"Providing Feedback in Computer-based Instruction: What the Research Tells Us"

Prepared by Beryl Jean Maxon and Roger Bruning

Previous Thought in Research
Feedback was believed to either weaken or strengthen the learner's responses.
It enabled educators to think more positively about feedback but was a narrow view of what feedback could ultimately contribute.

Types of Feedback
There are two basic types of feedback.
Verification labels an answer as correct or incorrect.
Elaboration gives more information to lead the learner to a correct answer.

• Informational elaboration gives information that the learner can use to reach the correct answer.
• Topic-specific elaboration gives more detailed information about the topic and leads the learner through the correct answer. It does not address incorrect responses.
• Response-specific elaboration gives information on both correct and incorrect responses. The learner finds out why the incorrect answer was wrong.

The current belief is that feedback needs both verification and elaboration.

Types of Feedback (used individually and in combination)

• No-feedback: no verification and no elaboration 
• Knowledge-of-response feedback: verification and no elaboration 
• Answer-until-correct feedback: verification and no elaboration (The learner answers it again and again until he or she chooses the correct answer.) 
• Knowledge-of-correct-response feedback: verification and no elaboration except telling what the right answer is 
• Topic-contingent feedback: verification and elaboration containing information with the correct answer 
• Response-contingent feedback: verification and elaboration based on the learner's response 
• Bug-related feedback: verification and elaboration on the error the student may have made (It allows the learner to discover where they went wrong and determine the correct answer.) 
• Attribute-isolation feedback: verification and elaboration on the general concept

Factors That May Influence the Feedback's Effectiveness
Elaboration. some studies show that elaboration has no effect, but more studies show that elaboration helps increase learning. There is a "trend for increased learning in response to more elaborate feedback," but there are apparently other variables at work.
Student achievement levels. immediate feedback better for low-achieving students while higher ability students may receive more benefit from feedback that allows them to process information on their own
Depth of knowledge. There weren't exact answers in this area (in fact, results conflicted) but learners with more understanding could get more out of answer-until-correct feedback than learners with less understanding. Feedback effects seem to be weaker for higher-order learning, but learners at a lower level could benefit more from delayed and knowledge-of-correct-response feedback than from answer-until-correct or no-feedback.
Attitude toward feedback. Students seem to want "more elaborative, immediate feedback" (11). So giving learners feedback that they believe helps is probably a good idea even though the attitude may not actually impact learning outcomes.
Learner control. It may be beneficial to allow high ability learners control over the type of feedback they receive. However, it may not be as good of a policy to allow low ability learners to choose their feedback type, because low ability learners may just want the correct answer instead of the feedback that will provide the highest amount of learning.
Response certitude. A correct answer may be from a guess or from total understanding. An incorrect answer may be from a simple mistake or from a total lack of understanding. The impact of the feedback would be different depending on the reason.

1. High certitude and correct answers. the learner does not spend much time on feedback.
2. High certitude and incorrect answers. the learner wants to find the reason for the error. (This situation is possibly when feedback is the most helpful.)
3. Low certitude and correct/incorrect answers. the learner may not receive as much benefit from the feedback if the understanding (and thus certitude) is low. However, a study by Mory (1994) found that learners with low certitude spent more time studying the feedback.

There was not a significant difference in learning when the feedback was adjusted for certitude and correctness.Timing. There is a debate between whether immediate feedback is better due to immediate correction before the error is ingrained or whether delayed feedback is better so that correct information can be studied without interference. It seems immediate feedback is better for applied studies, list learning, decision-making, novel information tasks, and lower level knowledge based tasks. It helps with learning the content. Delayed feedback is better for learning test content, abstract concepts, and application/comprehension skills. It helps with remembering content for the long-term. Both types could be used. For example, the student can know immediately whether they were correct (verification) but learn later about the information behind the correct answer after having some time to think for themselves (elaboration).

Conclusions and Recommendations
"The challenge therefore is to identify the type of feedback that is most effective in specific educational settings" (14).

Student Achievement	Lower Level		Higher Level
Task Level	Lower level task	Higher level task	Lower level task		Higher level task
Timing of Feedback	Immediate feedback		Immediate feedback		Delayed feedback
Prior Knowledge	Low prior knowledge	High prior knowledge	Low prior knowledge	High prior knowledge	Low prior knowledge	High prior knowledge
Type of Feedback/Level of Elaboration	Knowledge-of-correct-response feedback with response-contingent feedback	Knowledge-of-correct-response feedback with topic-contingent feedback	Knowledge-of-correct-response feedback with RC feedback	Knowledge-of-response feedback with TC feedback	Knowledge-of-response feedback with delayed knowledge-of-correct-response feedback + RC feedback	Answer-until-correct feedback and delayed TC feedback

Questions

1. Is it fair to split up lower order learners and higher order learners? Is it a reality or a label? How can we create higher order learners from low level learners? Can we instill it in them from the very beginning?
2. Is there a way to overcome the problem in student's scores reflecting guesses and simple mistakes? After all, a learner could memorize a chapter and spit it out but still not understand it. Is there a new type of "question" we haven't thought of?
3. Is our view of feedback still narrow? In what ways could it be broadened?