Study 1:  Mueller & Dweck (1998)

According to Dweck, our mindset has a big impact on how well we are able to achieve our potential—in school and in many other areas of our lives (for example, in sports, music, and business). But where do these differences in mindsets come from?

There can be many reasons that a person comes to believe that intelligence is fixed or changeable, but one obvious influence on our way of thinking about ourselves is the messages we hear from adults as we grow up. Dweck and her then-graduate student Claudia Mueller wanted to see if they could influence the mindset of children, even if only for a brief period of time, by selectively offering different kinds of praise. Praise is very motivating; when we do something and receive praise, we are more likely to want to do that same thing again. But Mueller and Dweck wondered if all praise was equal. In particular, they sought to investigate whether certain types of praise might actually reduce a child’s motivation to learn and erode that child’s resiliency when they encounter challenges.

The researchers recruited 128 fifth graders (70 girls and 58 boys ranging in age from 10 to 12) to participate in their study. Before we go into the details of the first experiment, please get a feel for the task that the children had to perform.

You will have one minute to solve as many of the problems below as you can.^[1] For each problem, you will see a set of patterns arranged in a 3×3 matrix. Each matrix has one item missing, and your task is to figure out what the missing item is based on the changing patterns in the rows, columns, and diagonals.

Before we start, here is one practice item. The 3×3 matrix is at the top and the pattern on the lower right is missing. Figure out which one of the eight patterns on the bottom, labeled 1 to 8, is the missing pattern.

Show Answer

The correct answer is pattern #7. The pattern on the right in each row combines the dots from the other two patterns in that row.

The problem-solving task you just tried out is based on a widely used psychological test called the Raven’s Progressive Matrices. Most people find the test to be challenging, requiring close attention to detail and careful logical thinking. Mueller and Dweck chose this task because it could be adapted to be relatively easy or extremely difficult by changing the complexity of the patterns required for the solution.

The experiment had three stages, each based around a different set of matrix problems like the ones you worked on. Each child was tested one-on-one in an otherwise empty classroom by a research assistant.

Stage 1: Pretest, Treatment, and Assessment of Motivation

Pretest

The children were given instructions and 10 problems that were fairly easy to solve. At the end of 4 minutes, they were stopped and the research assistant scored their answers. On average, the children attempted to answer 7.9 out of the 10 problems, and the mean number correct was 5.2.

Treatment

When you do something to manipulate an independent variable, that manipulation (administer a pill, tell the participant something that might affect performance, etc.) is called a “treatment.” In this case, the treatment was the feedback the child received about their performance on the progressive matrices task. This treatment involved a bit of deception because the children received randomly assigned feedback.

First, every child was told: “Wow, you did very well on these problems. You got _____ right. That’s a really high score.” Some mild deception was used here as well because all children were told that they got at least 80% correct, which is above the actual average score. If a child actually did solve more than 80% correctly, their actual score was used.

Then, children were randomly assigned to hear one of three statements:

• Some of the children were praised for their ABILITY: “You must be smart at these problems.”
• Other children were praised for their EFFORT: “You must have worked hard at these problems.”
• The remaining children were in the CONTROL condition. They did not receive any additional feedback, aside from the general praise shown above.

Assessment

After receiving feedback and, for children in two of the conditions, additional praise, the children were asked a series of questions. The experimenters wanted to know if the success the children experienced in the first set of problems, along with the type of praise they received, might influence the children’s choice of whether they would prefer to work on easier or more challenging problems in the future. The children were told they might get some additional problems to solve later on in the experiment, and they were asked to choose the difficulty of those additional problems. There were several options, but the choice essentially came down to:

• Give me easy problems: “Problems that I’m pretty good at, so I can show that I’m smart.”
• Give me challenging problems: “Problems that I’ll learn a lot from, even if I won’t look so smart.”

The children were then told that there might be some time at the end of the session to work on these problems they had chosen, but that the next problems they would work on had been determined before the experiment started. They were told this so they would not interpret the next problem set as being “easy” or “challenging” based on their selection.

The results showed that the children were genuinely influenced by the praise they had received. The figure below shows the percentage of children choosing EASY problems, broken down by treatment condition. The children who were praised for how smart they were (ability) were far more likely to choose easy problems than were the children praised for working hard (effort). The control condition, children who were told they did well but received no additional praise, were in the middle.

Stage 2: Failure, Negative Feedback, and Consequences

Failure

Next, the children tried to solve a new set of 10 matrix problems and again they had 4 minutes. On the surface, these problems looked about the same as the first set, but they were considerably more difficult. After the 4-minute test period, the researchers scored the answers and, regardless of actual performance, they told the children that they had done poorly (“a lot worse”). No one was told that they had solved more than 50% correctly. In fact, this feedback was often accurate. The results showed that the children found this second problem set difficult. On average, they attempted 5.8 of the 10 problems and correctly solved only 1.8 of them. There was no significant difference in the number of problems solved for the three groups (ability feedback, effort feedback, and no-feedback control).

Consequences

Now the experimenters wanted to know about the effect of “failure” on the children’s motivation (though the term “failure” was never used with the children).

Immediately after receiving feedback, the children were asked a series of questions:

• “How much would you like to take these problems home to work on?” [This was a measure of “task persistence”]
• “How much did you like working on the first set of problems? How much did you enjoy working on the second set? How much fun were the problems? [These measured “task enjoyment”]
• Using a somewhat more complicated measure, the children were also asked to explain their difficulties with the second problem set as being caused by either a lack of ability, a lack of effort, or some combination of the two.

Results

• “How much would you like to take these problems home?” The children answered on a 1-to-6 scale, where higher numbers meant more interest in taking the problems home to practice.

• “How much fun were the problems?” The children answered on a 1-to-6 scale, where higher numbers mean more enjoyment of the problems.

• Why did you perform poorly on this second set of problems? The children expressed their own explanation for their poor performance using a somewhat more complicated procedure. They were not simply asked to choose between ability vs. effort, but were also given the option to attribute their failure to a combination of the two (reference the original study for more details).

Stage 3: Posttest

For the last stage of the experiment, the children were given a new set of problems that was similar in difficulty to the first set. The problems were moderately difficult, and the children had 4 minutes to solve as many as possible. The figure below shows the change in the average number of problems correctly solved between the pretest (Stage 1, before feedback) and the posttest (Stage 3, after feedback).

Instructions: Click and drag the circles on the right (Posttest) to where you think they should be to reflect the results of the experiment. When you’re done, click the link below to see the actual results.

Click here to see the results.

The Mueller and Dweck experiment shows how a single comment to a child can have at least a temporary effect on their attitude and performance. It is unlikely that these children were still influenced by that one comment (“You’re smart!” or “You worked hard!”) a day later or even an hour later. But at least for a short time in a controlled setting, the children were apparently affected by what the adult researcher said to them. Why would this matter? If a child repeatedly and consistently hears one sort of encouragement or the other, the child might begin to internalize that way of thinking. Later, as an adolescent and then as an adult, the individual’s “mindset” can determine how that person approaches new opportunities to learn and to grow.

Before you go on, we’d like you to create a psychological theory. This may sound like a strange thing to do because theories are often presented to you in textbooks as being the final summary of some research. Sometimes that is true, but the primary use of theories in real scientific research is as a temporary and changeable summary of a researcher’s ideas.

Next, let’s read about a second study performed by Dweck’s research team nearly a decade later, although this one is described more briefly and with less detail. This second study was not an experiment because there were no manipulated variables. It was a longitudinal study, which means that the same participants (in this case, children) were tested repeatedly over a long period of time.