An earlier form of this essay was published in Gary K. Clabaugh & Edward G. Rozycki
Understanding Schools: the foundations of education 1990 New York. Harper & Rowe

A. Causal Fallacy in Teaching and Learning
©2000 NewFoundations


edited 11/27/18


Cognitive research demonstrates that people work best with and within a complex system if they have a "mental model" of the system - that is, an idea of all of its parts, what each does and how they work together, how changes in one part of the system cause changes in other parts. This mental model permits flexibility in responding to unexpected situations. ... One important function of schooling is to develop the knowledge and mental skills students will need to construct appropriate mental models of systems with which they will eventually work. ------Lauren B. Resnick (1)

How does what teachers do connect with what students learn? Bad teaching, so we are told in the mass media, has helped bring about the trade deficit. But why not blame someone or something else? Why should educators bear such responsibility? In this set of essays we will examine different models of learning and teaching and look at different conceptions of causation which might connect them. This will help clarify what it is reasonable to hold teachers accountable for. We will see that there are competing traditions in terms of which causation, learning and teaching are identified and assumed to be related. We will see that the relationships between teaching and learning, and the kind of causal connections believe to link them depend very much on the image of the school invoked.

Teachers face at least two problems. The first is where to get knowledge about learning and teaching. The second is whether or not they can put that knowledge to use. Consider the following scenario: it is discovered that reduction of class size to nine pupils per teacher produces substantial increase on all measures of student achievement. Can this knowledge be put to use? Who will rush to possibly triple the size of the school budget to take advantage of this discovery? We will see below that concerns about the nature of learning and teaching are not merely theoretical exercises, but bear heavily on who enjoys what benefits and who pays what costs in the socio-educational system.


In her 1987 Presidential Address to the American Educational Research Association, Lauren B. Resnick summarized the research on learning in and out of school. She described how very different learning is in school from what it is on the job.(2) School learning, Resnick pointed out, demands individual cognitive effort, generally unaided by tools. However, individuals in the work world are generally involved in cooperative thinking and use a variety of tools and artifacts to support their learning process. Reasoning in school is generally a matter of the manipulation of abstract symbols, whereas at work reasoning is contextualized and concrete. Learning in school tends to be of a very general nature. At work it is very specific to the job situation.

Thus the idea that school prepares students for the world of work does not seem supported by research. Does this mean that schooling is irrelevant? No. According to Resnick's research, schooling affects trouble-shooting ability. When the very specific, concrete situations of the work-a-day world become snarled, when the normalcy that supports job skills falls apart, then people with more schooling adapt more readily to the emergency situation. But how often do such emergencies arise? Should every student be prepared to deal with such irregularities? The differences between in-school and on-the-job learning seem to indicate that substantial reform in the school curriculum may be called for. Chart 1 below summarizes Resnick's findings.

from Resnick, 1987

Differences in


 Individual, Isolated
 Cooperative, Teamwork
 No tools.
"Pure mentation"
Tools for thinking.
Manipulation of media.


 Manipulation of symbols
 Contextualized reasoning

 Highly generalized
 Situation specific

 Unadaptive, stymied

Chart 1

Resnick's distinctions point to an important consideration: Learning may not be a uniform process. That is to say that learning English may not be the same sort of thing as learning math, or music. Indeed, the casual term "learning" may be little more than a slogan which is traditionally applied to a variety of processes or outcomes with little concern for their actual similarity. On the other hand, many theorists presume that "learning" indicates some sort of generic process so that different kinds of learning can all be understood as variations of the same thing. Hilgard's classic textbook in introductory psychology defines learning in this manner: learning is the process by which an activity originates or is changed when someone responds to a situation.(3) So important is it for some theorists to insist on a generic process, that they are willing to ignore contradictions. Hilgard defines "latent learning" as that which occurs even if the learner does not respond to a situation. We could just as easily say that nothing acquired "latently" is learning, since by Hilgard's own definition it did not involve a response to a situation.

Is this quibbling? What conceivable practical difference could a matter of definition make? We will see shortly. These issues are not mere abstractions. What is important to notice right off in considering the controversy about whether learning is multiplex or generic is that a decision on this issue is likely to affect the authority of different stakeholders in the curriculum controversy. If learning a language is a substantially different process than learning to shoot a basketball, then research in one area may not be relevant to another. People on the "front line" of schooling may well know things that university study alone cannot give them. We see then that what appears at one level to be a theoretical dispute, may at another have substantial consequences in terms of who benefits and who loses in the competition for school resources.


Philosopher Ludwig Wittgenstein asks us to consider two bushes both trimmed into the same shape; let's say, an elephant. We look and see the same two floppy ears and curled trunk on both of them. There, too, the curled tail and raised foreleg. The bushes have been expertly trimmed into the same configuration. But would we then expect to find the same pattern of branches, twigs and leaves, the same root structure in both bushes? Of course not! We recognize that even substantial surface regularities may have very different underlying structure. (See figure 2) Why then should we assume that learning to play the piano is the same thing as learning to do quadratic equations, or even that learning to play the piano is the same thing for different people? Why should we imagine that learners are alike? Why should we expect teaching to have the same effects even when motivation may be the same?


Figure 2

Do dissimilar appearances indicate dissimilarities? How do we come to judgment on these matters?(3B) Again, these may seem to be esoteric theoretical questions. But every teacher (every researcher) is faced with such "abstract" questions everyday on the job. All the students may have successfully completed a lesson in a workbook. Have they all learned the same? Not necessarily.

On the other hand, couldn't some skip that lesson or do something different and still learn the topic? It is very possible. Yet, for organizational reasons, we often plan the same lessons for all the students in a class. How are such plans connected to the results we are looking for? How is what we do connected to what we look for? And how is what we see connected to what we plan to do? In the next section we will look more closely at the notions of cause and effect in learning and teaching. It turns out they are not simple.

Causation in Learning and Teaching

It is perplexing, even annoying to teachers to hear would-be reformers tell them that their efforts at teaching affect the balance of international trade, the alcoholism rate among teenagers, or the prospects for American Democracy in the twenty-first century. Yet, when teachers ask in return, "How should I deal with children from broken homes, with alcoholic parents, with high needs for attention, or with low level reading skills?" the answers they receive are seldom explicit or pertinent. Many people seem to be able to discern all sorts of connections between the schools and social evils, yet far simpler connections between teacher treatments and student learning remain deepest mystery.


What connects with what? That is the general question of causation, posed in its most informal guise. More practically-minded people might put it, "We want some results, so how do we get them?" No fancy theories of causation will amount to much if they do not help us with everyday questions like these. But there are problems with finding causes. Not the least of these is that the word "cause" is ambiguous. It has at least the following meanings:

These rather informal, non-scientific senses contrast with the technical notion of cause generally used by scientific researchers:

A is the cause of B if A and B are distinct, highly correlative variables and A can be manipulated, e.g. "Turning the knob to the right makes the gas flame higher."(5)

It is a matter of dispute how this technical notion of cause relates to our traditional ones.

Are sarcasm and negligence variables? What about persons? It is unclear.(6)  The technical concept of causation is summative. That is, it requires that if all of certain conditions are met, i.e. the "cause" is present, certain results must occur, i.e. the "effect" must follow. If not, there is no causal relation. So if we manipulate A and the is no change in B, or if B changes independently of our manipulations of A, there is no causal relation between A and B.

The point of this apparent technicality will be seen below when we discuss hoped-for learning outcomes that are not summative. Controversies in education often rest on confusing different senses of cause. It is easy to find two items of interest and discover they covary, that is, that changes in one appear to relate to changes in the other. It is quite another thing to establish a causal relationship between them. For example, the following things have all increased over the last ten years:

Would anyone want to claim any item in the list has caused any other? What connection, if any, could we claim between them?

A major concern in dealing with educational problems is the extent to which stakeholders in a situation share the same perception of it. Those perceptions depend upon their sharing criteria and authorities in case of dissensus. Exactly the same problems exist for seeing a connection between items of interest. Understanding a causal relationship depends upon

People who believe in astrology see connections that others reject as "unreal." Teachers who believe in Freudian theory see connections that Behaviorists don't, and vice versa. Most Americans, having absorbed one lesson of our pluralistic society, readily recognize that different beliefs affect a person's notions of cause.

What is harder to see, however, is that our everyday language mixes concepts from different traditions so that they cannot be translated into the simple, uniform concepts required for precise measurement.(7) Many terms are taken from experimental contexts and used in schools applying traditional, rather than the experimental criteria which were their original context. So teachers talk about "reinforcing behavior" and mistakenly think they are doing the same thing with kids that Skinner did with the rats in his "Skinner boxes".


Educators often find themselves in a dilemma. The schools are looked to to make up for all kinds of social ills but teachers are seldom trained to look for causal linkages between what they might do and the benefits expected of their actions. Nor are they trained to be critical of those who claim to see a connection between changes they want in the schools and some imminent calamity. Because it is important that school people have some idea about expectations that might be based on ill-established causal claims, we will review quickly errors they might meet where a connection is claimed between school practice and some other kind of event. These errors are

We will look at each of these in turn.

Spurious Correlation

Two kinds of events may correlate yet be unrelated causally. The National Debt has recently risen along with SAT scores. Does this mere correlation establish a connection? Not at all. Very often mere change is accompanied by a increase in another variable. A teacher, for example, may find an increase in student motivation when new materials are handed out. This increase may have to do with the novelty, rather than the nature of the materials. Also, the special attention paid persons in experimental groups tends to bring about changes even though the special treatment itself has no effect. This is called the Hawthorne effect.(8)

Sometimes two items correlate because they have a common cause. Thus, reduction of pre-school activities may correlate with reduction of after-school activities because of a generally reduced school budget, not because pre-school and after-school activities are causally related. There is an extensive professional literature on "effective schools." Very often such articles give a list of characteristics that correlate with measures of school effectiveness. For example George and Oldaker (9) mention such a correlation between "learning climate", "student behavior" and "student achievement" What are we to make of this? First of all, we must know that the measures of the three variables, "learning climate","student behavior", and "student achievement" are independent. That is, one is not measured by the other. Otherwise, we are not talking about three distinct variables. Secondly, and most importantly, the correlations may not be causal. They may be the result of a common cause, for example, parental support. This means that attempts to make your school more effective by directly improving , "learning climate", "student behavior", and "student achievement" may not work.

Spurious correlation is troublesome at all levels of the schooling enterprise. For example, in an article by former Secretary of Education, William J. Bennett, called "Lessons From Great Schools" published in The Reader's Digest in November 1988.(10) He describes how a troubled Artesia High School in Lakewood, California was turned around:

  "...When Mara Clisby took over as principal in 1981, gangs and drugs plagued the school. Test scores were low, and a quarter of students were absent on any given day, many of them afraid to attend. Clisby and assistant principal Joe Quarles walked the neighborhoods, talking to parents, students and gang leaders. They laid down the law. No more drugs. No more gangs. Any students who were caught fighting would be suspended - and any caught in a gang fight would be expelled. The girls' rest room was a hangout, so Clisby moved her desk right into it, counseling and holding meetings there. Fear left the halls. Daily attendance soared to 98 percent. In 1981 a little over half of Artesia's graduates went on to college. By 1983, the number exceeded 80 percent, and remains at that level today."

If there is a lesson to be learned here for administrators, it is far from clear what that lesson is. Should they walk the neighborhoods? Talk with people? "Lay down the law?" Put their desks in the girls' room? What Mr. Bennett insinuates is that other principals have not tried such things. But what we have to know is how many have tried and how many have failed or succeeded. Then we might have some reasonable idea about what is effective.

Mr. Bennett commits the ancient logical fallacy of post hoc ergo propter hoc , "after this, therefore, because of this." Single examples of success, no matter how inspiring, do not prove the case. Mr. Bennett seems also to believe in the magic of command.  But what effect can we expect that a principal's laying down the law will have on the actions of drug dealers and gang leaders, when the police themselves cannot control them?

If we want to look for an explanation for the ease with which Artesia High School was "turned around" we should, perhaps, pay more attention to the fact that more than 50% of its graduates we going on to college despite these problems. This indicates a community substantially committed to education. This might be the real key to the principal's success.

At the risk of beating an obvious point to death, let's recall from our own experience that curious correlational superstition which equates success with control and failure with blame. There is a great deal of chance that determines the outcomes of as risky a venture as schooling. But it is a privilege of the successful to claim sole authorship of their successes. "If it turned out right, I did it!" On the other hand, woe to the person who is standing by the Xerox machine when it breaks down. Baleful glances convict where reason would not dare to accuse.

Experienced educators are well-aware of this fallacious thinking, even though they are often tempted to exploit its benefits. Consequently, there is no end of applicants for positions in "good" schools, while even substitute teachers cannot be found for "bad" ones. The easiest way to be a "good" teacher or a "good" administrator is to have "good" students or a "good" school.

Non-systemic Thinking

Perceptions of straightforward cause and effect are invariably the result of a stable underlying system. We turn the knob on the stove and the flame shoots up in direct relation to the angle of the turn. A direct, unilinear correlation: the paradigm of efficient causation. But only because the rest of the system is stable. If the gas coming into the house varied in its pressure, or if the pipes leaked, our nice causal relationship would disappear.

Most causes are multiple. We tend to focus on those that we can manipulate while maintaining the rest constant. Lecture is by far the most widely used method of instruction from seventh grade on. But how effective is the lecture method? It all depends. If you can control motivation and IQ -- often done by college admissions -- it seems to work as well as anything and is cheaper than most other techniques. (This is the real value of IQ; it identifies those who can survive educationally in crowded lecture halls.) Do better lecturers get more taught? That is not clear. Are students learning because of the lectures? That is not clear either.

Suppose now we transfer a competent lecturer to a situation where either motivation or IQ vary.  Should we be surprised if test grades fall? Successful teachers who are wise will recognize the contribution that both their students and their schools make to their success.

Developmental Misperceptions

Organizations are systems. But so are people, in many ways. Systems change and develop through time. So do the causal relations within them. Readiness is an educational concept that recognizes that a technique may be appropriate at one time in the development of a learner and inappropriate at another. Some researchers believe, for example, that reading problems are almost invariably a result of trying to push early literacy. Teachers used to working with kids within a certain age range often find their expertise reduced when they have to deal outside it. It is here that curriculum material organized along logical, disciplinary or organizational lines conflicts with what is pedagogically effective. (11)

Keeping the possibility of developmental difference in mind can help us through our failures. A technique may test out as ineffective, when in fact it is only inappropriate to the developmental level of the student. The notion of development leavens our judgment against rash decision.

Assumptions of Passivity

Some reforms are not worth the pain. Expecting to make substantial changes on a system assuming that the stakeholders will remain passive while these changes are effected, is no small mistake in judgment. Systems management models often used to try to explain schooling activities rest on substantial assumptions of passivity in the systems they describe. So do the great majority of reform proposals that receive airing in the press. We begin to understand the recurrent call for educational reform in the United States. It is based on false assumptions of passivity. What is less obvious is why such assumptions persist in the face of contrary evidence.

See Related article:
Moral Responsibility in the Education Industry




(1) Lauren B. Resnick "Learning in School and Out" Educational Researcher Dec.87

(2) Resnick, p.18

(3) Ernest R. Hilgard Introduction to Psychology 3rd Edition (New York: Harcourt, Brace and World, 1962) p.252

(3B) Problems which continue (2018) to beset researchers in Artificial Intelligence. See Artificial Intelligence Weirdness. Need categorizing relate to visual cues?

(4) Cf. H. L. A. Hart , "The Ascription of Responsibility and Rights" in Anthony Flew (ed.) Logic and Language (Garden City: Doubleday Anchor, 1965) pp.151-174

(5) The literature on the concept of cause is immense and its application to education problematic. See Edward. G. Rozycki, "The Functional Analysis of Behavior" Educational Theory 26, 3. Summer 1975 pp.278-302 Also, the classics: Donald T. Campbell and Julian C. Stanley, Experimental and Quasi-experimental Designs for Research (Chicago, Rand-McNally, 1963) Mario Bunge, Causality (Cleveland: World Pub., 1963) and Arthur L. Stinchcombe, Constructing Social Theories (New York: Harcourt, Brace and World, 1968)

(6) Edward G. Rozycki Human Behavior: Measurement and Cause. Can there be a science of education? Ed.D. dissertation, Temple University (Ann Arbor: Xerox University Microfilms, 1974) Chapter IV, "The Causes of Behavior"

(7) Edward G. Rozycki, "Rewards, Reinforcers and Voluntary Behavior" Ethics. Vol 84, October 1973. pp.38 - 47.

(8)  See literature on the "Hawthorne effect", e.g. Perrow, pp.90-92

(9) Paul S. George and Lynn I. Oldaker, "A National Survey of Middle School Effectiveness" Educational Leadership. December 1985/January 1986 pp.79-85. See also, M. Donald Thomas, "What is an effective school?" The Effective School Report. Jan/Feb 1989 Vol.17, No.1 (New York: Kelwyn Press, 1989) p.5 for criteria which conflict along the dimensions divisible-indivisible and absolute-positional.

(10) William J. Bennett "Lessons From Great Schools" The Reader's Digest November 1988. pp.121-125.

(11) For more on the relation of pedagogical to logical, disciplinary and organizational approaches to curriculum, see Clabaugh & Rozycki (1999)  "Curricular Structures".