Exploring Teacher Effects for Mindset Intervention Outcomes in Seventh-Grade Science Classes

Dweck and others found that significant numbers of school-age children believe that ability is fixed, particularly in science, technology, engineering, and mathematics areas, and that these beliefs predict achievement (Dweck, 2006; Hill, Corbett, & St. Rose, 2010). Incremental theories of intelligence (growth mindsets) have been found to predict greater achievement and effort in school than entity theories (fixed mindsets) from early childhood through college (Blackwell, Trzesniewski, & Dweck, 2007; Dweck, 2008; Yeager & Dweck, 2012). Much of the prior research has been conducted in the context of mathematics; we extended those findings to the context of science during middle school.

In multiple lab studies, researchers have shown that mindset can be changed (see Dweck, 1999). Those lab studies led to attempts to promote growth mindsets among students in schools. A mindset intervention with seventh graders, which was similar to the one used in this study, was successful at influencing students’ beliefs about the malleability of intelligence, increasing their mastery goal orientation, and improving their mathematics grades. Mastery goal orientation refers to the degree to which students take on academic tasks with the goal of learning something new, developing skills, and improving understanding. Mastery goal orientation is often contrasted with performance goal orientation, which refers to a focus on demonstrating one’s ability or competence, and a concern with how one’s ability will be judged compared to others (Ames, 1992; Dweck & Leggett, 1988; Elliott & Dweck, 1988; Nicholls, 1984). Students with a growth mindset tend to adopt a mastery orientation when approaching academic tasks (Dweck, 1999).

Our first study showed that middle school student participants in the mindset intervention developed more of a growth mindset (pre to post) than did students in the control group (Schmidt, Kackar-Cam & Shumow, 2015). For example, participants developed significantly stronger beliefs about the malleability of intelligence in science over the course of the intervention, while students in the control group did not. There was also a significant change in mastery goal orientation as a result of the intervention; students who participated in the intervention reported an increase in mastery goal orientation over the course of the intervention whereas nonparticipants in the control group reported a decrease in mastery goal orientation. The focus of this study is the beliefs and practices of two teachers in whose classrooms (n = 7) the intervention was conducted, and the beliefs of their students at follow-up (several months after the intervention).

The first purpose of the present study was to investigate whether the student outcomes associated with the intervention differed by teacher. Researchers have not yet fully considered the role that the teacher plays in implementing mindset interventions in classrooms, particularly in domains such as science. There are several reasons to expect that the impact of classroom interventions on student outcomes will vary by teacher. First, teacher characteristics such as educational background, experience, and beliefs might add to or detract from the impact of the intervention. Second, the instructional and classroom management practices that teachers use in their classrooms have been shown to contribute to student achievement and motivation in ways that might amplify or lessen the effects gained from participation in an intervention (Muijs, 2008; Patrick, Mantzicopoulos, & Sears, 2012). Variation in the extent to which the teachers reinforce, elaborate on, and send messages about mindset and other concepts related to the intervention during daily instruction also is expected to boost or curtail outcomes of the intervention (Cimpian, Arce, Markman, & Dweck, 2007; Kamins & Dweck, 1999; Mueller & Dweck, 1998). The second purpose of this study was to investigate such differences between the two teachers.

Teacher characteristics. There has been a tremendous amount of research conducted on whether teacher characteristics like years of experience, amount and quality of education, and certification impact students (Harris & Sass, 2011; Nye, Konstantopoulos, & Hedges, 2004). We asked teachers to provide such background information about themselves because those qualities continue to be considered important teacher characteristics despite the fact that studies of the impact of such characteristics on student motivation and achievement have yielded weak, mixed, and sometimes contradictory results (Kennedy, 2010).

Teacher beliefs, another type of teacher characteristic, have also been studied extensively (see Bryan, 2012 for a review of science teachers’ beliefs). In this study, we measured teachers’ mindset beliefs. Teachers’ mindset beliefs are likely to influence their students’ mindset beliefs through the teaching practices they employ and through the quality of their interactions with students. Teacher promotion of learning strategies, for example, is an important teaching practice especially in middle school, because learning strategies mediate the relationship between middle schoolers’ motivational orientations and academic achievement (McClintic-Gilbert, Henderlong Corpus, Wormington, & Haimovitz, 2013). In a laboratory experiment by Good, Rattan, and Dweck (2007, cited in Dweck, 2008), teachers who had been influenced to believe in a growth mindset in mathematics were more likely to encourage failing students to work harder and to recommend specific learning strategies that would help them to improve. On the other hand, teachers who had been influenced to believe in a fixed mindset tended to comfort students who had failed by telling them that some students are good in mathematics and others are not, thus reinforcing a fixed mindset. Individual differences within the group of teachers who had been influenced to believe in a growth mindset were not analyzed in that study.

Teachers’ beliefs about how to best motivate students to learn are likely to impact their instructional practice. The present study examined teachers’ beliefs about the effectiveness of various approaches that are most aligned with mastery and performance goal orientations. A recent study (Shim, Cho, & Cassady, 2013) found that teachers’ mastery and performance approach goal orientations for teaching predicted which type of goal orientation they established in their classrooms. Thus, we gathered indicators of teachers’ beliefs before we turned to investigating their actual teaching practices.

Instructional practices and interactions. Teachers exert influence on student motivation and achievement through the instructional practices they use, the feedback they give students, and other day-to-day interactions with students (Stipek, 1996). Middle school students whose teachers coconstruct learning experiences with them in a supportive classroom environment demonstrate a sense of academic progress, show increased engagement with academic tasks, and are more cooperative with peers and teachers (Strahan, Faircloth, Cope, & Hundley, 2007). It stands to reason that the classroom climate and the learning context that teachers and their students cocreate will impact the effectiveness of any intervention that is introduced with the intent of improving student outcomes.

Teachers who facilitate a positive emotional climate, organize and manage the classroom effectively, and express enthusiasm provide a context in which student learning and motivation flourishes and in which students are primed to cooperate and participate in lessons (Hattie, 2009; Patrick et al., 2012). In the present study, researchers conducted multiple classroom observations during times that the district science curriculum was being taught: Observers recorded global ratings of the classroom climate, organizational management, and teacher enthusiasm during the class period.

At the most basic level, student learning is impacted by how time is used in the classroom (Kyriakides & Creemers, 2008). In a recent study of high school science classrooms, our research team found that teachers used lecture and seatwork more than any other practice and that considerable time was spent in noninstructional activities like taking attendance, making announcements, or distributing and collecting papers (Schmidt, Zaleski, Shumow, Ochoa Angrino, & Hamidova, 2011). Furthermore, teacher-student interaction, students’ reported learning, and students’ motivational states varied by the type of instructional activity the class was engaged in (Shumow & Schmidt, 2013). In the current study, classroom observers recorded the types of instructional activities that teachers used, the amount of time students were engaged in various activities, and then rated each instructional activity on several dimensions, including: conceptual development, drill, and instructional feedback provided by the teacher. Each of these factors has been found to contribute to students’ perceptions of their ability, their learning goals, and their academic success (Hattie, 2009; Muijs, 2008; Patrick et al., 2012).

Mindset messages. Our interest in mindset led us to focus specifically on mindset messages within classrooms. Dweck and her colleagues have found that, in the years from preschool through college, the messages students receive from teachers impact their mindsets, their goal orientation, and, consequently, their academic achievement (Cimpian et al., 2007; Kamins & Dweck, 1999; Mueller & Dweck, 1998). For example, when teachers praised students for their intelligence or talent and made ability comparisons among students, students developed performance goal orientations (Dweck, 2007, 1999; Patrick et al., 2012). On the other hand, teachers who recognized students’ effort and study skills helped students develop a growth mindset, mastery goals, and tenacity in the face of challenge (Dweck, 2008, 2010). In the present study, researchers recorded and coded events in which teachers communicated mindset messages to ascertain possible variation in those messages.

Students’ beliefs about the nature of ability have been linked to a variety of motivational and achievement outcomes. In the context of a mindset intervention, our research team investigated whether student beliefs about the malleability of intelligence, their goal orientation, and their achievement in science improved more in one science teacher’s classes than in another’s. Characteristics and practices of the teachers are then compared using multiple sources of data to understand features of classroom context that may enhance the efficacy of classroom interventions designed to impact students’ beliefs about science ability.

The larger study from which these data were drawn was conducted in 14 middle school science classrooms in a diverse, public school district, and included 363 seventh graders and 4 teachers. Two of these teachers participated in a mindset intervention, and are the focus of the present study. Together, these teachers were responsible for 7 seventh grade science classes. The other 2 teachers (7 classrooms) participated in a different intervention and are not included in the present analyses.

The school district in which the study took place was located on the fringe of a large metropolitan area. Sixty percent of students in the school district were considered “low income.” The student population in the district was over 50% Hispanic (specific sample characteristics are provided below).

Students in seven classrooms (n = 160) and their teachers (n = 2) participated in a mindset intervention. The student sample for the study was 42% male, and 58% female. Racial and ethnic distribution was as follows: 34.4% Hispanic, 13.8% Black, 25% White, and 22.5% multiracial (4% did not report race/ethnicity). Fifty percent of the student sample received free or reduced lunch.

Teachers. We refer to the teacher participants as Celia and Donna (both are pseudonyms). Celia taught three of the intervention classes and Donna taught four classes. Donna was a White female who was 54 years old at the time of the study. She held a master’s degree and had 20 years of teaching experience (18 years at her current school); she had taught sixth-, seventh-, and eighth-grade science classes. Celia, who was also a White female, had been Donna’s student teacher. At the time of the study, Celia was 28 years old and had been at the school for 6 years, which comprised her total teaching experience. She held a bachelor’s degree and had experience teaching sixth-, seventh-, and eighth-grade science classes. Donna and Celia taught in different schools in the same district.

Intervention. The mindset intervention consisted of the Brainology program—a 6 week, web-based tutorial that teaches students that the brain is like a muscle and can grow in intelligence. Students are provided with information about how the brain responds to learning and how they can improve their learning through effort, study strategies, and behavioral choices (sleep, diet, and stress management). Researchers met with student participants once per week and took primary responsibility for delivering the web content and for leading both anticipatory and follow-up activities to each lesson. Teachers were always present during these “Brainology days.”

The program also included teacher education activities; teachers met on several occasions with the researchers for individualized sessions to discuss the mindset concept, how mindset develops, and practical details about implementation including discussion of extension activities to be done in class. The central role of the teacher in fostering mindset and in emphasizing the concepts during instruction and through statements made to students was emphasized. Teachers received a teacher’s manual containing numerous extension activities, online access to the Brainology program and to students’ work in the program, and two brief (3-5 page) supplementary readings on mindset.

Data collection. Students completed surveys that measured their beliefs about the malleability of intelligence, goal orientation, and interest in science before (pre), immediately after (post), and several months after (followup) the intervention. Prior to the start of data collection in classrooms, participant teachers completed a survey in which they provided information about their demographic characteristics, professional training, and current teaching assignment. The survey included questions in which teachers shared their beliefs about whether specific instructional practices (e.g., praising students, offering rewards) were effective in motivating students. Also included in the survey was a series of questions with established reliability and validity used by Blackwell et al. (2007) to assess mindset. These items exactly mirrored those in the student survey.

Classrooms were observed on 11 different occasions before, during, and after the intervention which was significantly more than the sufficient number of observations suggested by some studies (e.g., Shih, 2013) to effectively capture the qualities of classrooms. On each of these 11 occasions, a team of two to three trained observers recorded instructional activities and multiple dimensions of classroom context including event sampling of explicit and implied messages conveyed by teachers and students regarding goals, effort, evaluation, feedback, encouragement and study strategies. Observers were intentionally placed in different positions in order to capture student messages that might not be heard from across the classroom.

One of the observers was always a principal investigator. Principal investigators had extensive experience observing classrooms, and together they trained the other observers. Trainees received a field manual with detailed instructions, participated in several half-day training sessions, and practiced observing independently, using videotapes filmed in science classrooms similar to those participating in the study. Observers did not enter the field until they reached 90% or greater interrater reliability with the ratings the senior observation instructor had pre assigned on two classroom videos.

Reliability on classroom ratings among coders was high (see measures). Notes from all coders present were used to compile a comprehensive set of field notes documenting mindset messages expressed by teachers and students in the classroom. These field notes were later coded (see description of coding below).

Malleability of Intelligence. Four items were used to measure students’ beliefs about the malleability of intelligence or mindset. The items asked students to report on a 6-point scale (1 = disagree a lot, 6 = agree a lot) whether they believed it was possible to change one’s intelligence in science (two items) or whether science intelligence is fixed (two items which were reverse scored to create this variable). A factor analysis provided evidence of the construct validity of this subscale. Cronbach’s alpha for these items was: .60 in the initial survey, and .74 in both the post intervention and follow-up surveys. Items were drawn from published studies (Aronson, Fried & Good, 2002; Blackwell et al., 2007), which reported test-retest reliabilities ranging from .77 to .82, Cronbach’s alpha of .78, and good predictive validity.

Mastery goal orientation. A mastery goals scale was created from four items on the student survey (I do science work to learn new things, I want to work on hard science work, hard assignments mean I’ll learn, and my goal in science is to learn as much as possible). These items have been previously used by Blackwell et al. (2007) and Elliott and Murayama (2008) who obtained evidence of high internal consistency, minimal response bias, and good structural and predictive validity. Three of these items were measured on a 6-point scale (1 = disagree a lot, 6 = agree a lot), and one item was measured on a 5-point scale (1 = strongly disagree, 5 = strongly agree), therefore all scores were converted to z scores to create a composite score. Factor analysis and correlations with other variables provided evidence of construct and discriminant validity. Cronbach’s alpha was .79, .79, and .82 on the pre-, post-, and follow-up surveys respectively.

Achievement. School officials provided students’ quarterly grades in science from school records. First quarter grades served as the initial estimate of achievement, second quarter grades served as the postintervention estimate of achievement, and third quarter grades (which aligned with follow-up surveys and observations) served as the follow-up estimate.

Endorsement of motivational strategies. Teachers were asked to rate the efficacy of a number of different motivational strategies for males and for females on a scale of 1 (not at all effective) to 5 (very effective). Items on this rating list included both strategies that foster intrinsic or mastery goals and strategies that foster extrinsic goals. For the purpose of this study, the teacher ratings of strategies that were effective for males and females were averaged.

Global ratings by instructional activity. Observers were provided with a numbered list of classroom activities adopted from Duke (2000):

1. teacher presentation

2. individual seatwork,

3. small-group seatwork,

4. tests/quizzes,

5. whole-class discussion,

6. student presentations/demonstrations,

7. video/movie,

8. lab work,

9. noninstructional time,

10. off task-activity, and

11. activities related to the research study (e.g., completing brief surveys).

Each time one of these instructional activities was observed (excluding noninstructional, off task, and research study-related activities), the researchers made global ratings on three dimensions using a scale from 1 (almost none) to 4 (extensive). Conceptual development indicated the degree to which teachers promoted higher order thinking, elaboration (why, how, compare), and problem solving. Drill indicated the degree to which surface learning strategies like repetition were emphasized. Instructional feedback described the extent to which teachers supported and extended student learning through responses, scaffolding, promotion of student skills, and participation in activities. Spot checks of reliability of these ratings indicated good agreement (ranging from 88-94%).

Mindset messages. Observational eventsampled field notes were coded for the purpose of identifying specific teacher-provided messages related to mindset. For each teacher, we coded field notes from a total of 11 days per teacher: one day of regular instruction per week in each classroom for 2 weeks prior to the intervention, the 6 weeks in which the Brainology program was being implemented, and 3 weeks post intervention later in the school year. The day of the week we observed varied from week to week. Field notes were coded using the NVivo10 software program. Altogether, 29% of field notes were coded in pairs. After demonstrating greater than 90% agreement on what was a mindset message and greater than 85% agreement on the dimensions of that message, coders completed coding individually.

Mindset messages were identified as any explicit statement or behavior that referred to Brainology program content, task difficulty/ease, effort, study strategies, ability, or performance criteria, regardless of whether the reference explicitly mentioned mindset. Each mindset message was coded along multiple dimensions which recorded the nature of the messages as promoting or undermining a growth orientation. Messages that were coded as promoting a growth mindset specifically mentioned growth of intelligence, referenced Brainology content, emphasized effort, or suggested/modeled study strategies. Messages that were coded as undermining a growth mindset included those that clearly mentioned a fixed view of intelligence, valued low effort, and focused on task ease, difficulty, and ability without reference to effort. Once coding in NVivo10 was completed, data were analyzed using SPSS.

Malleability of intelligence. A mixed between-within subjects analysis of variance was conducted to assess whether the Brainology program impacted students’ beliefs about the malleability of intelligence differently by teacher. There was a significant interaction between teacher and time, Wilks’s λ = .91, F (2, 134) = 6.87, p =.001, partial η² = .09, indicating a moderate effect size for the difference in increase in belief about the malleability of intelligence by teacher. Figure 1 displays the results of this analysis. Table 1 presents the pretest, posttest and follow-up scores. As seen in the table, students of both teachers developed stronger beliefs about the malleability of intelligence after participating in the intervention, with Donna’s students increasing more than Celia’s. During the follow-up period, Celia’s students regressed nearly to the point where they had been prior to the intervention whereas Donna’s students fell slightly but nevertheless, maintained considerable gains.

Mastery goal orientation. A mixed between-within subjects analysis of variance was conducted to assess whether the Brainology program impacted students’ mastery goal orientation differently by teacher. There was a marginally significant interaction between teacher and time, Wilks’s λ = .96, F(2, 135) = 2.83, p =.06, partial η² = .04, indicating a small effect size for the difference in change in belief about mastery goals by teacher. As can be seen in Figure 2, students in Donna’s classroom increased in mastery goal orientation from pre to postintervention and maintained those gains through follow-up, whereas, students in Celia’s classroom increased from pre to postintervention, but decreased to a numeric level at follow-up that is below their preintervention mean. The pre and follow-up scores did not differ significantly, however. Means and standard deviations can be seen in Table 1.

Achievement. A mixed between-within subjects analysis of variance was conducted to assess whether the Brainology program impacted student achievement differently by teacher. There was a significant interaction between teacher and time Wilks’s λ = .79, F(2, 135) = 17.9, p = .000, partial η² = .21, indicating a large effect size for the change in students’ achievement by teacher from before intervention to follow-up. Figure 3 displays the results of this analysis. Table 1 presents the preintervention, postintervention, and followup scores on student assessments. Donna’s students’ achievement increased during the intervention and maintained those increases through follow-up whereas Celia’s students’ did not. Although Donna’s students’ mean achievement was higher numerically at followup than post intervention, the difference was not statistically significant.

Celia expressed deep interest in the mindset intervention; it was new information for her and she saw it as exciting. Donna was also committed to the project because she was familiar with the importance of the content students would learn during the intervention. She was also using her participation in the study as part of the professional development plan that her district required of all posttenured teachers.

Donna’s and Celia’s score on the mindset measure identified each of them as having a growth mindset in terms of science intelligence. As can be seen in Table 2, both teachers moderately endorsed practices associated with mastery goals (e.g., “teaching them strategies for learning”). Celia endorsed using strategies often associated with fostering a performance approach goal orientation (Patrick et al., 2012), while Donna endorsed such practices only weakly. For example, Celia moderately endorsed comparing students to one another and strongly endorsed telling students that they were one of the best in the class as a means of motivating them. Neither teacher recommended strategies such as embarrassing students for poor performance as being motivational. Celia was more affirming of the motivational value of using consequences like rewards and contacting parents than Donna was. Both teachers endorsed the motivational value of praise, which we interpreted as an approach that is not advantageous for motivating young adolescent students.

Instructional practices. The teachers both covered the state and district curriculum for seventh-grade science and the percent of time each teacher’s classes spent in various instructional activities was very similar. Ratings given by observers during the activities indicated that Donna was more likely than Celia to focus on drilling students and to facilitate students’ conceptual development (see Table 3). There were no differences in ratings for instructional feedback provided to the students.

Teacher involvement/role in mindset intervention. In discussing her role during the intervention, we asked Celia to use the first class period of each day in which the unit was implemented to familiarize herself with the lesson, what the students were asked to do, and to observe what they were actually doing (recall that the intervention was done in 3 of Celia’s classes). Further, she was asked to use the next two class periods to review the logs that displayed students’ responses to the unit lesson, which were available to her through the teacher login. Celia did not use the class time in that way, however. She was frequently observed using the computer to catch up on record keeping. Each week of the intervention, however, Celia extended the Brainology unit in her science classes by using a supplementary lesson from the guide provided for teachers by Mindset Works, the developer of the Brainology program. She also responded to several individual students who had strong fixed mindsets. She chose one male student in particular as her own special project; he manifested an extreme fixed mindset and she was determined to change it.

Donna was invited to participate in the same way as Celia was. Donna played a more active role with students on the intervention days than did Celia by monitoring the class and encouraging student engagement. We have little evidence that Donna monitored the logs of students’ work in the program, even though this information was available to her. We did not observe her accessing it and she did not talk about it with us. The content of the program was not new to Donna and, as will be seen in the next section, she applied the content, using the vocabulary during her teaching. In contrast to Celia, Donna appeared to take little interest in the teacher guide provided.

Mindset feedback when teaching science. Analysis of field notes suggests that the two teachers differed in terms of the way they communicated with their students about goals, effort, evaluation, feedback, encouragement, and study strategies. The reader will recall that field notes were taken on an ordinary instruction day—not on a day when the Brainology program was a focus. The purpose of these analyses was to examine the degree to which each teacher was supporting a growth mindset outside of the designated “Brainology days.” As seen in Table 4, Donna made more frequent references that could be construed as generally supporting a growth mindset by emphasizing the idea that ability can grow, focusing on mastery rather than performance goals, and highlighting the value of effort and strategy use. For instance, she told a male student who had not done as well as he wanted to on a test, “Did you pay attention in class? Did you do your homework? Think of how well you could have done if you made more of an effort.” Moreover, on average, Donna referenced Brainology once during every class period. On one occasion, for example, we observed her saying, “Remember, we are doing this to make more neural connections in your brain like we learned in Brainology last week”.

Celia’s comments, while well-intentioned, were unlikely to promote a growth mindset. During the class periods we coded, she rarely mentioned Brainology explicitly or referred to a strategy mentioned in the program. She was generally supportive of her students, and genuinely wanted to help them succeed, but her supportive comments did not emphasize effort. Rather than arming her students with strategies to help them complete their work as Donna did, Celia tended to alert students to whether particular tasks were “easy” or “hard.” For example, referring to a worksheet on speed and velocity, she told students, “I know this is boring but it’s easy peasy and you are going to be tested on it.” Another example of undermining was that when Celia’s students were challenged academically, her reaction was to offer immediate assistance, rather than emphasize the importance of challenge and effort for learning. This might send the message to students that they are incapable of addressing these challenges on their own. As well, she set up competitions between classes with rewards of doing less work or earning extra points. One day she told a class, “Do it (vocabulary) better than the other classes and you’ll only have to do it once.” On another occasion she said, “The other classes have been struggling with this. If you beat their score on the review game, I’ll give everyone five extra points on the quiz.”

In this study, we considered the role of the teacher in a widely used intervention to impact student mindset. There were significant teacher effects in that students’ beliefs about the malleability of intelligence, their learning goals, and their achievement improved and/or were sustained more in one teacher’s classes than in the other’s classes. The positive results of the intervention at follow-up compared to the control group (Schmidt et al., 2015) appear to be largely accounted for by Donna’s students. Thus, we found evidence that the teacher is an important factor in sustaining the positive outcomes of a popular mindset intervention.

The findings are especially salient because the intervention with the students was implemented by researchers. The teachers received some brief education about mindset, resources for learning about and fostering mindset, and access to each student’s Brainology log. Some teachers implement the intervention themselves and studies should be conducted to examine possible teacher effects in their students; it seems reasonable to assume that teacher influence would be an even greater factor in those cases than in this one.

There were considerable differences between the two teachers in terms of years and breadth of teaching experience and their educational backgrounds. The impact of those characteristics on student motivation and achievement has been difficult to substantiate in large-scale studies using general measures, but might be particularly important in specific contexts (Kennedy, 2010). In this small study, Donna, the teacher with more experience and education, was more effective in promoting growth mindset, mastery orientation and student achievement than was Celia. Donna was more often observed facilitating students’ conceptual development and growth mindset than Celia. Our finding warrants continued finegrained analysis of teacher characteristics and practices in the context of mindset interventions and expected outcomes.

In terms of teacher beliefs, both teachers had a strong growth mindset regarding intelligence in science on the survey measure. However, their practices reflected differences in the degree to which they promoted a growth mindset among their students. The teachers were very likely aware of the social desirability of endorsing a growth mindset, so the survey items might not have been an accurate measure of their beliefs. When it came to their beliefs about the motivational value of different practices, greater differences were found between the teachers. Celia endorsed a greater number of performance-oriented motivational practices than did Donna. For example, Celia reported to us at the outset of the study that she believed that comparing students to one another, emphasizing grades and points, and offering rewards for performance were all effective motivational strategies.

Donna and Celia planned and enacted similar activities, but differed in the way they communicated with their students about goals, effort, evaluation, feedback, encouragement, and study strategies. These differences are generally consistent with the differences in the two teachers’ beliefs about motivational practices noted above. Donna’s daily interactions with students implied an emphasis on mastery, learning, and growth more so than did Celia’s. Whereas Donna more often promoted conceptual development, modeled and encouraged strategy use, Celia’s classroom was rarely characterized this way and her lessons rarely promoted deep understanding. Celia was quick to help her students when they struggled with tasks, but did not often suggest strategies for dealing with these struggles, which could send the message to her students that she lacked confidence in their abilities.

Though they did not assume primary responsibility for delivering the Brainology program due to the nature of the study, the teachers supported the program in different ways. While Donna did not make much use of the program-provided supplementary materials, and did not often examine students’ logs, she was very much involved in students’ participation in the Brainology program itself. Celia, on the other hand made fairly regular use of the supplementary program materials, and occasionally consulted student logs, but she was completely uninvolved during the ongoing program itself. From the students’ perspective, Donna probably appeared to be more actively invested in the program.

Perhaps the most striking and impactful difference between the teachers was in their usage of mindset messages in their daily interactions with students. Donna interacted with her students in a way that promoted a growth mindset, while Celia did not. Even though Celia made greater use of the supplementary Brainology materials than Donna did, she failed to reinforce the development of adaptive beliefs about learning in the comments she made to her students. These small daily interactions may make a critical difference in the degree to which classroom interventions are effective, particular in the longer term. Donna supported the program more consistently in her daily routine than did Celia. While it is impossible to support this conclusion empirically given the current data, we believe these daily reinforcements are the likely reason that the intervention was ultimately more effective for Donna’s students than Celia’s.

This is a small intensive study conducted in classrooms that were taught by only two teachers. While the teachers differed from one another in terms of their age and years of experience, they were both women teaching in the same school district. The generalizability of these findings to the broader array of teacher and student populations is not currently known. It is possible that the teachers varied in ways other than those considered in the study, and that it was these differences, and not the ones studied here that impacted student outcomes. While the differences in students’ outcomes in the two teachers’ classrooms are undisputable, we cannot rule out alternative explanations as to why they differed from one another.

The measures used in this study, while replicating those used in other research, did not evidence particularly strong psychometric properties when used with the current study population. In particular, internal consistency for the variable measuring malleability of intelligence was rather low. As a result, findings regarding this particular outcome may be less robust than they would otherwise be with a more reliable measure. Future research should explore ways to improve measurement of this construct among this particular student population. The teacher sample size was not large enough to assess the internal consistency of the measures in the teacher surveys. While the items we used replicated previous research, we were unable to evaluate their psychometric properties with our sample. We suspect that survey items regarding the malleability of intelligence, may have suffered from strong social desirability bias. The teachers in the study likely perceived a “right” answer to questions about mindset, and may have answered in the way they believed they were supposed to. Both teachers endorsed beliefs consistent with a growth mindset. While this might be an accurate representation of their beliefs, it is also possible that these scores were inflated.

Despite the limitations of this study, the results suggest that teachers play a critically important role in supporting classroom interventions. Across the seven classrooms we studied, there were systematic differences by teacher with respect to student outcomes both immediately following the intervention, and at

the end of the year. Despite similarities in the teachers’ stated beliefs regarding growth mindset, one teacher clearly engaged in classroom practices that were consistent with this belief framework, while the other did not. When teacher behaviors were observed to be supportive of a growth mindset, students adopted stronger mindset beliefs and were more likely to maintain these beliefs over time. Program developers may want to design and study ways to impact teachers’ practices in order to maximize program impact. Implementing such programs without considering teachers may limit any enduring effects.