Six Ways to Use Examples And Nonexamples To Teach Concepts

Your brain likes to generalize. It forms generalizations out of individual examples, similar to fitting together the pieces of a puzzle. This is more or less how we form concepts—by grouping together things, ideas and events with shared characteristics so to classify them. It’s an efficient way to give meaning to your world.

For example, when you first saw a comic strip you most likely didn’t know the concept of “comics.” But over time, you learned that certain styles of line drawings formatted in a sequence were referred to as comics. Once you formed this concept, it became easier to classify other examples of line drawings as comics or non-comics.

At a young age, we learn to form generalizations, such as the concept for “comics.”

Examples or instances seem to be crucial for helping a learner form accurate concepts. Otherwise, according to educational researchers, the learner can overgeneralize, under-generalize or form misconceptions. Here are six ways to help learners acquire concepts. In overgeneralization, a learner applies a rule to a concept or classifies an example to a concept when it does not apply. Their concept is too wide. In under-generalizing, a learner’s understanding of a concept is too limited. They may not think that an example is part of the concept when it is. Misconceptions involve a failure to comprehend a concept accurately.

Rule #1: Use examples in which the irrelevant attributes vary widely.

As learning designers, we can support accurate concept formation by carefully selecting examples. Not just any examples will do. The instances we present should vary widely, especially with respect to irrelevant attributes, so learners form the right idea of what defines the concept and what does not.

Consider a course on sexual harassment. If every example shows a man harassing a woman, learners may incorrectly conclude that harassment cannot occur between people of the same gender or in the opposite direction. This kind of faulty generalization arises when examples are too similar in nonessential characteristics.

By varying the irrelevant features of your examples, such as the gender or role of the individuals involved, the setting, or the type of communication, learners build a more flexible and accurate understanding of the concept. When examples do not vary on these noncritical dimensions, learners tend to form concepts that are too narrow (Merrill & Tennyson, 1977).

Rule #2: Progress from simple to difficult examples.

It’s also important to start with simple, clear examples at first, followed by more complex instances. Research shows that if learners see only the easiest examples of a concept, they may fail to recognize more difficult or borderline instances later. This leads to under-generalization (Merrill & Tennyson, 1978).

Imagine a radiologist learning to identify bone fractures. If the training only includes obvious fractures—where bone fragments are completely separated—the learner may not recognize subtler cases. When presented with a fracture where the fragments remain partially connected, the radiologist could misclassify it. Beginning with simple examples and gradually adding complexity helps learners develop an accurate mental model.

Rule #3: Present instances of a concept in sequence or allow multiple instances to be viewed at the same time.

Examples should appear close enough together in time that they remain active in working memory. This supports the learner’s ability to see the commonalities across examples and form the intended generalization. When examples are separated by long time gaps, learners may not make the connection. Present examples in a tight sequence, or keep several visible simultaneously (Gagné, 1985).

Rule#4: Use matched examples and nonexamples for concepts with related attributes.

Your brain generalizes, but it also likes to discriminate. Discrimination helps learners narrow the range of instances that belong to a concept. To support accurate concept formation and to avoid overgeneralization, promote discrimination by presenting nonexamples alongside examples. A nonexample is an instance that is not part of the concept, and it should differ from its paired example in only one key attribute.

Some research points to the fact that concepts with clearly distinguishable attributes (like texture versus color) are better taught with examples only. The learner only needs to generalize in order to acquire the concept. In these cases, learners mainly need to generalize. But concepts that share overlapping attributes, such as the statistical concepts of mean, median, and mode, require discrimination skills and benefit from matched examples and nonexamples. For discrimination to occur, the pair should be available in working memory at the same time.

Back to the comics theme mentioned earlier, it’s difficult to fully define the concept of comics. A paired example and nonexample can help because recognizing the difference requires discrimination. The pair below demonstrates that the intent of the artist is one attribute for defining comics. The example on the left was intended to be read as a comic book. Roy Lichtenstein’s comic-based painting on the right was intended as an artwork that provided commentary on the mass media.

Present matched examples and nonexamples when concepts could be confused with each other.

Rule #5: Provide opportunities for learners to generate their own examples of a concept.

If you’re working in a setting that allows for interaction, ask learners to generate their own examples of a concept. Promoting reflection and a response helps to reinforce well-formed concepts and correct misunderstandings. It may also clarify fuzzy areas, as all concepts are not neatly defined. Learner-generated examples can reveal where additional guidance is needed.

Returning to the sexual harassment course mentioned earlier, a learner might generate an example that asking a co-worker for a date constitutes harassment. Others in the discussion might point out that the behavior becomes harassment only when the requests continue after repeated refusals. This kind of peer feedback helps correct misconceptions and strengthens a person’s understanding of the concept’s defining attributes.

Rule #6: Expose learners to a wide range of examples and nonexamples and allow them to discover the concept.

Humans discover concepts independently throughout their lives, from childhood on. Our pattern-seeking brains naturally look for similarities and differences across experiences. It’s not surprising, then, that one effective way to learn a concept is to discover it by interacting with a varied set of examples and nonexamples.

Using this inquiry strategy in eLearning is more challenging than in instructor-led settings, but it can be done. Present learners with a diverse collection of examples and nonexamples and prompt them to identify the characteristics that define the concept versus those that do not. Through exploration and guided discovery, supported by interactive activities, learners begin to infer the underlying structure of the concept.

In self-paced eLearning, use built-in questions, hints, and context-sensitive feedback to guide learners toward an accurate understanding and to correct misconceptions immediately. This combination of exploration and timely digital feedback helps learners infer the concept while keeping the experience active and engaging.

In a synchronous setting, ask learners to articulate their ideas. use targeted questions and context-sensitive feedback to refine their understanding and correct misconceptions. Research seems to point to the fact that discovery learning is most effective when learners already have a developed knowledge base.

References:

1. Gagne, Ellen. The Cognitive Psychology of School Learning. Brown Little, 1985.
2. Merrill, M. D., & Tennyson, R. D. (1977). Concept teaching.An instructional design guide. Educational Technology.
3. Merrill, M. D., & Tennyson,R . D. (1978). Concept classification and classification errors as a function of relationships between examples and non-examples. Improving Human Performance, 7, 351-364.
4. Tennyson, R.D. and Cocchiarella, M.J. (1986). An Empirically Based Instructional Design Theory for Teaching Concepts. Review of Educational Research, Vol. 56, No. 1.

Don’t miss one issue. Get my monthly instructional design newsletter with articles, resources, and freebies plus the Writing for Instructional Design eBook.