This Article Abstract Free Full Text (Free PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Cornelius-White, J. Search for Related Content Social Bookmarking                   What's this?

Learner-Centered Teacher-Student Relationships Are Effective: A Meta-Analysis

Missouri State University

	Abstract

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 
Person-centered education is a counseling-originated, educational psychology model, overripe for meta-analysis, that posits that positive teacher-student relationships are associated with optimal, holistic learning. It includes classical, humanistic education and today’s constructivist learner-centered model. The author reviewed about 1,000 articles to synthesize 119 studies from 1948 to 2004 with 1,450 findings and 355,325 students. The meta-analysis design followed Mackay, Barkham, Rees, and Stiles’s guidelines, including comprehensive search mechanisms, accuracy and bias control, and primary study validity assessment. Variables coded included 9 independent and 18 dependent variables and 39 moderators. The results showed that correlations had wide variation. Mean correlations (r= .31) were above average compared with other educational innovations for cognitive and especially affective and behavioral outcomes. Methodological and sample features accounted for some of the variability.

Key Words: person centered • learner centered • instructional relationships • meta-analysis • constructivist • humanistic

The literature is missing a meta-analysis of classical person-centered education, the learner-centered model, or teacher-student relationships. Theoretically, the classical and learner-centered models come from somewhat different traditions (humanistic and constructivist, respectively) and decades (the 1950s through the 1970s and the 1990s through the 2000s, respectively), and the learner-centered model includes more of a focus on student variables and learning processes as critical to positive student outcomes (American Psychological Association, 1997; Lambert & McCombs, 1998). However, when operationalized, the two models are remarkably similar in their teacher-student relational variables.

The classical approach emphasizes teacher empathy (understanding), unconditional positive regard (warmth), genuineness (self-awareness), nondirectivity (student-initiated and student-regulated activities) and the encouragement of critical thinking (as opposed to traditional memory emphasis). Carl Rogers was the founder of nondirective and client-centered therapy and influenced its spread as the person-centered approach to interpersonal relations, nursing, organizational functioning, and education (Cornelius-White & Cornelius-White, 2004; Rogers, Cornelius-White

Cornelius-White, & Cornelius-White, 2005). Freedom to Learn, in its three editions (Rogers, 1969, 1983; Rogers & Freiberg, 1994), is perhaps the best known presentation of the theory, though it was originally presented in Rogers’s (1951) Client-Centered Therapy.

Rogers’s theory of education has as its goal the facilitation of the whole and fully functioning person, who is a citizen and leader in a democratic society. Rogers (1969) saw the facilitation of citizens as vital, without which education will "doom us to a deserved and universal destruction" (p. 125). Elsewhere, I have summarized as follows:

Rogers saw "the facilitation of learning as the function which may hold constructive, tentative, changing, process answers to some of the deepest perplexities which beset man today" (p. 105). It was likely these perplexities that lead Rogers to facilitate international cross-cultural workshops [in the process garnering a Nobel Peace Prize nomination and] encouraging a revolutionary educational style for real world problems both in the classroom and at the diplomatic table (Rogers & Russell, 2002). Rogers (1969) was broad in his intentions: "Learning becomes life" (p. 115). (Cornelius-White, 2006b, p. 8)

This goal includes the fostering of the development of those who "take self-initiated action, . . . are capable of intelligent choice and self-direction, . . . are critical learners, . . . have acquired knowledge, . . . adapt flexibly, . . . utilize all pertinent experience freely and creatively, . . . cooperate effectively, . . . [and] work . . . in terms of their own socialized purposes" (Rogers, 1951, pp. 387–388).

Rogers (1969) held that "certain attitudinal qualities which exist in the personal relationship between the facilitator and the learner" yield significant learning (p. 106). Facilitation requires at least an initial genuine trust in learners by the facilitator, followed by the creation of an acceptant and empathic climate. Acceptance is "prizing," "non-possessive caring," and the "operational expression of his essential confidence and trust in the capacity of the human organism" (p. 109). The perception of care by the student is emphasized. Empathy is "the attitude of standing in the other’s shoes, of viewing the world through the student’s eyes, [which] is almost unheard of in a classroom" (p. 112). Classical person-centered education also includes facilitator flexibility in teaching methods; transparent compromise with learners, school administrations, the public, and the teacher’s own self; collaborative and student-self evaluation; and the provision of human and learning resources. Seeking and embracing a willingness to be changed are hallmarks of students and facilitators within the person-centered framework. Activities include but are not limited to solving relevant and real problems, providing resources, using contracts for planning and evaluation, forming learning groups, programmed instruction adapted to individual needs, encounter groups, using the community, and peer tutoring. Rogers summarized the theory: "Within the limitations which are imposed by circumstance and authority, or are imposed by the instructor as necessary for his own psychological comfort, an atmosphere of permissiveness, of acceptance, or reliance upon student responsibility, is created" (p. 397). Patterson (1973), Combs (1962), Axline (1947), and others have also contributed significantly to the theoretical formulations. Aspy, Roebuck, Carkhuff, and colleagues were the most prolific American researchers, while Tausch, Tausch, and colleagues were the most prolific European researchers (Rogers, 1983).

The learner-centered model refers to a "perspective that couples a focus on individual learners . . . with a focus on learning" (McCombs & Whisler, 1997, p. 9). It is rooted in the American Psychological Association’s (1997) 14 learner-centered principles, which are summarized by the four domains of metacognitive and cognitive, affective and motivational, developmental and social, and individual differences factors. An emphasis on quality teacher-student relationships has been further advocated by McCombs (2003b, 2004a, 2004b) and colleagues (McCombs & Whisler, 1997). The current synthesis focuses on these teacher relational variables and treats student and learning variables as outcomes that are facilitated by these relational practices. These relational practices include teachers’ honoring of students’ voices, adapting to individual and cultural differences, encouraging learning, thinking, and having learner-centered beliefs. The learner-centered model uses research that sees learning as

non-linear, recursive, continuous, complex, relational, and natural in humans. . . . Learning is enhanced in contexts where learners have supportive relationships, have a sense of ownership and control over the learning process, and can learn with and from each other in safe and trusting learning environments. (McCombs, 2004a, p. 7)

Developmental, feminist, and multicultural models also drive current research on teacher-student relationships, such as that by Pianta, Midgley, and Wentzel. Attachment theories (e.g. Bowlby, 1969; Stern, 1977) are influential on teacher-student relationship research and originate from perspectives on mother-child relationships. Attachment research emphasizes the long-lasting and personality-forming nature of relationships. Secure and reciprocal attachments are important for students to engage in their relationships with teachers, peers, and subject matter and develop healthy self-concepts and senses of well-being. Feminist and multicultural theories (e.g. Gilligan, 1982, Noddings, 1984, 2003; Walker, 2004) assert that children, women, and men exist only in relationships. Dominant culture, characterized by White, heterosexual, male leadership, underemphasizes relationships and overemphasizes separation, which is seen as the primary source of suffering and developmental problems. Positive interdependence is characteristic of this view and is somewhat similar to that found in Johnson and Johnson’s (2001) theory of cooperative learning. This perspective that persons only exist in, and in fact are, relationships is also found in contemporary person-centered theory (e.g. Cornelius-White, in press; Schmid, 2001).

However, perhaps more compelling than either the classical or current models are the voices from inside schools themselves. Poplin and Weeres (1994) reported a powerful qualitative study investigating the question "What is the problem with schooling?" and involving participants in four multiethnic school districts in California from every aspect of school life (students, teachers, cafeteria workers, security guards, parents, and administrators). The number one problem identified was relationships. They wrote,

Participants feel the crisis inside schools is directly related to human relationships. Most often mentioned were relationships between teachers and students. Where positive things about schools are noted, they usually involve reports of individuals who care, listen, understand, respect others and are honest, open, and sensitive. (p. 12)

They continued, "Students desire authentic relationships where they are trusted, given responsibility, spoken to honestly and warmly, and treated with dignity" (p. 20).

The present study synthesized these traditions’ correlational and causal associations between teacher-student relationships and affective or behavioral and cognitive student outcomes. It also included studies that did not research either whole model but looked at select variables, which are emphasized in either or both models, most frequently warmth or care (e.g., Noddings, 1984, 2003; Thayer-Bacon & Bacon, 1998; Voelkl, 1995).

The study aimed to answer the following questions:

1. What is the degree of association between all person-centered teacher variables and all positive student outcomes combined?
   
2. What is the degree of association between positive teacher-student relationships and positive student outcomes?
   
3. What is the degree of association between submodels of person-centered education and positive student outcomes?
   
4. What are the degrees of association between individual person-centered teacher variables and positive student outcomes?
   
5. How much are person-centered teacher variables associated with the cognitive versus affective or behavioral outcomes?
   
6. What moderators may account for the variability of correlations between person-centered teacher variables and positive student outcomes?
   

	Method

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 
The background and methods of this meta-analysis have been articulated in more depth in Cornelius-White (2006a, 2006b); Cornelius-White and Brown (2006); Cornelius-White and Cornelius-White (2005); Cornelius-White and Godfrey (2004); Cornelius-White and Harbaugh (in press); and Cornelius-White, Hoey, Cornelius-White, Motschnig-Pitrik, and Figl (2004). The synthesis was designed to satisfy all of the eight criteria Mackay, Barkham, Rees, and Stiles (2003) used to appraise the quality of 255 syntheses, only 11% of which met all eight criteria. They included clear questions, a comprehensive search for primary studies, inclusion criteria, the validity of primary studies, accuracy and bias control, the analysis of variation in findings, the appropriate combination of findings, and adequately supported conclusions.

Procedures
I used comprehensive search mechanisms to locate studies, including PsycINFO, the Education Resources Information Center (ERIC), reference lists from other included studies, published bibliographies (including those of Carkhuff, 1983; Schmid, 2006; and Stanley & Purkey, 2001), and experts’ knowledge of the existence of additional studies (e.g., Barbara McCombs, Judith Meece, Reinhard Tausch, Dave N. Aspy, R. R. Carkhuff, Howard Kirschenbaum, William W. Purkey, and Jerome Freiberg). Some studies (mostly unpublished) were identified but were not obtained during the data collection phase, between August 2002 and May 2004. The inclusion criteria cast a wide net, requiring only that a study have statistics, be written in English or German, and include one or more independent and one or more dependent variables. However, only independent variable words and synonyms in some cases (e.g., congruence for genuineness) were used as descriptors for the locations of studies. I entered the independent (teacher) variables, which included empathy, warmth, genuineness, nondirectivity, higher order thinking, encouraging learning/challenge, adapting to individual and social differences, and composites of these, as the independent variables into the PsycINFO and ERIC databases. I read the titles of the identified manuscripts, and moved on to abstracts and/or full text as needed to determine if the reference was an acceptable study (i.e., having one or more independent and dependent variables and statistics).

In addition to the 9 independent variables, there were 18 dependent variables: 9 cognitive and 9 affective or behavioral. Cognitive dependent variables included achievement batteries, grades/retention, perceived achievement, verbal achievement, math, science, social science, IQ, and creative/critical thinking. Affective or behavioral dependent variables included student participation/initiation, positive motivation, self-esteem/mental health, social connection, attendance/absences, global satisfaction, disruptive behavior, negative motivation, and drop-out prevention. There were 39 moderator variables: 16 concerned with sample qualities, 19 with methodological features, and 5 with publishing. Moderator variables are those that might potentially alter the size of the relationships between independent and dependent variables. A moderating effect was noted when studies showed that differences (e.g., in gender) on a moderator were associated with significant differences in the size of the relationship between independent and dependent variables. For example, teacher care (an independent variable) appeared to be more associated with student learning (a dependent variable) when demonstrated by a female (a moderator variable) teacher than a male teacher. Six moderators were concerned with the students, 5 with teachers, and 5 with the whole sample. Of the 19 methodological modifiers, 5 concerned the nature of the independent variables, 5 that of the cognitive dependent variables, and 6 that of the affective or behavioral variables; 3 were general. Table 1 includes specific coded features of studies.

I analyzed the results at both the study and finding levels. The study level was calculated by taking the mean of all findings from each study, a method that tends to conservatively estimate the overall relationship found from each study (Hunter & Schmidt, 2004). The finding level uses each separate findings, regardless of how many came from each study, thereby not keeping the assumption of independence. However, finding-level analysis provides a larger sample, allows the analysis of separate independent and dependent variables, and is used more frequently in syntheses. Hence, it was used to address the research questions when not otherwise noted, with the exception of sample moderator analysis.

Participants
The current synthesis included 119 studies in English and German conducted from 1948 to 2004, with 1,450 findings involving approximately 355,325 students, 14,851 teachers, and 2,439 schools. The numbers were approximate because not all studies reported all three of these sample numbers, and individual findings were based on slightly different numbers in original studies. The majority of studies involved both female and male teachers, though 13 had only female teachers and 5 had only male teachers. Caucasian, African American, Latino, and Filipino participants were included in large numbers. Native and Asian populations were under-represented. Locations included most areas of the United States, the Philippines, Brazil, Germany, the United Kingdom, and Canada. Participants came from urban, suburban, and rural communities, with low to high incomes. The median student sample for each study was 500. The smallest individual study sample had 20 students, and the largest had 81,000 students. Grade levels (and corresponding ages) included pre-K through 20, though the majority of students were in Grades 1 through 12. Economic class and intellectual ability included low to high levels. Experience levels of teachers varied from less than a bachelor’s degree through 40 years with graduate degrees.

Measures
Two persons coded information on 9 independent, 9 cognitive dependent, 9 affective or behavioral dependent, and 39 moderator variables. In English, 92% of studies were coded independently, but differences were unanimously resolved before data entry. Data entry was done by a third person and checked by myself. A random sampling of 12 studies (10%) yielded a coefficient of 0.85. The remaining 8% of studies in English and all studies in German were coded by two persons working together as proficiency in the coding process was developed. The Scientific Methods Score (Sherman et al., 1997) is a Likert-type scale ranging from 1 to 5 that is used in other meta-analyses. In brief, a score of 1 is characterized by having statistics, 2 by having an independent variable that was measured before a dependent variable, 3 by having comparison groups, 4 by having randomized or controlled moderators, and 5 by having large samples. A higher score generally necessitates the presence of features from lower scores. For example, a simple correlational study with 1,000 students would be scored 1, not 5, because although large, it does not include the additional features of a study scored 5. Independent and dependent variables were done from multiple perspectives, including teacher, observer, student, or composites of two or all of these. The inclusion of the specific modifying variables was brainstormed and decided on from the suggestions of 30 teachers (all of whom were graduate students), 4 professors in three colleges on two continents, and 1 high school student.

	Results

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 
The results are organized according to the six research questions.

What Is the Degree of Association Between All Person-Centered Teacher Variables and All Student Outcomes Combined?
The study-level analysis found an average correlation of r = .34 (SD = .20). The corrected correlation was r = .39 (SD = .22). The 95% confidence interval for the corrected study-level mean correlation was r = .35 to r = .43. Figure 1 shows the distribution of study-level corrected mean correlations. The finding-level analysis found an average correlation of r = .26 (SD = .25). The corrected correlation was r = .31 (SD = .29). The 95% confidence interval for the corrected finding-level mean correlation was r = .30 to r = .33. Studies with more findings had smaller correlations. Table 2 presents the study-level mean corrected correlations and select variables.

View larger version (26K): [in this window] [in a new window]   FIGURE 1 Distribution of study-level corrected mean correlations.

What Is the Degree of Association Between Submodels of Person-Centered Education and Student Outcomes?
The analysis compared classical person-centered, contemporary learner-centered, and isolated parts of these models (most frequently warmth). The classical model showed a mean correlation of r = .36 (SD = .29) and a corrected correlation of r = .41 (SD = .34). The learner-centered model showed a mean correlation of r = .26 (SD = .25) and a corrected correlation of r = .31 (SD = .29). The third model was composed of parts (most frequently warmth) of the first two models. It showed a mean correlation of r = .18 (SD = .20) and a corrected correlation of r = .23 (SD = .25).

What Are the Degrees of Association Between Individual Person-Centered Teacher Variables and Positive Student Outcomes?
Figure 2 shows the average corrected correlations for each class of the nine classes of independent variables. Aside from positive relationships, the highest corrected correlations were found for nondirectivity (r = .35), empathy (r = .32), warmth (r = .32), and the encouragement of higher order thinking (r = .29). Encouraging learning (r = .23) and adapting to differences (r = .20) showed moderate relationships. Genuineness showed a smaller relationship (r = .14). Learner-centered (and the inverse of non-learner-centered) beliefs (r = .05) showed almost no relationship to positive student outcomes.

View larger version (27K): [in this window] [in a new window]   FIGURE 2 Corrected correlations of each person-centered teacher variable with all positive student outcomes.

View larger version (26K): [in this window] [in a new window]   FIGURE 3 Corrected correlations of all person-centered teacher variables with each cognitive student outcome.

View larger version (26K): [in this window] [in a new window]   FIGURE 4 Corrected correlations of all person-centered teacher variables with each affective or behavioral student outcome.

Figure 5 shows that when observers’ (r = .40) and students’ (r = .33) perspectives were used to measure the person-centered teacher variables, substantial correlations were found. Composites of three of the perspectives (r = .27) showed a moderate correlation, and teachers’ own views of their behavior (r = .17) showed a smaller correlation. In contrast, for affective or behavioral outcomes, composites of the three perspectives (r = .48) and observers’ (r = .40) views showed large influence on the sizes of the correlations between teacher variables and student outcome. Teacher (r = .33), followed by student (r = .31), also showed substantial correlations.

View larger version (11K): [in this window] [in a new window]   FIGURE 5 Corrected correlations of each measurement perspective for person-centered teacher variables with all positive student outcomes.

	Discussion

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 
Summary of Findings
Overall, the meta-analysis found that person-centered teacher variables have an above-average finding-level association with positive student outcomes (r = .31), with wide variability (SD = .29). The overall corrected correlation between person-centered teacher variables and cognitive student outcomes was r = .31 (SD = .25), while the overall corrected correlation between person-centered teacher variables and affective or behavioral student outcomes was r = .35 (SD = .20). The correlation between positive teacher-student relationships and positive student outcomes (r = .36) appeared larger. Many prominent person-centered experts have argued the inseparable hypothesis, that individual facilitator conditions are not separate but merely parts of a gestalt (Bozarth, 1998; Schmid, 2003). Although empirically, the conditions can be separated and their effects in isolation can be beneficial, together, in the gestalt of a positive learner-centered teacher-student relationship, they are more effective.

Tying Findings to the Literature
Fraser, Wahlberg, Welch, and Hattie (1987), in their seminal synthesis of syntheses of 134 meta-analyses based on 7,827 studies, 22,155 correlations, and 5 to 15 million participants, supported by subsequent syntheses of syntheses (see Hattie, 1999), found an overall average correlation of r = .20 for all educational innovations. They asserted that any correlation greater than r = .20 is "well worth pursuing," and any correlation greater than r = .30 "should be of much interest" (p. 208). Hence, at a broad level, this meta-analysis shows that learner-centered relationships are well worth pursuing.

The effect of person-centered teaching on cognitive student outcomes offers intuitively logical results. First, the unusually high correlation (r = .45) of students’ critical and creative thinking (usually measured by observers as frequency of occurrence) follows from the models’ explicit attempts to encourage higher order thinking and respect for divergent opinions. The second finding is that basic learning (IQ, verbal, math) skills appear more associated with learner-centered teacher variables than success in more specific areas (grades, science, social science, batteries, and perceived achievement). Part of this effect may be due to a confound with the ages of participants in these comparisons, whereby all IQ studies concerned early elementary school students, and science, social science, and grades often are not tested at these earlier, more potentially influential ages. Nevertheless, grade level was not found to be an overall moderator.

In terms of specific affective or behavioral outcomes, person-centered education is associated with large increases in participation/initiation (r = .55), satisfaction (r = .44), and motivation to learn (r = .32). These findings seem to indicate that students become very engaged in learner-centered classrooms. The effects on self-esteem (r = .35) and social connections and skills (r = .32) seem to indicate that students make better relationships with both themselves and others. Such prosocial outcomes are the more traditional aims of counseling, from which classical person-centered education evolved. Affective or behavioral outcomes that have become larger concerns in recent years involve the reduction of oppositional or resistant behavior. Reduction in dropout (r = .35), disruptive behavior (r = .25), and absences (r = .25) seem to be associated with a learner-centered environment. As in findings from the psychotherapy literature (Beutler, Rocco, Moleiro, & Talebi, 2001; Cornelius-White, 2003), resistance appears best dealt with by nondirectivity. Avoiding power struggles through empathy and the encouragement of self-initiated learning seems to help. However, negative motivation (r = .06) appears to be negligibly related. In other words, if students are actively not complying with school structure, increased positive relationships seem to curtail these behaviors, but if students are only passively avoiding effort, person-centered education appears to have virtually no effect. Students (like clients) who have effort-avoidance goals may continue to have them, though their curiosity and compliance with basic structures (i.e., attending and not disrupting) will likely increase.

Fraser et al. (1987) found that researchers emphasized affective or behavioral outcomes less, but they have become increasingly more important to both researchers and especially the public. For example, according to the Public Agenda Foundation (1994), 71% of Americans feel that affective or behavioral education is more important than academics, especially motivation variables such as work habits and curiosity and interpersonal variables such as respect. Similarly, 88% of Americans feel that school’s mission is not only to teach the three R’s but also to build productive citizens, and 78% feel that schools need good teacher and student behavior to flourish (Public Agenda Foundation, 2004). Fraser et al. also found that educational innovations are less successful in influencing affective outcomes. The overall mean was r = .11. Although affective outcomes are usually more narrowly defined than in the current meta-analysis (often meaning only motivational variables), person-centered education appears to be more efficacious (r = .35) than any educational innovation reviewed by Fraser et al. More recent research has found that for specific affective outcomes, there are significantly more effective innovations. For example, Johnson and Johnson’s (2001) meta-analysis found that their Peacemaking program had an average correlation of r = .60 for reducing conflict in schools. Interestingly, Johnson and Johnson’s models bear similarity to person-centered education in their emphasis on cooperation, empowerment, and respect.

The inseparable hypothesis (that the variables coexist and function most effectively together) may also be supported by the analysis of submodels, which found that isolated parts appeared to be less potent than a whole model. An alternative and antithetical explanation is that researcher allegiance (the consistency between the theoretical orientation of the researchers and the theory being tested) biases the amount of differences in models tested. Elliot (2002) found that researcher allegiance accounted for as much as r = .20 in his meta-analytic comparison of humanistic and cognitive behavioral therapies. The difference in magnitude between the classical person-centered and contemporary learner-centered model mean correlation is small but becomes smaller when the one negligible teacher variable (beliefs) is not considered.

Although understanding and holding learner-centered beliefs rather than non-learner-centered beliefs seems heuristically and personally important for teacher development, this lack of empirical association is consistent with one of the largest trends found in educational productivity research. The distal is less potent than the proximal, where students (or their current behavior) are the center. Teachers’ beliefs are more distant from students than their lived behaviors, even as physical attributes of a school (r = –.02) is a more distant variable than students’ prior achievement experiences (r = .75) when correlated with current achievement behavior (Fraser et al., 1987). All of the other person-centered teacher variables (empathy, respect, nondirectivity, encouraging learning and thinking, and adapting to differences) individually showed above average effects (r > .20), with the exception of genuineness (r = .14).

A comparison of the measurement perspective of teacher variables shows that students’ and observers’ perspectives are more predictive of student success than teachers’ views of themselves. This finding mirrors that found in the psychotherapy literature, in which clients’ and observers’ perspectives of therapists’ empathy and alliance are better predictors of client success than counselors’ own views (Bergin & Garfield, 1994). This supports the widespread use of student-rated teacher evaluations. In contrast, when concerned with affective or behavioral student outcomes, all perspectives appear to show the effects of a person-centered environment. The larger correlation found in studies that controlled for the effects of pretest (usually aptitude or achievement) is expectable given that these pretest standings have been shown to have very large correlations with future achievement.

The findings regarding ethnicity seem to suggest that person-centered teacher variables are universal, because a comparison of the associations of teacher-student relationships with student outcomes between teachers of color and Caucasian teachers did not show a significant difference. Learner-centered relationships with female teachers seemed to show more impact than those with male teachers, indicating congruence with traditional gender roles as effective nurturers of human development.

Limitations
This meta-analysis has limitations. One significant limitation for inferring from the studies to classroom practice is the meta-analytic strategy used. Although it contained inferential procedures to correct error and bias for independent and dependent variable unreliability and used progressively more specific methods of analysis, it loses much of its inferential potential in averaging across all variables because there is large heterogeneity across the variable categories. This strategy is more effective at providing a broad and descriptive picture of the literature than it is in showing the specific effects of each individual teacher variable on each student outcome.

Although the overall analyses were based on very large samples, one limitation is the small sample size of some of the specific categories of independent and dependent variables, especially dropout (based on two findings from one study conducted in Brazil). The independent teacher variable categories had a range of 35 (for genuineness) to 605 (for positive relationships and composites) findings. The cognitive findings had a range of 16 (social science) to 134 (verbal). Excluding dropout, affective behavioral outcomes had a range of 42 (satisfaction) to 288 (positive motivation). Moderator analysis had even smaller numbers at times, though clustering categories into the smaller number of categories seen in Table 1 allowed for sufficient sample size.

Limitations of the original studies, seen through their Scientific Methods Scores, also exist. A majority of the studies were correlational in design, and more still did not have randomization and large sample sizes. In these cases, the primarily researched direction, that teachers’ good relational practices influence students toward broad beneficial outcomes, may be misleading. It is possible that these teacher relational variables may actually be caused by student processes and outcomes rather than the other way around. This potentially bidirectional phenomenon is not just a limitation but an important idea. Such reciprocity was actually hypothesized in Rogers’s (1959) original theory statement. Skinner and Belmont (1993) explored these reciprocal relationships using path analysis. In other words, teachers may behave in empathic ways to students who participate, which in turn may increase student participation, setting up a mutually beneficial cycle.

Conceptually, congruence is similar to learner-centered beliefs in that it is mostly distal to students’ direct experiences, being originally hypothesized as a condition of personality integration within the teacher. However, when operationalized, it becomes more proximal, because it is usually measured from an observer’s perspective. As with teacher beliefs, it is arguable that genuineness is heuristically and developmentally useful, even though its correlation with student success is smaller than average. Also, both beliefs and genuineness are difficult to measure, particularly with regard to their proximal effects, and may be indirectly measured through other independent variables as part of the gestalt of positive (learner-centered) relationships discussed above. Hence, measurement difficulties may be a factor in these two small relationships as well.

Implications for Future Research
Future meta-analytic research might focus on specific subsets of learner-centered behaviors to reduce heterogeneity in synthesizing results and increase the inferential potential for future syntheses. For individual studies, integrating classical and current learner-centered models in terms of convergent validity and operationalized definitions of relational variables would help researchers and practitioners understand the strengths of and integrate the specific submodels. Reciprocal effects of teacher and student behavior need further exploration. It is likely that much of the correlation seen between learner-centered teaching and student success is bidirectional. Students’ cooperation and success likely lead to better teacher behavior, just as better teacher behavior leads to student success. Likewise, the effects of relationships in preventing school dropout need further examination. Additional controlled experiments with the current learner-centered model are also called for, because most of these newer studies were more correlational in nature. Given the lack of association found for learner-centered beliefs with positive outcomes and for teacher variables with reduction in students’ negative motivation, the further revision of related measures, methods, and heuristics might be useful. Finally, the long-term effects of positive teacher-student relationships need to be explored.

Perhaps more important than further research is advocacy for the robust associations of positive teacher-student relationships with student success, especially as concerned with the growing importance of affective or behavioral outcomes and learner-centered education’s unusually high relationship with affective or behavioral outcomes (Lambert & McCombs, 1998).

	Conclusion

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 
Overall, learner-centered teacher variables have above-average associations with positive student outcomes. The classical and contemporary models as wholes both appear more supported than looking at variables in isolation. Positive relationships, nondirectivity, empathy, warmth, and encouraging thinking and learning are the specific teacher variables that are above average compared with other educational innovations. Correlations for participation, critical thinking, satisfaction, math achievement, drop-out prevention, self-esteem, verbal achievement, positive motivation, social connection, IQ, grades, reduction in disruptive behavior, attendance, and perceived achievement are all above average and are presented in decreasing order. The majority of moderators did not show differences in the magnitude of associations, though measurement perspective, pretest control, teacher gender, and teacher ethnicity showed small but consistent effects. Researchers, policy makers, teachers, administrators, students, parents, and others involved in schooling can advocate for increasing the awareness and practice of positive learner-centered relationships.

	Figures and Tables

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 

	Footnotes

 
JEFFREY CORNELIUS-WHITE is an assistant professor and director of school counseling at Missouri State University, 901 National Avenue, Springfield, MO 65897; e-mail: jornelius-white{at}missouristate.edu. He is editor of the Person-Centered Journal, associate editor of the Journal of Border Educational Research, and a board member of the World Association for Person-Centered and Experiential Psychotherapy and Counseling. His research typically focuses on person-centered and social justice education and counseling.

	References

TOP Abstract Method Results Discussion Conclusion Figures and Tables References

 

• American Psychological Association. (1997). Learner-centered psychological principles: A framework for school redesign and reform. Washington, DC: Author
• *Amos, LW, & Purkey, WW. (1988, June). Teacher practices and student satisfaction in dental hygiene programs. Dental Hygiene, 286-294
• *Asch, MJ. (1951). Nondirective teaching in psychology: An experimental study. Psychological Monographs, 65(4). 1-24[Medline] [Order article via Infotrieve]
• *Aspy, DN. (1965). A study of the three facilitative conditions and their relationships to the achievement of third grade students (Doctoral dissertation, University of Kentucky, 1965). Dissertation Abstracts International, 30(05A). 1853
• *Aspy, DN. (1972). Toward a technology for humanizing education. Champaign, IL: Research Press
• *Aspy, DN, & Hutson, B. (1972). Promotion of student success. Journal of Educational Research, 66, 57-60[ISI]
• *Aspy, DN, & Roebuck, FN. (1972). An investigation of the relationship between student levels of cognitive functioning and the teacher’s classroom behavior. Journal of Educational Research, 65, 365-368[ISI]
• *Aspy, DN, & Roebuck, FN. (1975). The relationship of teacher-offered conditions of meaning to behaviors described by Flanders’ interaction analysis. Education, 95, 216-222[ISI]
• *Aspy, DN, & Roebuck, FN. (1977). Kids don’t learn from people they don’t like. Amherst, MA: Human Resource Development Press
• *Aspy, DN, & Roebuck, FN. (1982, April). Affective education: Sound investment. Educational Leadership, 39, 488-493[ISI]
• *Aspy, DN, & Roebuck, FN. (1984). Productivity in education. In Aspy, DN, Roebuck, FN, & Aspy, CB (Eds.), The third century in American education: Education in the age of productivity, (p. 91-98). Amherst, MA: Human Resource Development Press
• *Aspy, DN, Roebuck, FN, & Black, B. (1972). The relationship of teacher-offered conditions of respect to behavior described by Flanders’ interaction analysis. Journal of Negro Education, 41, 370-377[CrossRef][ISI]
• Axline, VM. (1947). Play therapy: The inner dynamics of childhood. Boston: Houghton Mifflin
• *Bensley, ML. (1970). Value enhancement for children through non-directive in-service teacher training (Doctoral dissertation, U.S. International University, 1970). Dissertation Abstracts International, 31(05A). 2224
• *Berenson, DH. (1971). The effects of systematic human relations training upon classroom performance of elementary students. Journal of Research and Development in Education, 4, 70-85[ISI]
• Bergin, AE, & Garfield, SL. (1994). Handbook of psychotherapy and behavior change (4th ed). New York: John Wiley
• *Bernieri, J. Frank. (1991). Interpersonal sensitivity in teaching interactions. Personality and Social Psychology Bulletin, 17, 98-103[Abstract]
• Beutler, LE, Rocco, F, Moleiro, CM, & Talebi, H. (2001). Resistance. Psychotherapy: Theory/Research/Practice/Training, 38, 431-436
• *Birch, SH, & Ladd, GH. (1997). The teacher-child relationship and children’s early school adjustment. Journal of School Psychology, 35, 61-79[CrossRef][ISI]
• *Boak, RTR, & Conklin, RC. (1975). The effect of teachers’ levels of interpersonal skills on junior high school students’ achievement and anxiety. American Educational Research Journal, 12, 537-549[Free Full Text]
• Bowlby, J. (1969). Attachment and loss. New York: Basic Books
• Bozarth, JD. (1998). Person-centered therapy: A revolutionary paradigm. Ross-on-Wye, UK: PCCS Books
• *Brekelmans, M, Wubbels, T, & Creton, H. (1990). A study of student perceptions of physics teacher behavior. Journal of Research in Science Teaching, 27, 335-350[CrossRef][ISI]
• Carkhuff, RR. (1983). IPS: Interpersonal skills and human productivity. Amherst, MA: Human Resource Development Press
• *Christensen, CM. (1960). Relationships between pupil achievement, pupil affect-need, teacher warmth, and teacher permissiveness. Journal of Educational Psychology, 51, 169-174[CrossRef][ISI]
• *Cogan, ML. (1958). The behavior of teachers and productive behavior of their pupils: I. "Perception" analysis and II. "Trait" analysis. Journal of Experimental Education, 27, 89-124[ISI]
• Combs, AW. (1962). Perceiving, behaving and becoming: A new focus for education. Washington, DC: Association for Supervision and Curriculum Development
• *Combs, AW, & Taylor, C. (1952). The effect of the perception of mild degrees of threat on performance. Journal of Abnormal and Social Psychology, 47, 420-424[CrossRef][ISI]
• *Conners, CK, & Eisenberg, L. (1966). The effect of teacher behavior on verbal intelligence in Operation Headstart children. Baltimore: John Hopkins University School of Medicine: (U.S. Office of Economic Opportunity Headstart Contract No. 510)
• Cornelius-White, JHD. (2003). The analyzed nondirectiveness of a brief, effective person-centered practice. Person-Centered Journal, 10, 23-30
• Cornelius-White, JHD. (2006a). A learner’s guide to person-centered education. Person-Centered Journal, 13, 38-43
• Cornelius-White, JHD. (2006b). A review and evolution of Rogers’ theory of education . Manuscript submitted for publication
• Cornelius-White, JHD. Congruence. In Cooper, M, Schmid, PF, O’Hara, M, & Wyatt, G (Eds.), The handbook of person-centred therapy. New York: Palgrave Macmillan: (in press)
• Cornelius-White, JHD, & Brown, RD. (2006). Politicizing school reform through the person-centered approach: Mandate and advocacy. In Cooper, M, Malcolm, B, Proctor, G, & Sanders, P (Eds.), Politicizing the person-centered approach: An agenda for social change, (p. 263-269). Ross-on-Wye, UK: PCCS Books
• Cornelius-White, JHD, & Cornelius-White, CF. (2004). Diagnosing person-centered and experiential psychotherapies: An analysis of the PCE 2003 programming. Person-Centered and Experiential Psychotherapies, 3, 166-175
• Cornelius-White, JHD, & Cornelius-White, CF. (2005). Nondirective education. In Levitt, BE (Ed.). On being nondirective, (p. 314-323). Ross-on-Wye, UK: PCCS Books
• Cornelius-White, JHD, & Godfrey, P. (2004). Pedagogical crossroads: Integrating feminist critical pedagogies and the person-centered approach to education. In Proctor, G, & Napier, M (Eds.), Encountering feminism: Intersections of feminism and the person-centered approach, (p. 166-178). Ross-on-Wye, UK: PCCS Books
• Cornelius-White, JHD, & Harbaugh, AP. Learner-centered instruction: Building relationships for student success. Thousand Oaks, CA: Sage: (in press)
• Cornelius-White, JHD, Hoey, A, Cornelius-White, CF, Motschnig-Pitrik, R, & Figl, K. (2004). Person-centered education: A meta-analysis of care in progress. Journal of Border Educational Research, 3, 81-96
• *Crick, RD, & McCombs, BL. (2004, April, The case for learner-centered practices across the K-12 and college level: Enhancing the research credibility of the ALCP’s by an international study of upper elementary through seconding students in Bristol, England, Paper presented at the annual meeting of the American Educational Research Association, San Diego, CA
• *Daniels, DH, Kalkman, DL, & McCombs, BL. (2001). Young children’s perspectives on learning and teacher practices in different classroom contexts: Implications for motivation. Early Education & Development, 12, 253-273[CrossRef]
• *Diskin, P. (1956). A study of predictive empathy and the ability of student teachers to maintain harmonious relations in selected elementary classrooms (Doctoral dissertation, University of Michigan, 1956). Dissertation Abstracts International, 16(08). 1399
• *Dixon, WR, & Morse, WC. (1961). The prediction of teaching performance: Empathic potential. Journal of Teacher Education, 12, 322-329[CrossRef][ISI]
• Elliot, R. (2002). The effectiveness of humanistic therapies: A meta-analysis. In Cain, DJ, & Seeman, J (Eds.), Humanistic psychotherapies: Handbook of research and practice, (p. 57-81). Washington, DC: American Psychological Association
• *Elmore, PB, & LaPointe, KA. (1975). Effect of teacher sex, student sex, and teacher warmth on the evaluation of college instructors. Journal of Educational Psychology, 67, 368-374[CrossRef][ISI]
• *Faw, V. (1949). A psychotherapeutic method of teaching psychology. American Psychologist, 4, 104-109[CrossRef]
• *Finn, JD. (1993). School engagement and students at risk. Washington, DC: U.S. Department of Education, Office of Educational Research and Improvement, National Center for Education Statistics
• *Fortune, JC. (1967). A study of the generality of presenting behavior in teaching. Memphis, TN: Memphis State University: (ERIC Document Reproduction Service No. ED 016 285)
• Fraser, BJ, Wahlberg, HJ, Welch, WW, & Hattie, JA. (1987). Syntheses of educational productivity research. International Journal of Educational Research, 11, 144-252
• Gilligan, C. (1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press
• *Gross, L. (1948). An experimental study of the validity of the non-directive method of teaching. Journal of Psychology, 26, 243-248
• Hattie, JA. (1999, August 2). Influences on student learning. New Zealand: Inaugural lecture, University of Auckland
• *Hefele, TJ. (1970). The role of teacher interpersonal communication factors in the graduate education of prospective teachers of the deaf and the influence of such factors on the academic achievement of deaf students (Doctoral dissertation, University of New York at Buffalo, 1970). Dissertation Abstracts International, 31(09A). 4598
• *Hefele, TJ. (1971). The effects of systematic human relations training upon student achievement. Journal of Research and Development in Education, 4(2). 52-69
• *Hoeder, J, Joost, H, & Klyne, P. (1975). Zusammenhange zwischen Hauptdimensionen des Lehrerverhaltens und Merkmalen des Erlebens von Schulern im Unterricht. Psychologie in Erziehung und Unterricht, 22, 88-96[ISI]
• *Hoeder, J, Tausch, R, & Weber, A. (1979). Die qualitat det schulerbeitrage im unter-richt und ihr zusammenhang mit 3 peronenzentrierten haltungen ihrer lehrer. Zeitschrift fuer Entwicklungspsychologie und Paedagogische Psychologie, 11(3). 232-243
• *Howe, RW. (1964). The relationship of learning outcomes to selected teacher factors and teaching methods in tenth grade biology classes in Oregon (Doctoral dissertation, Oregon State University, 1964). Dissertation Abstracts International, 25(01). 0190
• Hunter, JE, & Schmidt, FL. (2004). Methods of meta-analysis: Correcting error and bias in research findings. Thousand Oaks, CA: Sage
• Johnson, DW, & Johnson, RT. (2001, April, Teaching students to be peacemakers: A meta-analysis, Paper presented at the annual conference of the American Educational Research Association, New Orleans, LA. (ERIC Document Reproduction Service No. ED 460 178)
• *Joost, H. (1978). Forderliche Dimensionen des Lehrevenhaltens im Zusammenhang mit emotionalen und kognitiven Prozessen bei Schulern. Psychologie in Erziehung und Unterricht, 25, 69-74[ISI]
• *Kendrick, HK. (1988). The relationship between invitational teaching behaviors and students’ affective outcomes in the clinical setting. Louisville, KY: Bellarmine College: Unpublished master’s thesis
• *Kratochvil, DW, Carkhuff, RR, & Berenson, BG. (1969). Cumulative effects of parent and teacher offered levels of facilitative conditions upon indices of student physical, emotional, and intellectual functioning. Journal of Educational Research, 4, 161-164[Medline] [Order article via Infotrieve]
• Lambert, NM. (1998). In McCombs, B (Ed.). How students learn: Reforming schools through learner-centered education. Washington DC: American Psychological Association
• *Lambeth, CR. (1980). Teacher invitations and effectiveness as reported by secondary students in Virginia (Doctoral dissertation, University of Virginia, 1980). Dissertation Abstracts International, 41(05A). 1923
• *Landsman, T. (1950). An experimental study of a student-centered teaching method. Syracuse, NY: Syracuse University: Unpublished doctoral dissertation
• *Lauer, PA, & McCombs, BL. (1998, July, Self-assessment and reflection tools for enhancing learner-centered practices in higher education: Validation data and their implications, Paper presented at International Conference on Improving University Teaching, Dublin, Ireland
• *Lewis, WA, Lovell, JT, & Jessee, BE. (1965). Interpersonal relationship and pupil progress. Personnel and Guidance Journal, 44, 396-401[ISI]
• Mackay, HC, Barkham, M, Rees, A, & Stiles, W. (2003). Appraisal of published reviews of research on psychotherapy and counseling with adults 1990–1998. Journal of Consulting and Clinical Psychology, 71, 652-656[CrossRef][ISI][Medline] [Order article via Infotrieve]
• *McCombs, BL. (2003a). How effective is Community for Learning (CFL) as an example of a learner-centered program? Philadelphia, PA: Temple University, Center for Research on Human Development and Education
• McCombs, BL. (2003b). Providing a framework for the redesign of K-12 education in the context of current educational reform issues. Theory Into Practice, 42(2). 93-101[CrossRef][ISI]
• *McCombs, BL. (2004a, April, The case for learner-centered practices: Introduction and rationale for session, Paper presented at the annual meeting of the American Educational Research Association, San Diego
• McCombs, BL. (2004b). The learner-centered psychological principles: A framework for balancing a focus on academic achievement with a focus on social and emotional learning needs. In Zins, JE, Weissberg, RP, Wang, MC, & Walberg, HJ (Eds.), Building academic success on social and emotional learning: What does the research say? (p. 23-39). New York: Teachers College Press
• *McCombs, BL, & Dunlap, C. (2003). Year 3 report—Third party evaluation of the capacity building grant for schools implementing the Community for Learning (CFL) Program: Efficacy of training and technical assistance in creating learner-centered practices for improving student academic achievement and non-academic outcomes. Philadelphia, PA: Temple University, Laboratory for Student Success
• *McCombs, BL, & Lauer, PA. (1997). Development and validation of the Learner-Centered Battery: Self-assessment tools for teacher reflection and professional development. Professional Educator, 20, 1-21
• McCombs, BL, & Whisler, JS. (1997). The learner-centered classroom and school: Strategies for increasing student motivation and achievement. San Francisco: Jossey-Bass
• *McKeachie, WJ, Moffet, MM, & Daugherty, M. (1978). Effective teaching: Facilitative vs. directive style. Teaching of Psychology, 5(4). 192-193
• *Meece, JL, Herman, P, & McCombs, BL. (2003). Relations of learner-centered teaching practices to adolescents’ achievement goals. International Journal of Educational Research, 39, 457-475[CrossRef]
• *Meffert, H, Steinbach, I, & Tausch, AM. (1976). Anderungen bel kindergarnerinnen und deren kindergruppen durch elnen erzleher-kurs in psychologisch hilfreichem verhaltern. Psychologie in Erziehung und Unterricht, 23(6). 328-336[ISI]
• *Midgley, C, Feldlaufer, H, & Eccles, JS. (1989). Student/teacher relations and attitudes toward mathematics before and after the transition to junior high school. Child Development, 60, 981-992[CrossRef][ISI][Medline] [Order article via Infotrieve]
• *Morgan, GA. (1972). Effects of a less prescriptive, student-centered college curriculum on satisfaction, attitudes, and achievement. Washington, DC: U.S. Department of Health, Education, and Welfare, Office of Education: (ERIC Document Reproduction Service No. ED 069 798)
• *Motschnig-Pirtnik, R. (2004, April, Person-centered e-learning in a major academic course: What are the results and what can we learn from them? Paper presented at 4th International Conference on Networked Learning, Lancaster, UK
• Noddings, N. (1984). Caring, a feminine approach to ethics and moral education. Berkeley: University of California Press
• Noddings, N. (2003). Caring, a feminine approach to ethics and moral education (2nd ed.). Berkeley: University of California Press
• *Ozer, EJ, Weinstein, RS, Maslach, C, & Siegel, D. (1997). Adolescent AIDS prevention in context: The impact of peer educator qualities and classroom environments on intervention efficacy. American Journal of Community Psychology, 25, 289-323[CrossRef][ISI][Medline] [Order article via Infotrieve]
• Patterson, CH. (1973). Humanistic education. Englewood Cliffs, NJ: Prentice Hall
• *Perry, K, & Daniels, DH. (2004, April, Elementary students’ perceptions of teachers and school adjustment, Paper presented at the annual meeting of the American Educational Research Association, San Diego, CA
• *Pianta, RC. (1994). Patterns of relationships between children and kindergarten children. Journal of School Psychology, 32, 15-31[CrossRef][ISI]
• *Pianta, RC, & Steinberg, M. (1992). Teacher-child relationship and the process of adjusting to school. In Pianta, RC (Ed.). Beyond the parent: The role of other adults in children’s lives, (p. 61-80). San Francisco: Jossey-Bass
• *Pianta, RC, Steinberg, MS, & Rollins, KB. (1995). The first two years of school: Teacher-child relationship and deflections in children’s classroom adjustment. Development and Psychopathology, 7, 295-312[ISI]
• *Pierce, JW. (2001, April, Changing practices in higher education: Facilitating the development of learner-centered faculty, Paper presented at the annual meeting of the American Educational Research Association, Seattle, WA
• *Pierce, JW, Holt, JK, & Kolar, C. (2004, April, Testing the learner-centered model with data at the college level, Paper presented at the annual meeting of the American Educational Research Association, San Diego, CA
• *Pierce, JW, Kalkman, D, & Dean, C. (2002, April, Validating the learner-centered model with undergraduates in teacher education, Paper presented at the annual meeting of the American Educational Research Association, New Orleans, LA
• Poplin, M, & Weeres, J. (1994). Voices from the inside: A report on schooling from inside the classroom. Claremont, CA: Institute for Education in Transformation at the Claremont Graduate School
• Public Agenda Foundation. (1994). First things first: What Americans expect from public schools. New York: Author
• Public Agenda Foundation. (2004). Teaching interrupted: Do discipline policies in today’s public schools foster the common good? New York: Author
• *Reed, HB. (1961). Teacher variables of warmth, demand and utilization of intrinsic motivation related to pupil’s science interest: A study illustrating several potentials of variance-covariance. Journal of Experimental Education, 29, 205-229[Medline] [Order article via Infotrieve]
• *Robinson, EH, Wilson, ES, & Robinson, SL. (1981). The effects of perceived levels of warmth and empathy on student achievement. Reading Improvement, 18, 313-318
• *Robinson, RY. (1995). Affiliative communication behaviors: A comparative analysis of the interrelationships among teacher nonverbal immediacy, responsiveness, and verbal receptivity on the prediction of student learning . (ERIC Document Reproduction Service No. ED 386 767)
• *Rocha, L. (1984). Educational productivity in Brazil. In Aspy, DN, Roebuck, FN, & Aspy, CB (Eds.), The third century in American education: Education in the age of productivity, (p. 99-102). Amherst, MA: Human Resource Development Press
• *Roebuck, FN, & Aspy, DN. (1974). G