Does Too Much Instruction Interfere with Creativity?
Recent articles like the one from Alison Gopniks entitiled, Why Preschool Shouldn't Be Like School, suggest that children's creativity is stifled when they are directly instructed from teachers. While, I agree that it may be partially true that using direct teaching methods with children may result in less novel responses, I do not agree that this type of research provides evidence that teachers should not directly instruct children in preschool. In order to discuss this issue it is important to be familiar with 3 types of instructional methods that have been heavily researched in the science of learning.
The first type of learning is referred to as trial-and-error learning or contingency-shaped teaching. With this type of learning the child learns on his own via the consequences of his behavior. The type of consequence present after the behavior occurs will shape or change the future behavior. This type of learning typically occurs in the natural learning environment. For example, a child will learn not to reach out for the roses if the consequence is being pricked by a thorn. Or on a more positive note, a child will continue to twirl and dance when the consequence is an audience clapping and cheering. Behaviors that result in unpleasant consequences will be less likely to occur again and those that result in pleasant consequences will be more likely to occur again. This type of learning is referred to as trial-and-error because the child is not instructed to the right answer or given any hints or prompts, he must discover the right answer on his own.
The second type of learning is referred to as errorless learning. Errorless learning is similar to trial-and-error learning in the sense that the consequences shape or change the future behavior, but with errorless a cue or prompt is provided to help the student choose the correct answer, thereby, reducing or eliminating errors during the learning process. Contrary to popular belief, research in errorless learning has shown that it is not necessary to experience errors in order to learn. In fact, errorless learning has similar learning results as trial-and-error, with the added benefit of having less frustration with the learner. As long as the teacher, quickly fades out the type of cue or prompt she is providing, errorless learning is a great teaching strategy with many benefits to the learner.
The third type of learning is referred to as rule-governed behavior. This is the category that I think most people are referring to when they are talking about directly instructing the child. Rule-governed behavior is behavior that is taught by instruction. Instead of directly experiencing the consequences of the behavior, an instructor gives a rule to follow, such as, "Don't touch the thorns, they will prick your finger." One of the perks of being a part of a species that has language is that we can teach our young via rules so that they do not have to directly experience everything in order to learn.
Now this brings us back to our original question. Does too much instruction interfere with creativity? If one is to measure creativity by the number of novel responses made to a toy after being taught one response using one of the three methods of teaching (see example in Gopniks article), then the child will make the most novel responses when taught via trial-and-error, a slightly smaller amount of novel responses when taught via errorless, and is likely to make very few, if any novel responses when told what to do. The problem with this, is I don't think we are actually measuring creativity with this example. We are measuring how many errors the child makes after being taught a correct response. In the example, Alison Gopniks referrs to in her article, they demonstrated (i.e., used a modeling prompt) one of many ways to interact with a toy with one group of children, then instructed one of many ways to play with a toy with another group of children, then observed the children playing with the toy and measured how many ways the child played with the toy other than what was demonstrated or instructed. Essentially, this is comparing errorless, with the teacher demonstrating the correct response and rule-governed instruction, with the teacher instructing the correct response. They measured novel responses, which could also be looked at as errors, which in most learning situations novel responses would be errors because they are responses outside of the response modeled by the teacher or the instructions of the teacher.
If one wants to compare instructions versus trial-and-error or errorless to measure creativity the initial teaching would have to teach all desired responses either by reinforcing correct responses during trial-and-error or errorless and compare that to directly instructing the child to all the responses, then one could present the child with a new toy and measure the number of novel responses to a novel item to measure creativity. To my knowledge this hasn't been done, so I don't know what the results would be, but based on the principles of learning, I would guess that there wouldn't be much of a difference between the three groups in the number of novel responses observed to a novel stimulus. Where a difference would be found, is the length of training (trial-and-error would take more training trials to teach the skill) and if the child were presented with changes to the consequences. For example, if a part of the toy broke and no longer produced the desired effect, the children who were taught via trial-and-error and errorless will adjust their behavior (no longer play with that part of the toy) faster than the children who were taught via instruction. The children in the instruction group would be more likely to continue to attempt to make that part of the toy work longer. This characteristic of rule-governed behavior versus contingency-shaped behavior could be argued to be a case for creativity versus rigidity. The children who were taught via instructions are more likely to continue engaging in the same response in the same way whereas children who were taught using contingency-shaped procedures (trial-and-error and errorless) will be more likely to try new things.
Both trial-and-error learning and errorless learning result in more flexible responding. What that means is that if there are changes to the environment or to the consequences, behaviors taught via trial-and-error or errorless learning will adapt and change more readily to suit the environment. Rule-governed behavior, on the other hand, is more persistent and rigid in the face of environmental fluctuations or changes. If this is used as a measure of creativity, then, yes, instruction based learning will result in less variability and flexibility of responses. If your goal of teaching is to have variable and flexible responding, then a contingency-shaped teaching procedure (trial-and-error or errorless) should be used to teach the skill. If your goal is to have rigid and accurate responding then instruction based teaching should be used to teach the skill.
But, this doesn't mean we should throw instruction based teaching out just yet to teach creativity. It is the case that creativity can be directly taught via instruction. For example, you can tell a child he has to respond differently than his last response and as a result he will engage in novel responses (i.e., creativity). He is likely to do so in a very systematic way, but this may be a useful teaching strategy for children who seem to struggle with creativity. Some teaching programs, such as Fit Learning, have found success directly teaching verbal flexibility which is a form of creativity (e.g., saying increasingly abstract things about an item). Curriculum for children with autism also focuses on directly teaching creativity using both instruction based teaching and contingency-based teaching methods with success.
The decision to use instruction based teaching or play-based teaching methods (contingency-shaped teaching), really depends on the skill you are trying to teach, the child's learning abilities, and what works best for the learner in the situation. A strong teacher should be familiar with learning literature and should know the characteristics associated with each teaching strategy so that he can modify the curriculum for each learner as needed.
The first type of learning is referred to as trial-and-error learning or contingency-shaped teaching. With this type of learning the child learns on his own via the consequences of his behavior. The type of consequence present after the behavior occurs will shape or change the future behavior. This type of learning typically occurs in the natural learning environment. For example, a child will learn not to reach out for the roses if the consequence is being pricked by a thorn. Or on a more positive note, a child will continue to twirl and dance when the consequence is an audience clapping and cheering. Behaviors that result in unpleasant consequences will be less likely to occur again and those that result in pleasant consequences will be more likely to occur again. This type of learning is referred to as trial-and-error because the child is not instructed to the right answer or given any hints or prompts, he must discover the right answer on his own.
The second type of learning is referred to as errorless learning. Errorless learning is similar to trial-and-error learning in the sense that the consequences shape or change the future behavior, but with errorless a cue or prompt is provided to help the student choose the correct answer, thereby, reducing or eliminating errors during the learning process. Contrary to popular belief, research in errorless learning has shown that it is not necessary to experience errors in order to learn. In fact, errorless learning has similar learning results as trial-and-error, with the added benefit of having less frustration with the learner. As long as the teacher, quickly fades out the type of cue or prompt she is providing, errorless learning is a great teaching strategy with many benefits to the learner.
The third type of learning is referred to as rule-governed behavior. This is the category that I think most people are referring to when they are talking about directly instructing the child. Rule-governed behavior is behavior that is taught by instruction. Instead of directly experiencing the consequences of the behavior, an instructor gives a rule to follow, such as, "Don't touch the thorns, they will prick your finger." One of the perks of being a part of a species that has language is that we can teach our young via rules so that they do not have to directly experience everything in order to learn.
Now this brings us back to our original question. Does too much instruction interfere with creativity? If one is to measure creativity by the number of novel responses made to a toy after being taught one response using one of the three methods of teaching (see example in Gopniks article), then the child will make the most novel responses when taught via trial-and-error, a slightly smaller amount of novel responses when taught via errorless, and is likely to make very few, if any novel responses when told what to do. The problem with this, is I don't think we are actually measuring creativity with this example. We are measuring how many errors the child makes after being taught a correct response. In the example, Alison Gopniks referrs to in her article, they demonstrated (i.e., used a modeling prompt) one of many ways to interact with a toy with one group of children, then instructed one of many ways to play with a toy with another group of children, then observed the children playing with the toy and measured how many ways the child played with the toy other than what was demonstrated or instructed. Essentially, this is comparing errorless, with the teacher demonstrating the correct response and rule-governed instruction, with the teacher instructing the correct response. They measured novel responses, which could also be looked at as errors, which in most learning situations novel responses would be errors because they are responses outside of the response modeled by the teacher or the instructions of the teacher.
If one wants to compare instructions versus trial-and-error or errorless to measure creativity the initial teaching would have to teach all desired responses either by reinforcing correct responses during trial-and-error or errorless and compare that to directly instructing the child to all the responses, then one could present the child with a new toy and measure the number of novel responses to a novel item to measure creativity. To my knowledge this hasn't been done, so I don't know what the results would be, but based on the principles of learning, I would guess that there wouldn't be much of a difference between the three groups in the number of novel responses observed to a novel stimulus. Where a difference would be found, is the length of training (trial-and-error would take more training trials to teach the skill) and if the child were presented with changes to the consequences. For example, if a part of the toy broke and no longer produced the desired effect, the children who were taught via trial-and-error and errorless will adjust their behavior (no longer play with that part of the toy) faster than the children who were taught via instruction. The children in the instruction group would be more likely to continue to attempt to make that part of the toy work longer. This characteristic of rule-governed behavior versus contingency-shaped behavior could be argued to be a case for creativity versus rigidity. The children who were taught via instructions are more likely to continue engaging in the same response in the same way whereas children who were taught using contingency-shaped procedures (trial-and-error and errorless) will be more likely to try new things.
Both trial-and-error learning and errorless learning result in more flexible responding. What that means is that if there are changes to the environment or to the consequences, behaviors taught via trial-and-error or errorless learning will adapt and change more readily to suit the environment. Rule-governed behavior, on the other hand, is more persistent and rigid in the face of environmental fluctuations or changes. If this is used as a measure of creativity, then, yes, instruction based learning will result in less variability and flexibility of responses. If your goal of teaching is to have variable and flexible responding, then a contingency-shaped teaching procedure (trial-and-error or errorless) should be used to teach the skill. If your goal is to have rigid and accurate responding then instruction based teaching should be used to teach the skill.
But, this doesn't mean we should throw instruction based teaching out just yet to teach creativity. It is the case that creativity can be directly taught via instruction. For example, you can tell a child he has to respond differently than his last response and as a result he will engage in novel responses (i.e., creativity). He is likely to do so in a very systematic way, but this may be a useful teaching strategy for children who seem to struggle with creativity. Some teaching programs, such as Fit Learning, have found success directly teaching verbal flexibility which is a form of creativity (e.g., saying increasingly abstract things about an item). Curriculum for children with autism also focuses on directly teaching creativity using both instruction based teaching and contingency-based teaching methods with success.
The decision to use instruction based teaching or play-based teaching methods (contingency-shaped teaching), really depends on the skill you are trying to teach, the child's learning abilities, and what works best for the learner in the situation. A strong teacher should be familiar with learning literature and should know the characteristics associated with each teaching strategy so that he can modify the curriculum for each learner as needed.
Posted by Dr. Eilers at 9/7/2011 4:56 PM 
Categories: Education
Tags: teaching strategies creativity education
Categories: Education
Tags: teaching strategies creativity education
Great post. I think one of the stigmas against ABA is the idea that it stifles creativity. I believe a study by Goetz and Baer in the Journal of Applied Behavior Analysis in the 60s or 70s discussed this. It measured both variability in words and new words said per session.
Reply to this