Classic Learning Research

Ultimately, learning entails a neurobiological response to a stimulus of some sort. This unobserved neurological response is translated to an observable behavioral response, which can encompass anything from avoiding the stimulus in the future to correctly answering a test item. Classic learning research (as well as educational research in general) primarily concerns itself with changes in such behavioral responses ( learning ) following the presentation of visual or oral stimuli ( instruction ). Thus, if students are able to correctly answer test questions following instruction that they couldn’t answer correctly beforehand, then we infer that learning has occurred.

Just as all learning basically involves some type of observable/ measurable behavioral response, instruction also always boils down to a stimulus that is capable of eliciting such a response. From this perspective, then, instruction can take the form of (but is not limited to) such diverse stimuli as:

• Being lectured to in a classroom setting
• Completing computerized/online instructional modules
• Being presented a word, phrase, or nonsense syllable and told to memorize it
• Completing homework
• Engaging in self-study
• Reading
• Being read to
• Watching television
• Surfing the internet
• Listening to others (whether in class or at the dinner table)
• Being the beneficiaries of direct parental teaching
• Being corrected by parents
• Observing and subsequently modeling parental or peer group behaviors
• Observing the environment
• Visiting institutions with instructional agendas such as a churches, museums, and science centers

To control as many factors as possible in their research and to avoid teaching something that their subjects had already learned, classic learning studies often employed the visual presentation of nonsense syllables via a technique called paired-associate learning trials . Experimental subjects (typically, college undergraduates) were taught, via repeated presentations — often involving a slide projector or its equivalent — to “pair” these syllables (or sometimes conceptually unrelated words) until this arbitrary association was successfully “learned.” To avoid as much error as possible in inferring that learning had occurred (and to measure it as precisely as humanly possible), testing involved exactly the same processes that were used in instruction (i.e., the syllables, words, or whatever, presented via the same medium in which they were learned).

As obsolete as current classroom instruction is, present-day teaching isn’t quite this rote. Still, unlike classroom research, these experiments employed a form of instruction and a method of measuring learning that could be controlled and repeated quite consistently. This meant that scientists could have a great deal of confi dence in any learning principles they unearthed. Whether these principles would apply to all types of learning, no one knew for sure, but the best guess was (and is) that the same neurobiological processes are associated with all types of learning resulting from all types of instruction, rote or creative, interesting or dull.

So, at the risk of oversimplifi cation, three facets of learning were inferred by these studies, based on how many trials (or how quickly) students mastered the paired-associate tasks for which they received “instruction.” These learning facets or parameters were:

• Original learning, which is identical to what we mean when we refer to school learning;
• Retention, which refers to how long what was learned is remembered — or to the circumstances under which forgetting occurs; and
• Transfer of learning, which in classic learning theory refers to the fact that previous learning can sometimes facilitate (and sometimes even impede) subsequent learning.

And, if you think about it, these three behaviors pretty much reflect what we expect students to take from the schooling process: learning what is taught (otherwise attending school is a total waste of time), remembering what is taught (because if we don’t remember what we’ve learned, we might as well have not learned it in the fi rst place), and being able to apply what is learned to new situations (because supplying correct responses to test items would be worthless if we can’t assume that this will ultimately be related to other types of innovative, creative, or compliant behaviors of societal importance).

In a nutshell, then, the principles emanating from this type of research that were most relevant to classroom instruction and student learning were:

1. The more times the paired-associate tasks were repeated (that is, the more instructional time supplied), the more learning occurred. This was the strongest and most consistent relationship that this line of investigation ever uncovered: more relevant time on task (or more presentations of the stimuli) results in more learning. It was so pervasive, in fact, that some researchers embraced a “total-time hypothesis,” which basically postulated that, within reasonable limits, the same amount will be learned in a given amount of time regardless of the number of trials presented within that time period.
2. Some forgetting almost always occurs, but the more time on task (or the more presentations of the stimuli), the longer the association (or learning) was retained (remembered). Retention can also be improved by (a) increasing the meaningfulness (or relevance) of the content and/or (b) continuing to present the stimuli even after they are learned (which was called over-learning ). Of course, this still reduces to time on task (or increased instructional time) since the presentation of a stimulus is a form of instruction.
3. Transfer of learning (one form of which was called “learning to learn”) proved to be a more tenuous affair, but it does occur as a function of instruction under certain conditions. For example, transfer was facilitated by over-learning, and it occurred most reliably when the training conditions were most similar to the ultimate testing conditions (which in schooling terms is reflected by practices such as teaching to the test or teaching test-taking skills) and when the original learning task possessed certain components in common with the transfer task (such as teaching a child the sound representing a certain vowel to facilitate the learning of a word containing that vowel). However, we still haven’t learned enough about this concept to stretch it to what we mean by such attributes as creativity (or innovativeness), and this remains a major gap in our understanding of the instructional–learning process. Suffice it to say that the occurrence of learning is a prerequisite for both retention (and transferring that learned knowledge to novel applications), but learning is no guarantor of either.

Now, admittedly, this brief overview does not do justice to classic learning research. Other variables were involved 4 but, generally speaking, most of the work in classical learning research, as in educational research in general, never transcended what educational researchers in my day called the “grandmother principle,” which can be summed up in the following succinct generalization:

You never discover anything in educational research that your grandmother didn’t already know .

Still, our grandmothers weren’t always right about everything, so it doesn’t hurt to subject some of their opinions to scientifi c tests. Thus, in summary, far and away the most important finding emanating from this classic research (as well as from learning research that involved rats navigating mazes) was that the strongest determinant of laboratory learning is the amount of instruction delivered. More instruction, more learning; more time spent studying, more learning; more time on task, more learning; the more time an author spends repeating something, the more likely the reader is to learn it — to remember it — and to apply it.
Read More : Classic Learning Research