v3.02
Material used solely for educational purposes in ESLE 301, Theory and Methods of Research, at the University of Connecticut.
chapter 14

Theories, Maps, and Models

from: The Science Game by Neil Agnew, 1969

A theory, or a map, or a model is an abstraction, and abstractions involve representing the part of the world under examination by a model of similar but simpler structure. The important point is that theories are simpler than the data they are designed to represent. Theories are built (1) by squeezing some parts of experience together-all blacks, or all smokers, or all Southern Democrats-and (2) by ignoring or omitting some information, such as the differences that exist between blacks, or between different smokers, or between different Southern Democrats.

WHY BOTHER WITH THEORIES? No Other Choice

We propose that the most important reason for bothering with theories is that we have no alternative. We would have an alternative if we were like some wondrous computer that (1) saw, heard, and felt everything, (2) had a massive and unlimited memory file so we could store each bit of information separately and permanently, and (3) could draw at will from our memory file each bit of information and experience for examination. Obviously, we are not like this. We listen, look, and feel selectively. We also forget, condense, and distort the information we do have. We tie bits of information together in our heads that are not necessarily tied together in the world around us. We are, in fact, information screening, condensing, and relating machines. The end results of this process are theories or mental maps of what goes together and what leads to what. If it were otherwise, we would go crazy from the great avalanche of detailed hits of information and experience to which we are exposed both from the outside and from the inside. You will recall that when the psychiatrist is faced with too much information or too many alternative suspects he develops biases or theories about what to look at and about what leads to what. Some psychiatrists focus on biochemical information and some on early training. The important point is that by focusing on certain types of information, by developing biases, and through forgetting, we reduce the avalanche of information and experience to manageable size. Like the psychiatrist, the scientist and the citizen increasingly face information and opinion overload. Like the psychiatrist, we too must develop simplified pictures, stories, models, or theories of what goes together and what leads to what. Through experience, training, and bias we group together certain people, objects, or events, at the same time ignoring many differences. We may group all blacks together. Even if we do it on the basis of color, we must ignore the fact that many non-blacks are as dark as many blacks. We may go beyond simple classification and, on the basis of selective experience, bias, or training, conclude that blacks and low intelligence or laziness go together without considering the question of how many exceptions there are to this theory or picture. Others may say that low intelligence and laziness are strongly tied to early environment, ignoring the role which inheritance may play. We develop theories about women, men, communists, Democrats, Jews, WASPS, and Irishmen, the rich, the foreign, teachers, electrons, liquor, and LSD. To the extent that we omit critical information, the results of our oversight or error come home to roost, causing us varying degrees of trouble. To the extent we can screen out information contrary to our theory or bias, we continue blithely on our way. So we propose that the first reason for bothering with theories and maps, those simplified pictures or stories of our world, is that with the information-processing equipment and ability that is built into us, we have no other choice.

Simplify Decision Making

Another reason for bothering with theories, maps, hypotheses, or hunches is that they simplify our decision making. This point is closely tied to the fact that we have no choice but to develop and act on theory because of the nature of the beast (who distorts, omits, rejects, and selectively receives and combines information) and the nature of the world around us (with the fantastic amount of information bombarding us each second). All of us base many of our day-to-day decisions on hunches, points of view, biases, or theories that we have developed over the course of time. Some of us have the ability to formulate a point of view very quickly, particularly if we have little information about the issue. It is much easier to develop a theory if you are relatively ignorant about the area in question. As one national leader replied to a reporter, "I would have a ready solution if I didn't know so much about it." In any case, there is no doubt that theories do simplify decision making by providing us with decision guidelines. These guidelines are generally in the nature of a reduction in the amount and kind of information we consider, and also in the number of possible alternative decisions we might formulate. In other words, the thousand and one potential decisions we might make on any question are reduced to one, two, or three merely because we have a theory which channels our thinking along certain lines and not others.

Theories Are Fascinating

A third reason for bothering with theories is that, for many people, theories are exciting to develop and discuss, regardless of whether the theory concerns why Cathy and Bill split up or why the stock market took an unpredicted and painful downward trend just after we made our purchases. Most of us, when faced with an unexpected event, have a strong desire to explain what caused the event or how it occurred. It's like a puzzle; we worry away at it until we've got some explanation that satisfies us. It's particularly gratifying if our theory is some dramatic or startling recombination and others can be swayed to accept our new interpretation. Predicting the Future A fourth reason for bothering with theories is our need to predict the future. We need to predict what will lead to what even when we lack necessary information. The. predictions may range from attempting to decide which of three job offers likely gives you the best chance of promotion and development, to predicting the future effects of continued smoking or of air pollution. Scientist and non-scientist all daily act on predictions. We use the weatherman's predictions to help us decide what to wear and whether to leave earlier because of predicted snow or sleet. We predict that the car will start, or if this prediction, much to our dismay, is disconfirmed, we predict that the public transportation system will be operating, that the driver is sober, that the brakes are sound. We predict that the water supply is pure, the cook is clean, the judge is honest. We predict that in our absence our roommates will not die, and our apartment won't be burgled. Thus we're constantly operating on the basis of a variety of predictions of varying probability, often without recognizing it.

Rewards from the Scientific Community

Finally, theories are important for scientists because the academic and scientific communities offer a variety of rewards to people who build acceptable theories and who test them. The eminent men and those to whom the accolades of science are awarded are usually the theory builders. In summary, we have theories, or simplified pictures of various parts of the world, because we have no choice. Second, we use them as decision making guides through a storm of information, opinion, and experience impinging on us from all directions. Third, for many, constructing stories about the past, present, and future is exciting. Fourth, theories about what is likely to happen assist us to move into the future with some confidence. Finally, the academic and scientific communities reward theory builders and testers. In the final section of this chapter we will develop some yardsticks which can be used to help us evaluate the adequacy of a theory. Before doing that, however, we will briefly examine one way science can be viewed-as a love affair.

BUILDING MODELS

Theory building becomes a game of building models, or pictures of the world. Some models are based on how we see the world through our own eyes, other models are based on how we see the world through an ordinary microscope. There are models based on how we see the world through an electron microscope, and models based on how we see the world through a telescope, or how we see the world through the eyes of a protestant biochemist, or how the world looks through the eyes of a Catholic gynecologist. Once a researcher or a theorist has stabilized a picture or a model, it usually requires a great deal of time, data, argument, and turmoil before he changes its super-structure. In other words, if the data coming in through his own research, or the research of others, does not fit his picture or model, he is more likely to question the adequacy of the data, or the competence of the researcher than he is to question the adequacy of his model or picture.

A Love Affair

Since any experiment is open to criticism, the theorist always has a way out in rejecting unwelcome data. The rejection of a theory once accepted is like the rejection of a girl friend once loved-it takes more than a bit of negative evidence. In fact, the rest of the community can shake their collective heads in amazement at your blindness , your utter failure to recognize the glaring array of differences between your picture of the world, or the girl, and the data. You will perhaps find it easier to understand some of the excitement and despair in the world of science if you view research as a love affair between the investigator and his project. During the initial courting stage he is open to certain kinds of information. But as he invests more time, energy, and money in the courtship, he becomes almost hostile to any information threatening the relationship. Perhaps as the data comes in, he has private moments of uneasiness which he shares with no one-as he analyzes the data he may have moments of agony-but it takes more than one or two lovers' quarrels to break up a love affair, which is just us well. If it were otherwise there would he almost no marriages, just as there would he very few, if any, worthwhile results from research. Fly-by-night relationships in almost any field yield little that is memorable or of lasting interest. We suspect there are those who would disagree violently with this love ;affair model of research. They will say that the researcher must be completely dedicated to objectivity, that he/she is only interested in the truth. Perhaps there are researchers like that. We haven't met enough to fill a phone booth. We have, however, met many researchers who can be brutally objective about someone else's research project, or about someone else's girl, but not about their own. But you are no doubt wondering what happens after a while- does the love affair model shift to the marriage model? Yes, we believe so. After working with a project for a long time we gain some objectivity and can accept some of the limitations and restrictions of our model. Many of us will still keep our serious family quarrels to ourselves. The very serious and established researcher can afford to joke in public about the possible limitations of his model, and speak philosophically about the relativity of truth in science. But if you are wise and humane you'll no more join him in making fun of his model than you would join him in his very personal game of making pseudo-fun of himself, his wife or his dog. As you know, making fun of something you love or cherish is one thing, having someone else do it is quite another.

THEORY EVALUATION

We proposed earlier that we bother with theories for a variety of reasons: (1) We have no choice in that we are theory-manufacturing or information-condensing organisms; (2) theories make decision making simpler; (3) theories are fascinating; (4) theories help us make predictions about the future, and; (5) theory production and testing are rewarded by the academic and scientific establishments. Therefore, while there seems to be a variety of reasons for bothering with theories, how are we to decide on the adequacy of one theory as opposed to another? Generally speaking a theory is useful to the extent it provides us with acceptable information in a shorthand or economical way, that assists us in making decisions and approaching our goals, or at least appears to help us avoid frequent high cost errors. More specifically, if we are comparing one theory with another we can use the following guidelines: (1) Which theory is the simplest to learn and use? (2) Which theory is more readily open to test? (3) Which theory provides us with sufficiently relevant and precise information at each step in our decision making in dealing with the question at hand? (4) Which theory provides us with the most unique and original information or allows us to predict the most new facts or solutions? (5) Which theory best fits with other accepted facts on theories? (6) Which theory is internally consistent, that is, doesn't contradict itself?

Simplicity

We like simple theories because they are easy to remember and to apply. Scotsmen are stingy, Englishmen are cold, blacks are lazy, Jews are pushy, Wasps are self-righteous, spare the rod and spoil the child, GM products are better than Ford products. While these theories or condensations are simple, they are obviously very imprecise; nevertheless, some are widely held. So, right or wrong, the simplicity yardstick is one of the most important. This is particularly so when, even though the theory is wrong, we personally don't suffer from its application. If, on then other hand, the application of an imprecise theory obviously does lead to our discomfort, we become interested in examining its relevance or its precision. Thus parents come to psychologists saying, ''even though I've whaled the daylights out of him, he still misbehaves. In fact, he seems to be getting worse." Such people are then ready for a more precise and more complex theory. Such theories might state, "On some occasions some children respond better to reward than to punishment," and for a more detailed background of this theory we might even be prepared to invest in and study a book on child rearing. Or, if the problem at hand is buying a new car, we may, through bitter experience, be forced to subscribe to the theory that only in certain years are some GM models better than Ford models, and acquaint ourselves with the theories or condensations of such publications as Consumer Reports. Notice, then that increased precision is usually purchased at the price of increased complexity in our theories. The general rule is: The more accurate we want to be, then usually the more complex the theory, and the more information we have to include in reaching our conclusion. You can usually spot a simple theory by its emphasis on one or two bits of information: behavior depends essentially on race, or behavior depends essentially on early training, or behavior depends essentially on biochemical factors, or behavior depends essentially on punishment, or behavior depends essentially on the institution that you are working for. While any one of these one-cylinder theories may have some validity, the more precision we want in our solution, the more likely we will need to combine these one-cylinder theories into multi-cylinder theories; behavior depends on intelligence, and on early training, and on genetics, and on biochemistry, and on work institutions. So we face the problem of reaching a balance between simplicity and precision. If we want to predict the developing behavior of a child, we may have to combine the series of one-cylinder theories. If, however, we merely want to predict the behavior of a particular bus driver (e.g., when he will come to our stop), we will usually be adequately accurate by consulting the bus schedule published by the institution that employs him. In most cases it would be irrelevant to worry about his intelligence, his early training, his genetics, and his biochemistry. Therefore it is not a question of simplicity versus complexity, it is a question of whether a theory or condensation includes enough information to meet our needs or help us make a decision. If we want to understand or predict complex human behavior, then theories, of necessity, must be fairly complex because (a) people differ, (b) they learn, (c) their behavior changes from one situation to another, and (c) they change with age. Therefore any time you encounter a theory about human behavior which is based on the assumption of stability, you will realize that such a theory leaves out a great deal of information. Examples include theories about introverts and extroverts, depressives and nondepressives, laziness and activity, responsibility and irresponsibility, etc. The reason why these theories have some appeal is: (1) they are simple, and (2) there are a few people who fit them. But they leave out most people. Most people are sometimes extroverted and sometimes introverted, sometimes lazy and sometimes active, sometimes responsible and sometimes irresponsible, sometimes security-seeking and sometimes risk-taking. Therefore the next time you hear a speaker describe a theory about human behavior that divides people into a few simple classifications you can bet that he is leaving out a great deal more than he is including, and that the major appeal of his theory is its simplicity.

Testability

While simplicity and personal appeal are two very important yardsticks in theory evaluation, the yardstick of testability is one which is presumably the most important for science. Think for example how you would go about testing the theory that southerners are less intelligent than northerners. It would be a relatively simple matter if we wanted to test the theory that that northerners are taller than southerners. We would obtain a nonelastic ruler and measure a large random sample of northerners and a large random sample of southerners. But intelligence tests are elastic rulers-the scores depend on (a) how the tests are administered and by whom, and (b) whether the people being tested and compared have been exposed to similar educational opportunities. Perhaps eventually some researchers will develop an acceptable test of intelligence that is as nonelastic as a tape measure, or at least not as elastic as the tests we now have, but, until better tests are available, the theory is relatively immune to test. Similarly if we wanted to test whether GM products are hotter than Ford products, we need a series of nonelastic yardsticks of what we mean by better and then we need to measure (test) a large number of different cars manufactured by the two companies. The reasons that make some theories difficult to test or evaluate include the following: (1) lack of nonelastic yardsticks; (2) inability to agree on which of the available yardsticks to use; (3) inability to measure a large representative sample of the population we are theorizing about; and (4) citizens and scientists who refuse to change their minds even in the face of new information. With some theories it is difficult to agree how they should be tested, and with others we are not prepared to invest the resources necessary to test them.

Novelty

Theories which lead to surprising or novel information are highly valued. Thus theory A raises few eyebrows, stating, "Students will get higher marks if they do one hour's study a day for ten days rather than if they do ten hours a day for one day." Eyebrows shoot high, however, if the theory states, "Students will get higher marks if they are exposed to a tape recording of lecture notes for one hour a night while they sleep, for ten nights, than by spending one hour a day for ten days studying the same material." Of course, theories that lead to novel or surprising information are not necessarily readily accepted. Unless such theories lead to an overwhelming amount of evidence, or permit a large number of researchers to test them readily and obtain the same results, these new findings may languish for many years in obscure journals until more and more evidence accumulates, or until the biases and attitudes of a sufficiently large number of the population change so that the new information becomes acceptable.

Goodness of Fit with Other Facts and Theories

As we noted at the outset, a few new facts do not change a well-established or lead to the acceptance of a new theory. This is not so merely because some scientists are biased, but rather because we all use familiar theories to evaluate new information_to help us in our decision making. If we gave every new bit of information or theory careful consideration, we would be overloaded with work in ten seconds. For example, drug companies put hundreds of new drugs on the market each year, and only a few can be adequately evaluated merely because of the time and effort required. Furthermore, thousands of research studies are published each year, but only relatively few are repeated (replicated) by other investigators. Most investigators are busy preparing to publish their own research. There is no simple solution to this problem other than using our own personally accepted theories, or small modifications of them, as guide-lines in helping us decide what we will read and examine carefully.

Internal Consistency

Another way of classifying or evaluating theories is to assess their internal consistency. This is perhaps one of the minimum conditions a theory must fulfill if it is to bo seriously considered. A theory which contradicts itself on any specific question proves embarrassing at times, even though, taken separately, each part is acceptable. Consider the following examples: (1) He who hesitates is lost and fools rush in where wise men fear to tread; (2) "Colonel Catheart was conceited because he was a full colonel with a combat command at the age of only thirty-six; and Colonel Catheart was dejected because although he was already thirty- six he was still only a full colonel." There are contradictions in the above examples and with such contradictions the theory provides no overall guidelines to aid us in predicting or in making decisions. According to theory (1) we should both buy and not buy the speculative stock, and according to theory (2) we really can't predict whether Colonel Catheart is happy or dejected. Until we have more information about the conditions surrounding lost opportunity because of hesitation, we can't use theory (1) effectively. Scientific or philosophical theories, as a rule, do not have such gross or obvious inconsistencies; however, often inconsistencies exist and are fair game for long arguments.

Understanding and Prediction

You will no doubt have heard that the purpose of theories is to help us both understand certain parts of our world and make predictions leading to new information and new solutions. Understanding is a difficult term to define. Perhaps one of the most common meanings of the term `'understanding'' is that we develop, sometimes with the help of others, a more satisfying picture of some part of the world. For example, we may ask "What's University like?" The reply may be, "University is like high school except that no one cares if you come to class, and in most classes you don't get any comments on the work you hand in." Or you may ask, "Why is Harry so cranky?" And you may get a reply, "He had a fight with his girl friend." Typically you will respond, "Oh I understand." You understand because you can now combine some bits of information that you already have in a new way- you now feel you have a better picture to go on. Our point is that understanding may give you a more acceptable picture about a part of the world, but, nevertheless. that picture may be quite erroneous. Unfortunately, you can walk away feeling you understand from a wide variety of different replies to the same question. It is proposed then that the term understanding implies a personally-acceptable picture of what goes together or what leads to what, but need not imply an accurate picture. The accuracy is subsequently, if ever, determined by more direct personal experience or by more accurate additional data from another source. The term "prediction" on the other hand, if stated in testable form is a more scientifically useful concept. Theories that enable us to understand are personally useful, while theories that enable us to predict are both personally and scientifically useful. This does not mean that the term "understand" cannot be redefined to include tests of the adequacy of the information; it is merely that the term, as commonly used, does not usually imply such tests whereas the term "predict" more frequently does. Therefore, we propose that theories that predict new information should be ranked more highly than those which lead to understanding, as defined above. We have stressed the importance of the testability of a theory. Testing a theory can be done in several ways. A theoretician may make a specific prediction growing out of his theory and may state his theory in a testable form which we then subsequently test and support or refute; or be can do an experiment growing out of his theory, and we can attempt to repeat his experiment. Replication in science is probably the foundation of testability, but, as noted earlier, we prefer the concepts of exportability or perishability of data to that of replicability. Summarizing, we are more and more inclined to view theories as decision-making aids, perhaps not as whimsical as the toss of a coin, nor as crude as a race tout's tip, but decision aids, nonetheless. In the face of ignorance, but forced to act, we build theories-these simplified and crudely drawn maps of the past, present, and future-that give us some semblance of confidence as we race or stagger through life's great maze of alternatives: Are theories true or false? No one knows, since most theories are designed to cover mammoth areas or massive populations, and yet we can usually explore in detail only a tiny corner or a few instances. Thus whether a theory is true or false is anybody's guess, whereas everybody knows that a decision aid is worthy if it helps you make even one decision that is not immediately followed by a disaster. End of story