Involving the teachers: Effects of a growth mindset intervention with teacher involvement

Last registered on February 01, 2019

Pre-Trial

Trial Information

General Information

Title
Involving the teachers: Effects of a growth mindset intervention with teacher involvement
RCT ID
AEARCTR-0003844
Initial registration date
January 31, 2019

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
February 01, 2019, 3:43 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Affiliation
University of Stavanger

Other Primary Investigator(s)

PI Affiliation
University of Texas - Auestin
PI Affiliation
University of Stavanger

Additional Trial Information

Status
On going
Start date
2018-08-01
End date
2025-12-31
Secondary IDs
The Research Council of Norway: 260407 and 227004
Abstract
Believing that intelligence and talents can be developed and changed through effort, perseverance and hard work is what psychologists refer to as a growth mindset (Dweck, 2006; Yeager and Dweck, 2012; Dweck and Leggett, 1988). Experiments have demonstrated that protocols from psychology, teaching students about the brain’s malleability, can help students develop a growth mindset and improve motivation and school outcomes (Aronson et al., 2002; Blackwell et al., 2007; Good et al., 2003; Paunesku et al., 2015; Yeager et al., 2014; Yeager et al., 2016). We build on the work by Bettinger et al. (2018) who adapted the intervention in Yeager et al. (2016) to the Norwegian context. To date, the mindset interventions with positive effects do not include teachers, which is critical for scaling up the interventions. Moreover, involving the teachers may enforce positive effects of mindset interventions by changing teachers’ practice. We address this gap in the literature by testing an intervention, which provides teachers with knowledge about growth mindset research, and gives them ownership of the intervention and guidance on how to foster a classroom climate that supports a growth mindset. Our hypothesis is that the intervention increases students’ believes in own abilities to learn and make them better at utilizing the learning opportunities provided to them (Bettinger et al., 2012; Carrell et al., 2016; Castleman and Page, 2015; Koch et al. 2015; Lavecchia et al., 2014). We expect that in turn, this will make students more likely to choose challenging assignments, show more interest in school activities, have higher expectations about own performance and increase their achievement in school (Burnette et al., 2013).

We test our hypothesis using a similar approach to Bettinger et al. (2018). First, we provide teachers with knowledge about growth mindset research and give them tools to help them foster a growth mindset among their students. These tools include the mindset intervention designed by Yeager et al. (2016) and adapted to the Norwegian context by Bettinger et al. (2018), boosters to remind students about growth mindset and guidelines on how to perform classroom discussions based on the intervention and the boosters. We then test whether these tools lead to increased incidence of growth mindsets among students using an experimental design with control and treatment groups. We measure students’ mindset using questionnaires both before and after the intervention. In addition to questions meant to detect students’ mindset, we ask the students questions to map their expectations about own performance, interest in school, behavior in school, attitudes towards school work, challenge seeking behavior and how they perceive their teacher. We also collect data on grades both before and after the intervention and behavioral outcomes in school (absence, passing, etc.). We use the data to compare students in the control and treatment group, and by doing this we get to uncover some of the mechanisms behind how growth mindset changes student behavior.
External Link(s)

Registration Citation

Citation
Auestad, May Linn, Mari Rege and David Yeager. 2019. "Involving the teachers: Effects of a growth mindset intervention with teacher involvement." AEA RCT Registry. February 01. https://doi.org/10.1257/rct.3844-1.0
Former Citation
Auestad, May Linn, Mari Rege and David Yeager. 2019. "Involving the teachers: Effects of a growth mindset intervention with teacher involvement." AEA RCT Registry. February 01. https://www.socialscienceregistry.org/trials/3844/history/40930
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
Our starting point is the web-based intervention and the same measures of mindset as Bettinger et al. (2018), who adapted the intervention and measures in Yeager et al. (2016) to the Norwegian high school context. However, in Bettinger et al. (2018), teachers were not a part of the intervention. In this trial, we include teachers by providing them with knowledge about the brain’s malleability, growth mindset, and tools that help them create a classroom climate that supports a growth mindset.

Together with the participating school districts, we first invite all teachers assigned to the treatment group to a full day seminar, with lectures about growth mindset and the intervention. We cover what past research has shown, why we should care about this in the Norwegian context, what to keep in mind when giving students feedback and characteristics of classrooms that support a growth mindset. In addition, we go through the entire intervention and the boosters together with the teachers and the teachers get to practice the classroom discussions. The goal of the seminar is to make the teachers confident in conducting the intervention with their students, and we encourage them to ask if they have any questions or concerns.

We provide the teachers with a detailed educational program with a description of every session they are going to perform, when each session is going to occur and the tools they need to perform the intervention with their students. The program consists of six separate sessions (five in the fall and one in the spring semester).

In the first session, the students log on through the school districts webpage and access the first of two web-based programs (the same as used in Bettinger et al., 2018). In this first session, students get to learn about what happens in the brain as you learn new things, why it is important to challenge yourself (especially during high school) and how a growth mindset have helped other students. In addition to looking at pictures and reading text, we ask students several questions during the program. The students answer these questions and hand in the answers to the teacher at the end of the session. These answers make up the basis for the classroom discussion in the next session.

Before the second session, the first of four classroom discussions, the teacher read through the students’ answers from the first session and create three to four questions to discuss in the classroom. We instruct the teachers to base these questions on the following themes: different learning strategies, it is normal to struggle, be frustrated and make mistakes when learning, and motivation to continue even though things are difficult. During the classroom discussions, students first discuss in small groups (3-4 students), before all students participate in a full class discussion. The goal of the discussions are that students get comfortable reflecting and talking about a growth mindset, that they get exposed to other students’ thoughts and ways of thinking, and that they get used to that struggling, getting frustrated and making mistakes are a normal part of the learning process.

The third session consists of the second of the two web-based programs. Similar to the first one, students log on through the school districts webpage. In this session, students get to learn about what scientists say about holding a growth mindset, what a growth mindset have helped other people do, examples of famous people who possesses a growth mindset and other high school students’ thoughts on growth mindset (explained in detail in Bettinger et al., 2018). Similar to the first web-based session students are asked questions, which they answer and hand in to the teacher at the end of the session.

In the fourth session, the second classroom discussion, teachers first read through the students’ answers on the questions from the second part of the web-based program (the third session). Based on these answers they create three to four questions for the classroom discussion (as in the second session). The themes we ask the teachers to base their questions on in this session are what students find important and how to encourage yourself and others to use a growth mindset. The layout and the goals of the session are the same as in the previous classroom discussion (in the second session).

Before the fifth and sixth session, we ask the teachers to give the students a hand-in assignment, a booster, which they are to read through and base the final two classroom discussion on. The layout and the goal of the sessions are the same as the previous classroom discussions. The theme in the fifth and sixth classroom discussion are “cooperation” and “after a holiday”, respectively. In addition to these overall themes of the session, we ask the teachers to make the questions based on the subthemes: it is normal to struggle, get frustrated and make mistakes when learning and the importance of effort, good strategies and to know when to ask for help.
Intervention Start Date
2018-09-20
Intervention End Date
2019-02-10

Primary Outcomes

Primary Outcomes (end points)
Mindset: In order to examine the effect of the intervention on student mindset we measure mindset the same way as Burnette et al. (2013), Yeager et al. (2016) and Bettinger et al. (2018) using previously used and validated measures of mindset.

Challenge seeking: We measure challenge seeking behavior similar to Bettinger et al. (2018) and Yeager et al. (2016). First, we ask the students, if given a choice between an easy and a more challenging assignment, which one they would choose. In addition, we have the students make their own worksheet, where they can choose between easy assignments, medium assignments or assignments that are more challenging. At the end of the survey, we randomly choose two of the chosen assignments for them to work on. The relationship between challenge seeking behavior and mindset have also been pointed out by Blackwell et al. (2007) and Mueller and Dweck (1998).

Achievement: Achievement is measured by having the students perform a math test consisting of ten math questions (similar to the math test used in Betting et al., 2018). In addition, we also collect data on national tests and grades in lower secondary school and grades in high school.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Interest: Interest is measured by asking the students to which degree (on a six-point scale) they agree with several statements regarding their emotions concerning activities in school (e.g. “I find schoolwork boring”, “I like the activities we are doing at school”, etc.). In addition, as in Bettinger et al. (2018), we also ask them more specifically about math with a question concerning to which degree they agree with the following statement: “Math in high school is interesting”.

Expectations: We measure expectations similar to how it is done in Bettinger et al. (2018), by asking students to which degree (on a six-point scale) they agree with statements about their goals in school (e.g. “My goal is to not make a fool of myself in class”), their thoughts if getting a bad grade on a math test (e.g. “This shows that I am not very good in math”) and the reasons they have for working hard at school (e.g. “I feel like I am doing something important for myself”). In addition, we also ask two questions specifically concerning math; “How do you think you will perform in high school math?” and to which degree they agree with the statement “I get insecure and anxious when thinking about math”.

Behavioral outcomes: We collect measures to look at behavioral outcomes in school from the school districts in the end of the school year (following the intervention). Behavioral outcomes include passing first grade in high school, days and hours of absence and grades in order and behavior both in high school and in lower secondary school.

Behavior in class: Behavior in class is measured by asking the students to which degree (on a six-point scale) they agree with statements concerning their behavior in class (e.g. “I refuse to do what the teacher tell me to”, “I am concentrating on what I am supposed to learn”, etc.)

How the students think others perceive them: How students think that others perceive them are measured by asking them to which degree (on a six-point scale) they agree with the several statements concerning this (e.g. “I think others perceive me as helpful”, “I think other perceive me as friendly”, etc.).
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We investigate the effect of the mindset intervention in an RCT in two participating school districts. Within each school district, we invited all high school teachers responsible for one first grade class (also referred to as contact teachers) to participate in the trial. 208 teachers, responsible for 173 different classes, from 32 schools signed up. Within each school, it varies how many teachers signed up for the project, but the requirement for signing up was that at least two teachers within the same educational program within the same school (responsible for separate classes) had to sign up in order to participate (if there were more than one class within the educational program within the school). After signing up for the research project, the participating teachers had their students complete a survey. Within each block (a block corresponding to all participating teachers within the same educational program within the same school), we randomly assigned the teachers to control and treatment group and informed the teachers about this following the completion of the survey.

We invited all participating teachers assigned to the treatment group to a full day seminar, where they got information about growth mindset, tools to use in order to develop a growth mindset and input on how to create a classroom climate that supports a growth mindset (explained under intervention). The purpose of the full day seminar was to get the teachers to take ownership of the intervention and provide them with guidance on how to perform the intervention with their students. Following this full-day seminar, teachers in the treatment group went through the mindset intervention with their students, consisting of six sessions spread out over five months during the first year of high school (explained under intervention). We matched students’ treatment status with their teacher’s treatment status, meaning that all students of the teachers assigned to the treatment group got access to the mindset intervention. Teachers and students in the control condition continued as originally planned, without any seminar or intervention.

Before the full day seminar (in September 2018) we had all the students of the participating teachers (both teachers in the control and treatment group) answer a set of questions in a survey, a pre-test, and after the intervention (in February/March 2019) we ask the students the same questions again, in a post-test. In the post-test, we also asked the students some additional questions about their teacher and some math specific questions. We do this in order to be able to investigate the effect of the mindset intervention.

We expect to experience some attrition between the pre- and the posttest because of absence being a major concern in Norwegian high schools. However, we do not expect this to be correlated with treatment status, and will check this in a similar balance test as used in Bettinger et al. (2018).

We conduct both the pre- and the post-test using Qualtrics surveys, and the students log on using their high school login on the school districts webpage and are then redirected to the survey. Before starting the pre-test (in August) the students were asked to give an informed consent in order to participate in the research project. This means that they got information about what it means to consent to the research project, and what to do if they at some point want to withdraw their consent. They also got information that whether they chose to consent or not would not influence their relationship with their teacher or others. If they did not consent they still got the treatment, but we are not going to collect any data on them.

Following the completion of the intervention, we will collect information about students’ test scores on the national tests and grades in lower secondary school and their grades in high school from the school districts. In addition, we will also collect information about behavioral outcomes in school (e.g. passing, days and hours of absence and behavioral grades) both in lower secondary school and in high school.
Experimental Design Details
Not available
Randomization Method
Prior to randomization, we divided teachers into blocks. Each block consisting of all participating teachers within the same educational program within the same school. Randomization was conducted within each block by assigning a random number (using a random number generator) to each teacher. After sorting teachers based on this random number, the first teacher within the block was assigned to the treatment group, and the second to the control group, etc. If there were more than two teachers within the block, we also sorted based on gender of the teacher, meaning that if there were two female teachers and one male teacher within a block, we randomized between the two female teachers. The excess teachers (those without any parings in the block) were placed in a residual group and randomized again (to get approximately the same number of teachers in the control and treatment group). This was done in the office by a computer with a witness.
Randomization Unit
Classes (or teachers)
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
Two school districts – 32 schools – 80 blocks of teachers (that is teachers within the same educational program and school).
Sample size: planned number of observations
About 3000 students.
Sample size (or number of clusters) by treatment arms
We have about 1500 students in the treatment and 1500 students in the control group.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We have 32 schools in our sample, the average number of students within a class is 17 and the average number of clusters within a school is 2. In each cluster, there is at least two classes. We use a 95 percent confidence interval and calculate the minimum detectable effect size using 80 percent power. Without any controls, the minimum detectable effect size is 0.20 of a standard deviation. If we include control variables and account for the fact that the educational program and school (the block) may explain some of the variations, lower effect sizes could be detected.
IRB

Institutional Review Boards (IRBs)

IRB Name
NSD Data Protection Official for Research
IRB Approval Date
2018-08-15
IRB Approval Number
61219