I gave a talk at the Seneca Virtual Science Conference on applying the principles of faultless communication (as described in Engelmann & Carnine’s Theory of Instruction) to Science. With it being a 5 minute talk and after a few requests, here’s a blog elaborating on the key ideas.

## Inductive reasoning

Engelmann and Carnine state that we learn from our physical surroundings by inductive reasoning. That is reasoning based on a series of examples and non-examples that we encounter. A good example of this is how we learn colours. Initially, when our parents point out something that is red, we assume red means just the one example that was used (e.g. a fire engine). As we are exposed to more and more examples (and non-examples such as non-red fire engines) we build up a picture of precisely what red is and is not.

The danger of this method of course, is that we may not induce the correct interpretation from these examples and non-examples. If we are only ever shown fire engines, cars and lorries that are labelled as red, we may incorrectly induce that the colour red only applies to vehicles. Engelmann and Carnine refer to this as stipulation.

## Faultless communication

And so we come to the principles of faultless communication. By adhering to these 5 principles in our teaching we will avoid stipulation and prevent students from inducing the wrong interpretation or embedding misconceptions.

The example I will use is a correlated feature concept. This is where two items are linked (correlated). This is most commonly seen in science as cause and effect – “if __________ then __________”. For example:

**If**a cell has a lower water potential than its surroundings**then**water will enter the cell by osmosis.**If**a gene locus has more than one possible allele**then**it is polymorphic.**If**the air humidity increases**then**the rate of transpiration decreases.

All of these examples are abstract or invisible and many students will have difficulty grasping these alongside an explanation of why the correlation occurs. Engelmann proposes that we first make the correlation fluent before then explaining it or providing experimental observations to prove it.

This would be done with a series of examples and non-examples (which I will refer to as items to avoid confusion). I’ll use the same sequence from my talk and highlight how each of the principles have been applied. The correlation I am trying to induce with this sequence is:

- If the air humidity increases then the rate of transpiration decreases.

### The Wording principle

Item 1 is a non-example of this correlation. The air humidity does not change and so neither does the rate of transpiration. Item 2 is an example of the correlation. The red box highlights that the wording used to introduce both non-example 1 and example 2 is identical (the difference being the response of “Yes” or “No” that denotes it as an example / non-example). Students are forced to attend to the difference in the examples rather than any difference in how the examples are worded.

### The Setup principle

The items used share the greatest number of irrelevant features as possible. The only difference we want students to attend to is a change in the air humidity. The items used are all % measurements of air humidity and are multiples of 10. The only change is a single number meaning that there is almost guaranteed success that students can attend to the change in air humidity.

### The Difference principle

As previously mentioned, item 1 is a non-example. But, importantly, it is the smallest possible non-example. There is no change in air humidity and so students will infer that any non-increase in air humidity will not cause a decrease in the rate of transpiration. We call this a boundary non-example.

Item 2 is the smallest possible example. It is an increase in air humidity of 10% which is the smallest possible increase with the scale I have chosen using the setup principle. Students will infer that any increase, no matter how small, will result in a decrease in the rate of transpiration. This is known as a boundary example.

With these two being the first items shown, students are immediately exposed to the boundaries of the correlation. They can now extrapolate that any non-increase (not just a decrease) is a non-example and any increase is an example.

### The Sameness principle

Item 6 is one of the largest items that can appear. It is 60% increase in air humidity from the previous item. This creates an outer limit, within which students can interpolate that all increases within the boundary set in item 2 and this limit are examples of the correlation.

### The Testing principle

Finally, students need to be tested so that the teacher can diagnose whether the instructional sequence has been successful or not. Items 3 to 6 are all posed to the students who should respond using the language demonstrated in item 1 and 2. This may seem simple but this is, to a certain extent, the point. The responses from the student will allow the teacher to know **for sure** whether they understand the correlation or not. The sequence of these test items should be random and not produce a pattern (yes, no, yes, no etc.) that students could game and produce false positives.