim of this review? This Campbell systematic review assesses the overall impact on student achievement of processes and methods that are more student-centered versus less student-centered. It also considers the strength of student-centered practices in four teaching domains. Flexibility: Degree to which students can contribute to course design, selecting study materials, and stating learning objectives. Pacing of instruction: Students can decide how fast to progress through course content and whether this progression is linear or iterative. Teacher's role: Ranging from authority figure and sole source of information, to teacher as equal partner in the learning process. Adaptability: Degrees of manipulating learning environments, materials, and activities to make them more student-centered. Teaching in K-12 classrooms involves many decisions about the appropriateness of methods and materials that both provide content and encourage learning. This review assesses the overall impact on student achievement of processes and methods that are more student-centered versus less student-centered (and thus more teacher-centered, i.e., more under the direct control of a teacher). It also considers in which instructional dimensions the application of more of these student-centered practices is most appropriate, and the strength of student-centered practices in each of four teaching domains. This review presents evidence from 299 studies (covering 43,175 students in a formal school setting) yielding 365 estimates of the impact of teaching practices. The studies spanned the period 2000–2017 and were mostly carried out in the United States, Europe, and Australia. What is the overall average effect of more versus less student-centered instruction on achievement outcomes? Which demographic variables moderate the overall results? More student-centered instructional conditions have a moderate positive effect on student achievement compared to less student-centered. Which dimensions of instruction are most important in promoting better achievement through the application of more versus less student-centered instruction? Do these dimensions interact? The teacher's role has a significantly positive impact on student achievement; more student-centered instruction produces better achievement. Pacing of instruction/learning—where learners have more choice over setting the pace and content navigation of learning activities—has a significant effect in the opposite direction; i.e., a significantly negative relationship. There is no relationship between adaptability and flexibility and student achievement. There are interactive effects. The teacher's role combined with adaptability produces stronger effects, whereas flexibility (greater involvement of students in course design and selection of learning materials and objectives) has the opposite effect; it reduces the effectiveness of teacher's role on learning outcomes. Special education students perform significantly better in achievement compared to the general population. Three other factors—grade level; Science, technology, engineering, and mathematics (STEM) versus non-STEM subjects; individual subjects—do not have any effect on the impact of the intervention. This review confirms previous research on the effectiveness of student-centered and active learning. It goes further in suggesting the teacher's role promotes effective student-centered learning, and excessive student control over pacing appears to inhibit it. An important element of these findings relates to the significant combination of teacher's role and adaptability, in that it suggests the domain in which the teacher's role should focus. Since adaptability relates to increasing the involvement of students in more student-centered activities, the evidence suggests that instruction that involves activity-based learning, either individually or in groups, increases learning beyond the overall effect found for more student-centered versus less student-centered activities. Various student-centered approaches, such as cooperative learning and peer-tutoring, have been found to accomplish this goal. This meta-analysis contains studies that date from 2000–2017. The question of how to best deliver instruction to k-12 students has dominated the educational conversation, both in terms of theory and practice, since before 1960. Two predominant models have clashed: (a) Traditional teacher-directed instruction (referred to here as teacher-centered Teacher-Centered instruction), where there is little methodological adaptation for individual differences in ability, skills, interests, etc. among students; and (b) so-called student-centered instruction (referred to here as Student-Centered instruction), deriving much of its theoretical justification and methodological intricacies from constructivist thought embodied in the works of Jean Piaget, Lev Vygotsky, Jerome Burner, and many others. While radical constructivism has never become dominant in k-12 schooling (except in a relatively small number of demonstration schools), there has been considerable interest in embedding some of the principles of constructivism into k-12 schooling. This is often referred to as individualized or adaptive instruction, meaning an operational concern for individual students, their abilities, interests, etc., which is nearly the opposite of Teacher-Centered instruction. A great deal of research has demonstrated that approaches to individualism, such as mastery learning, collaborative and cooperative learning, problem-based learning, peer tutoring, and computer-based instruction, are effective in promoting achievement and attitudinal gains, as contrasted with Teacher-Centered instruction, where mastery of content or subject matter is of the greatest concern, and the teacher is the “delivery mechanism.” More recently, this has been extended to include video-based lectures often delivered through the internet, as proposed by proponents of blended learning and its variant the flipped classroom (e.g., Baepler, Walker, and Driessen (2014). Research has also demonstrated that Teacher-Centered instruction is particularly useful in developing basic skills in areas such as reading, spelling, and math (Stockard, Wood, Coughlin, & Khoury, 2018). More recent theory and practice concerning Teacher-Centered (more conventional) and Student-Centered (more adaptive and individualized) instruction suggest that neither perspective is entirely sufficient and that some combination of Teacher-Centered and Student-Centered instruction is possibly more productive. This notion of combined teaching methods (i.e., Teacher-Centered plus Student-Centered) is one of the defining characteristics of the flipped classroom (Baepler et al., 2014). Certainly, students need to acquire skills and knowledge, but they also need to develop their own personal preferences, creativity, problem-solving abilities, and evaluative and self-evaluative perspectives. The current meta-analysis aims to determine if the advantage endowed by Student-Centered instruction also affects content achievement (i.e., content achievement is the outcome measure in this meta-analysis). The current meta-analysis was designed to explore teaching and learning in k-12 classrooms and the achievement benefit that derives from more Student-Centered versus less Student-Centered classrooms. Several perspectives informed the basis for the research approach described here, but none more so than the words of Gersten et al. (2008) while exploring through meta-analysis the question of Teacher-Centered versus Student-Centered instructional practices in elementary mathematics instruction. In the final report of their study, the group stated: “The Task Group found no examples of studies in which learners were teaching themselves or each other without any teacher guidance; nor did the Task Group find studies in which teachers conveyed … content directly to learners without any attention to their understanding or response. The fact that these terms, in practice, are neither clearly nor uniformly defined, nor are they true opposites, complicates the challenge of providing a review and synthesis of the literature …” (p. 12). The current meta-analysis intends to investigate variations of more versus less Student-Centered instruction and the four domains of the instructional process in which they are more or less profitable. Overall, does more Student-Centered instructional practices lead to a significant advantage in the acquisition of content (subject matter) knowledge (i.e., measured learning achievement)? Do any of the four primary (substantive) moderator variables (entered into multiple meta-regression), Teacher's Role, Pacing, Adaptability, and Flexibility, predict an increase or decrease in achievement across degrees of Student-Centered use (From less Student-Centered to more Student-Centered)? Is there a difference in categorical levels of less Student-Centered to more Student-Centered for each of the dimensions of instructional practice listed above, tested in mixed moderator variable analysis? Do any of the secondary (demographic) moderator variables interact with each other (i.e., combine) to produce more versus less Student-Centered instructional practices? Following the guidelines of the Campbell Collaboration (Kugley et al., 2017), in order to retrieve a broad base of studies to review, we started by having an experienced Information Specialist search across an array of bibliographic databases, both in the subject area and in related disciplines. The following databases were searched for relevant publications: ABI/Inform Global (ProQuest), Academic Search Complete (EBSCO), ERIC (EBSCO), PsycINFO (EBSCO), CBCA Education (ProQuest), Education Source (EBSCO), Web of Knowledge, Engineering Village, Francis, ProQuest Dissertations & Theses Global, ProQuest Education Database, Linguistics and Language Behavior Abstracts (ProQuest). The search strategy was tailored to the features of each database, making use of database-specific controlled vocabulary and search filters, but based on the same core key terms. Searches were limited to the year 2000–2017 and targeted a k-12 population. Database searching was supplemented by using the Google search engine to locate additional articles, but principally grey literature (research reports, conference papers, theses, and research published outside conventional journals). Be publicly available and encompass studies from 2000 to the present; Feature at least two groups of different instructional strategies/practices that can be compared according to the research question as Student-Centered and Teacher-Centered instruction; Include course content and outcome measures that are compatible with the groups that form these comparisons; Contain sufficient descriptions of major instructional events in both instructional conditions; Satisfy the requirements of either experimental or high-quality quasi-experimental design; Be conducted in formal k-12 educational settings eventually leading to a certificate, diploma, degree, or promotion to a higher grade level; Contain legitimate measures of academic achievement (i.e., teacher/researcher-made, standardized); and Contain sufficient statistical information for effect size extraction. Means and standard deviations for each treatment and control group; Exact t value, F value, with an indication of the ± direction of the effect; Exact p value (e.g., p = .011), with an indication of the ± direction of the effect; Effect sizes converted from correlations or log odds ratios; Estimates of the mean difference (e.g., adjusted means, regression β weight, gain score means when r is unknown) Estimates of the pooled standard deviation (e.g., gain score standard deviation, one-way ANOVA with three or more groups, ANCOVA); Estimates based on a probability of a significant t test using α (e.g., p < .05); and Approximations based on dichotomous data (e.g., percentages of students who succeeded or failed the course requirements). Effect sizes were initially calculated as Cohen's d (Cohen, 1988) and then converted to Hedges'g (i.e., correction for small samples; Hedges & Olkin, 1985). Standard errors (SEd) were calculated for d and then converted to standard errors of SEg applying the correction formula for g. Hedges’ g, SEg, and sample sizes (i.e., treatment and control) were entered into Comprehensive Meta-Analysis 3.3.07 (Borenstein, Hedges, Higgins, & Rothstein, 2014) where statistical analyses were performed. The effect sizes were coded for precision and these data were analyzed in moderator variable analysis. Overall weighted random effects analysis with the statistics of g ¯ , SEg, Vg, upper and lower limits of the 95th confidence interval, zg, and p value; Homogeneity is estimated using Q-Total, df, and p value. I2 (i.e., percentage of error variation) and tau2 (i.e., average heterogeneity) is also calculated and reported. Meta-regression (single and multiple) is used to determine the relationship between covariates and effect sizes; and Mixed-model (i.e., random and fixed) moderator variable analysis is used to compare levels (categories) of each coded moderator variable. Q-Between, df, and p value are used to make decisions about the significance of each categorical variable. Question 1: Overall, does more Student-Centered instructional practices lead to a significant advantage in the acquisition of content (subject matter) achievement (i.e., measured learning). Result: Answering the basic question, more Student-Centered instructional conditions (i.e., the treatment described above) outperform less Student-Centered to a moderate extent. The average effect, g ¯ = 0.44, k = 365, z = 4.56, p < .00, SE = 0.03, Q = 3,095.89, I2 = 88.22, tau2 = 0.27, between the mean of the more Student-Centered treatment and the less Student-Centered control, suggesting that teachers who promote and enact active classroom processes (more Student-Centered instruction), can expect to see better student achievement than in classrooms where teachers employ less Student-Centered instruction. Also, a linear trend was found in meta-regression when Hedges’ g ¯ was regressed on degree of Student-Centered instruction (β = 0.04, SE = 0.02, z = 2.41, p = .032). The distribution remains significantly heterogeneous. Question 2: Do any of the four moderator variables (entered into multiple meta-regression), Teacher's Role, Pacing, Adaptability, and Flexibility, predict an increase or decrease in achievement across degrees of Student-Centered use (From less Student-Centered to more Student-Centered)? Result: In meta-regression, Teacher's role produces a significant linear trend (β = 0.06, SE = 0.04, z = 4.42, p < .001) and Pacing (β = −0.14, SE = 0.04, z = 3.18, p = .002). Adaptability, and Flexibility are not significant (p > .05). However, the trend for Teacher's role and Pacing is opposite (note the opposite signs on β). Teacher's role is significantly positive (i.e., more Student-Centered instruction produced higher achievement), while Pacing produces the reverse (i.e., a significantly negative trend). For Pacing, more Student-Centered methods produce lower achievement. Question 3: Do any of the moderator variables interact with each other (i.e., combine) to produce more versus less Student-Centered instructional practices? Result: Yes, Teacher's Role compared to two dimensions added to the Teacher's Role produce significantly different results (Q-Between = 7.76, df = 3, p = .02: Teacher's Role and Teacher's Role plus Adaptability significantly outperformed Teacher's Role plus Flexibility. Question 4: Is there a difference in categorical levels of less Student-Centered to more Student-Centered for each of the dimensions of instructional practice listed above, tested in mixed moderator variable analysis? Result: Only one of five moderator variables produced a significant differentiation among levels. Among four moderator variables (i.e., grade level; STEM versus Non-STEM subjects; individual subjects; and ability profile) only ability profile significantly differentiated among levels. Special education students demonstrated significantly higher achievement compared to the General population of students. This meta-analysis provides strong evidence that Student-Centered instruction leads to improvements in learning with k-12 students. Not only is the overall random effects average effect size of medium strength ( g ¯ = 0.44), but there is also a demonstrated (subtle but significant) linear relationship between more Student-Centered classroom instruction and effect size (p = .03). Taken together, these results support the efficacy of allowing students to engage in active learning or other forms of Student-Centered enterprise as part of a comprehensive educational experience. The question of how to provide the best-quality instructional conditions for students of all grade levels has been scrutinized extensively since the early 1960s, principally from two major perspectives: Teacher-centeredness (Teacher-Centered) and student-centeredness (Student-Centered). Student-Centered education initially arose from the writings of early progressive educators like John Dewey, and was carried on subsequently, in various forms, by Jean Piaget, Lev Vigotsky, Jerome Bruner, and Carl Rogers, to name only a few. The ideas were radical when first introduced, but the notion of Student-Centered education resonated in educational circles, where lecturing and rote memorization was still the standard for quality education and led to vast amounts of theorizing and research to show that students could succeed in learning of all sorts without a strongly transmissive approach on the part of the teacher. Today, the terms individualized instruction and adaptive teaching have become a popular expression for current practice and are used nearly synonymously with Student-Centered learning. However, since their inception, Student-Centered practices have inspired resistance, both from the public and from educational theorists. Thus, after Student-Centered practices were widely introduced, a dichotomy arose in the literature, with one side promoting the continuation of Teacher-Centered learning and on the other side the adopting Student-Centered learning practices. This was argued as a dichotomy for many years. However, the arguments have abated somewhat now with the general recognition that there is value in both approaches. Generally speaking, educators no longer aspire to a pure implementation of either approach, but now discuss questions of which method, when, and for what purpose is best. Conceptual understanding of individualized learning and adaptive teaching varies broadly, encompassing a multitude of instructional strategies, approaches, and activities. It stretches from accounts of specific systems of instruction such as mastery learning (Bloom, 1968) and scaffolded adaptive feedback in computer-based instruction (e.g., Azevedo & Bernard, 1995) to more general conceptions of active learning and individualization that involve approaches such as cooperative learning (e.g., Johnson & Johnson, 2002; Johnson, Johnson, & Maruyama, 1983), collaborative learning (e.g., Bernard, Rojo de Rubalcava, & St-Pierre, 2000), problem-based learning (e.g., Zhang et al., 2015), and project-based learning (e.g., Bernard & Lundgren-Cayrol, 2001). It also includes educational concepts, largely derived from elements of constructivism, such as discovery learning, inquiry-based learning, activity-based learning, experiential learning, and other forms of Student-Centered education (Tobias & Duffy, 2009). Notions of unguided Student-Centered learners have not been free from detractors. Dewey criticized this approach in Experience and Education (Dewey, 1938), and, more recently, Kirschner, Sweller, and Clark (2006) published an influential piece that argued that the practice of turning kids loose to learn defies many of the tenets of the psychological principles of working memory and that guided instruction is both more efficient and ultimately more profitable to long-term learning outcomes. A flurry of responses and rejoinders ensued with no clear resolution, but the educational community was left with the strong impression that a teacher's role in Student-Centered learning was better as a guide on the side rather than a silent witness (King, 1993). The learning sciences have further contributed to the distinction between social constructivism and individual constructivism providing a theoretical grounding for teacher versus learner-based strategies (Kolodner, 2004). Current and developing applications, informed by pedagogical principles espoused by case-based learning (e.g., Kolodner et al., 2003). The earliest large-scale research project, aimed at exploring the efficacy of so-called progressive education, was conducted between 1933 and 1941 by the Progressive Education Association (funded by the General Education Board and other foundations). Twenty-nine model schools were selected for curricular experimentation with the security that over 200 colleges and universities would accept their students upon recommendation by their principals. Changes in these schools included more individualized instruction and more access to alternative and cross-disciplinary programs, which emphasized greater access to arts and extracurricular programs. Results indicated that students graduating from the 200 schools scored on par in basic courses (e.g., mathematics and science) with students from traditionally oriented schools and that there was more activity in artistic, political, and social engagement in students from the alternative experimental schools. The long-term impact of these experiments is generally described as influence on its participants and subsequent reformers rather than dramatic change. The intervening conservatism brought about by World War II and the ensuing Cold War are often cited as deterrents to widespread change in the overall educational system in the United States (Aiken, 1942). Examples of further attempts to make teaching and learning more individualized and adaptive can be found in both the early and current research literature. They include, but are not limited to, mastery learning (e.g., Bloom, 1968), Personalized System of Instruction (PSI; e.g., Keller, 1968), assorted forms of peer instruction (e.g., Mazur, 1997), various practices of reciprocal reading/writing activities (e.g., Huang & Yang, 2015), collaborative and cooperative learning, problem and project-based learning and, more recently, Intelligent Tutoring Systems (ITS; e.g., Huang & Shiu, 2012). Several of these approaches are summarized in the following paragraphs and a number of the most common group-based Student-Centered approaches are depicted in a Venn Diagram (Figure 2) that shows their inter-relationship and approximate overlap (Bishop & Verleger, 2013, p. 6). The benefits and limitations of so-called systems of instruction (i.e., mastery learning, PSI, and ISI) are summarized separately in both qualitative and quantitative reviews. In the late 1970s and early 1980s, several relevant meta-analyses were published on mastery learning and its variant PSI. First, Lysakowski and Walberg (1982), Guskey and Gates (1986), Guskey and Pigott (1988), Slavin (1987), and Kulik, Kulik, and Bangert-Drowns (1990) each performed successive meta-analyses (Slavin's was the best evidence synthesis) on the efficacy of mastery learning. The studies produced equivocal and highly debatable findings. Kulik, Kulik, and Cohen (1979) reviewed 75 individual comparative studies of Keller's Personalized System of Instruction (PSI is a spin-off of mastery learning) college teaching method. In comparison to conventional instruction, the PSI approach was demonstrated to have a positive effect on student achievement and course perception (mean effect size of nearly 0.70sd for both). Bangert and Kulik (1982) looked at the effectiveness of the Individualized Systems of Instruction (ISI, a spin-off of PSI) in secondary school students. They broadened the list of outcomes to account not only for student achievement (e.g., final exams), but also critical thinking, attitudes toward subject matter, and student self-concept. For all outcome types, the findings were inconclusive. For example, for the achievement data, only 8 out of 49 studies demonstrated statistically significant results in favor of ISI (four studies favored more conventional teaching methods and the rest were inconclusive). Finally, Kulik (1984) attempted a wider research synthesis (encompassing over 500 individual studies) of the effectiveness of programmed instruction and ISI, paying special attention to the moderator variables of study dates and grade levels. The most promising findings indicated that more recent studies showed higher effects than the earlier ones and that college-level students benefited significantly from using ISI compared with elementary and secondary school students. In summary, as stated earlier, these meta-analyses produced inconclusive results. Moreover, they are rather outdated and practically none of the above-mentioned instructional methods exists now in their original forms (e.g., Eyre, 2007 was able to identify fewer than 50 studies of PSI for the period between 1990 and 2006 in the PsycInfo database). Much of the preceding discussion has been about systems of individualized instruction, designed and intended as self-contained approaches to individualizing student learning. Because of their rule-based nature, they may be thought to be individualized, but insufficiently adaptive (systems often are not very adaptive). Several meta-analyses addressed the topic of individualized and adaptive instruction (i.e., instructional approaches that can be applied as local circumstances dictate), though in very specific narrowly focused forms. Aiello and Wolfle (1980) summarized research on individualized instruction in science education compared with traditional lectures and found that individualized instruction was more effective. Horak's (1981) meta-analysis of self-paced modular instruction of elementary and secondary school math (1981) produced a wide variety of both positive and negative effect sizes. A highly cited meta-analysis of active learning in science, engineering, and mathematics subject matters (Freeman et al., 2014) found a moderate effect size ( d ¯ = 0.47) based on 158 studies. The authors also state that “The results raise questions about the continued use of traditional lecturing as a control in research studies, and support active learning as the preferred, empirically validated teaching practice in regular classrooms” (p. 8410). This sentiment appears to add support to the comparative approach that is employed in the current meta-analysis. Kraft, Blazar, and Hogan (2018) examined the effects of teacher coaching (i.e., tutoring) on student achievement and found minor effect on achievement ( d ¯ = 0.08). Though these instructional approaches are not “adaptive,” per se, at least peer tutoring opens the educational process to much greater involvement of students, and thus accounts more for their individual inputs in learning. The effect size tended to be relatively small in middle school students, but higher at elementary and high school levels. There have been numerous reviews and meta-analyses of various forms of computer-assisted instruction (CBI). Ma, Adesope, Nesbit, and Liu (2014) meta-analyzed studies of ITS in a variety of subject matters, from reading and math to law and medical education. More specific reviews have been conducted on the effectiveness of feedback and scaffolding in CBI and ITS. The list of moderator variables included the type of both experimental and comparison treatments, as well as outcome type, student academic level, study discipline, etc. The highest achievement effects of using ITS were found in comparison with non-ITS computer-based instruction ( d ¯ = 0.57) and teacher-centered, large-group instruction ( d ¯ = 0.42), whereas in comparison with human tutoring it was even negative ( d ¯ = −0.11), though not statistically significant. ITS-based practices were similarly effective when used either alone or in combination with various forms of teacher-led instruction in many subject domains. In particular, certain specific aspects of instruction like feedback and scaffolding in CBI and ITS systems have come under scrutiny. Azevedo & Bernard (1995) examined studies testing the effectiveness of computer-provided feedback against no feedback, and Belland, Walker, Olsen, and Leary (2015) synthesized studies investigating feedback in computer-based scaffolding. In both cases, the average effect size was around d ¯ = 0.50 in favor of feedback conditions. Overall, the research literature paints a positive picture of Student-Centered learning. There has been considerable research in Teacher-Centered education as well over the years. In the 1960s, during the Lyndon Johnson administration in the United States, a massive experiment called Project Follow Through was initiated to test the efficacy of a range of instructional strategies. The intent was to evaluate the relative advantages of models of instruction that ranged from Direct Instruction (i.e., DISTAR) to so-called Open Education (i.e., based on the British Infant School Model). After years of testing and millions of dollars spent, only one really striking finding emerged: That direct instruction advantaged learners in terms of both measures of achievement and affect, outperforming other models by as much as 1.5 SD (standard deviation). While a great deal of controversy surrounds the conduct and findings of this large-scale educ