Experimental Methods in Psychology

Gustav Levine, Stanley Parkinson

• Publisher: Psychology Press
• Year: 2014
• ISBN:
  9781317781035
• Edition: 1
• Topic: Psicología
• Subtopic: Historia y teoría en psicología

This text focuses on the experimental methods and the associated terminology encountered in the research literature of psychology. Initially, the content is kept simple, so as not to distract from the information on research technique and philosophy. Interesting psychological questions from well researched areas are then examined in detail, permitting a fuller discussion of the problems encountered in specific paradigms. It is in this fashion that the book offers both methods and content. Unique features of this text include:
* a detailed discussion of the process of theorizing, coupled with a close examination of psychological constructs, offers the reader an opportunity to see how psychologists think about, develop, and modify their theories, and the part played by research in changing explanations of behavior.
* Although it is common for psychologists to be self-conscious in their reasoning, it is uncommon to see an analysis of the logic that they use to draw conclusions. Presenting material that is rarely verbalized but readily acknowledged by experienced researchers, the text contains an overt analysis of the logic of drawing conclusions from research.
* Instructors are given a choice among 15 chapters to focus on or combine to suit the course's concentration. For example, instructors have the option of focusing on experimental psychology or a broad-based course including material on research methods in experimental, social, clinical, and applied psychology.
* Courses in experimental psychology or research methods are required for every psychology major. Statistical understanding is vital for this curriculum, and this text contains a comprehensive chapter on statistics making it ideal for courses that combine statistics and experimental methods. Other important coverage includes:
* an all-inclusive summary of the material found in an introductory statistics class. Although courses in research methods and experimental psychology usually have a statistics prerequisite, the students rarely remember the material when entering the research course. This text provides the instructor with the option of simply assigning the statistics information as a review, rather than repeating the lectures. If the course requirements are such as to necessitate a joint statistics and research methods course -- with the instructor lecturing on both topics -- this text could serve as the single text for the course. A helpful discussion -- accompanied by a valuable table -- demonstrates how to choose an appropriate statistic. All necessary formulas and other familiar statistical procedures -- illustrating computational steps -- are also featured.
* a detailed discussion of how to develop tests for use in research. Aside from the value of this information for any researcher, it can be particularly helpful to students who are required to develop original experiments.
* an elaborate discussion of methodological issues in outcome research, using smoking cessation and weight reduction programs as examples. Test bank disks for Experimental Methods in Psychology, -- free to adopters -- consist of an average of six short-answer, 11 fill-in-the-blank, and 11 multiple-choice questions for each chapter. The files are in both ASCII and Word-for-Windows formats.

CHAPTER 1

Introduction: The Function of Research Methods

People were gathering information, and successfully using much of this information, long before science came along. The survival of the species is evidence that humans are capable of obtaining useful information without invoking the scientific method. Yet we are all familiar with incorrect information that was believed for long periods of time by intelligent people. For example, there was the belief in bloodletting to reduce fever (which weakened the patients and sometimes killed them), and there was the use of boiling oil to cauterize battlefield wounds when soldiers or sailors lost limbs (which created a toxic reaction from the burnt tissue, increasing the death rate). Remedial procedures sometimes appear reasonable, and are reported by observers to be effective, yet are in fact useless, or even harmful. It must be concluded that there are circumstances where experience and observation can be misleading, and that prior beliefs, or other factors, can affect what is observed or remembered (Nisbett & Ross, 1980). Some of the observational circumstances that are most vulnerable to such errors have been identified. Ways of rearranging the collection of observations to avoid these errors have been developed. These better arrangements for making (and interpreting) observations constitute the largest part of what is meant by research methods. In fact, an informal definition of science might be that science is a series of techniques to help people avoid fooling themselves.

The special arrangements for collecting observations that are integral to research procedures are detailed throughout this text, but an example is useful at this point. There is an interesting accidental experiment that occurred naturally and resulted in the abandonment of the use of boiling oil for cauterization. Boiling oil cauterization was used during the 15th and 16th centuries, having been recommended by Giovani Divego (1460–1520) in his classic text on surgical procedures. During military battles army and navy surgeons always had boiling oil handy, which they would apply after amputating a mangled arm or leg. Civilian surgeons followed the same procedures when accidents produced similar injuries.

The 16th century French army physician Ambroise Paré ran out of oil during a battle. The result was two groups of wounded men, one group receiving the burning oil cauterization, and another group, otherwise similarly wounded and similarly treated, who did not have the burning oil applied after the amputation procedure. Contrary to expectations, the surgeon found far more survivors, and better healing, among the noncauterized group, and reported this finding to his colleagues. Paré was something of a nay-sayer, and so was probably happy to be able to show that an established treatment was incorrect. If this dissident physician were simply to claim that in his experience cauterization was an undesirable treatment, his colleagues would have ignored him. As discussed in the following section, it was the form of the evidence that caused the profession to treat the new information seriously.

CONTROL GROUPS

The contrast of two similar groups, the only difference being that one group receives a specific treatment and the other does not, is an excellent situation for testing a treatment's effectiveness. It permits two conditions to be compared, where only one contains the component being tested. This is an example of one kind of observational circumstance that a researcher would deliberately create in order to test a treatment. In this research design, the group without the treatment is called the control group or control condition, and the group with the treatment is called the experimental group or experimental condition. It took the presence of a simultaneous control condition, along with an experimental condition (the subjects in the experimental condition receiving the treatment), for the surgeon to be able to recognize that the boiling oil treatment was a poor one. The opportunity, simultaneously, to compare the results in a control group with the results in an experimental group made it easier for other physicians with faith in the treatment to attend to the new contrary information.

Control Variables

In planned experiments care is taken to maintain identical conditions and influences in the experimental and control conditions, to see to it that the situations are the same. Only the treatment that is being tested is permitted to vary from one condition to the other.

The influences and details that could vary if permitted to do so, but that are deliberately kept identical in the two groups, are called control variables. If an experimenter wished to test boiling oil cauterization, he or she would include the same kinds of amputations, same ages of patients, same conditions of hygiene, same aftercare, etc. in both the control and experimental conditions. The control variables would then include types of amputations, ages of patients, hygienic conditions, and aftercare.

CONDITIONS THAT FOSTER MISPERCEPTION OF EVIDENCE

A reasonable question is, why do people need such sharp contrasts, that is, the simultaneous presence of both a no-treatment control group and an experimental group, to recognize when a treatment does or does not work? For example, surgeons saw many oil-cauterized patients prior to the accidental experiment of Ambroise Paré, and occasionally saw patients who lost limbs under conditions where no oil was available at the time of their accidents. Why couldn't those surgeons use their memories and accumulated experience to recognize that the use of boiling oil increased deaths? There are a number of conditions that can foster misperception of evidence.

A major cause of distortion of evidence is an observer's prior belief that a treatment works (or does not work). People attend more to information that is consistent with their beliefs and are prone to more often miss evidence that contradicts their beliefs (Greenwald, Pratkanis, Leippe, & Baumgardner, 1986). When evidence accumulates slowly over time, prior belief can permit a person to recall events selectively, forgetting some of the evidence. One way of reducing the effects of prior belief is to have all of the evidence (outcomes) simultaneously present for examination. With planned research, all of the outcomes are recorded and then simultaneously viewed at the end of the experiment.

An additional basic ingredient in self-conscious use of research design is the careful maintenance of similarity between the control and experimental conditions, so that the only difference between the two groups is the presence or absence of the treatment being tested. This requires scrutiny of the two conditions, and searching for control variables that should be held constant. This can be compared with informal (nonexperimental) gathering of information and experience by a physician in practice in Paré's time. It is possible that a particular surgeon only used boiling oil when he removed a limb in his office, but usually did not when patients arrived with a severed limb from a street accident (because the bleeding had stopped). In this case, there would be a number of differences between those treated and not treated with boiling oil. Street accidents could introduce more filth into the wounds, or there could be greater loss of blood, and so on. Thus, extraneous factors (that should have been kept identical in the two conditions for a proper comparison) could be responsible for making a useless or harmful treatment appear effective. The maintenance of identity between the experimental and control conditions (with the exception of the treatment being tested) is probably the single most important principle in good research design. Unfortunately, it can sometimes be a difficult task, with unwanted differences being subtly introduced. Undesired factors that additionally differentiate the experimental and control conditions are called confounding factors. A confounding factor is differentially present in the two conditions, just as the treatment is present in one condition but absent in another. Thus, with a confounding factor present, there are two differences between the experimental and control conditions, which then make it impossible to know whether the treatment or the confounding factor is responsible for any effects. The avoidance of confounding factors is discussed in chapters 2 and 3.

There is another interesting phenomenon that operates to prevent people from reaching proper conclusions from evidence. Sometimes when an accepted treatment is applied, if it fails, the patient is blamed. For example, those who die despite receiving a standard treatment are sometimes declared to have been constitutionally weak. On the other hand, if the patient gets well, the treatment is credited for the success. Prior beliefs are similarly maintained when the treatment is not applied. Those who die are thought to have died because they could not receive the treatment. Those who live are thought to have lived because of unusually strong constitutions, or perhaps because of the intercession of prayer.

Similar explanations have operated quite dramatically to maintain confidence in some inadequate forms of psychotherapy. Many different kinds of psychotherapeutic treatment have been developed over the years, and although a number of them can be helpful, some of them lack convincing evidence of effectiveness, yet all of them have their devoted advocates. The claim has generally been made by those using most treatments that patients who do not respond are not motivated, or are not ready for therapy, or do not want to get well. This offers a ready explanation for any individual lack of success. Because a certain percentage of people spontaneously recover with or without therapy (more favorable circumstances at work, a new love or friendship, discovery of a new source of pleasure, a new philosophy or faith, etc.), even an ineffective but busy therapist will see some people whose state of mind appears to improve. By dismissing those who did not improve as in some sense not ready for therapy, and accepting those who improve as having benefited from the therapy, any therapy, and any therapist, can appear efficacious.

The point is that given an assumption that something works, the occasions when it does not work are often explained away. It is rare to find a phenomenon that human beings cannot explain to their own satisfaction, when they are motivated to do so. Research methods are arrangements for making and interpreting observations that make it difficult to explain away evidence selectively—that make it difficult to ignore information that contradicts old assumptions.

A DEFINITION OF SCIENCE

The preceding discussion should make it clear that science is not just the organized information that has accumulated from the application of research methods, but includes the research methods themselves. The research methods are arrangements for making and interpreting observations in a manner that minimizes the probability of reaching incorrect conclusions. Thus, science could be defined as both an accumulation of knowledge and an accumulation of research methods that function to limit self-deception by serious observers. It offers some protection against the influence of prior beliefs and the various forms of bias that often are spawned by enthusiasm about an original idea.

OBJECTIVE METHODS, NOT OBJECTIVE SCIENTISTS

The point has been made that good research design is necessary because personal bias can distort observations and conclusions. Yet it is occasionally implied that scientists are objective individuals, and that it is their scientific detachment and tradition of careful observation that yields the more reliable information on which the society depends. Nothing could be further from the truth. Scientists, including psychologists, like other human beings, can (and usually do) fall in love with their own hunches, ideas, and theories. This is in fact desirable. Interesting new ideas are often difficult to validate. It frequently takes a great deal of passionate conviction and excitement about an idea to pursue the difficult path of confirmation. Therefore, the passion that scientists have for their own ideas is in fact useful, and probably necessary. Scientists can be expected to be protagonists for their own explanations, and to seek to find data in support of their own theories. On the other hand, this is what makes it so necessary for the scientist to use a set of techniques that can identify useless concepts and incorrect conclusions, and that can force the scientist to recognize unexpected facts. The scientific method and its research procedures perform this function, permitting the scientist to maintain his or her enthusiasm and the bias that this sometimes entails. Although the goal of avoiding the effects of the researcher's bias is not always realized by the use of standard research procedures, these procedures are far more effective than just a reliance on careful observation and the best of intentions. An additional factor strengthening the scientific method is the requirement of recording not only what was observed, but what was done. This too involves special procedures and concerns, because descriptions of what procedures were followed could themselves be affected by bias. Thus, in each scientific specialty, characteristic ways have developed for how an experiment is to be described, permitting other scientists to check and see if the scientist doing the research has overlooked possible confounding factors or has otherwise compromised the experiment. This standardization of procedural description allows other scientists to repeat the experiment and see if the results do in fact come out as reported by the prior observer. Thus, the entire enterprise is geared to permit the rest of the scientific community to peer over the shoulder of the person doing the research, and to minimize the influence of the expectations or hopes of the researcher.

The goals of avoiding bias and confounding factors are not always met, so future researchers sometimes have to ferret out the experimental errors of earlier researchers. But the requirement of full description of what was done permits the members of the scientific community within a discipline to keep track of the details of each other's research, and to criticize each other, and to redo or improve each other's experiments. The result is that the scientific community, in its research role, does offer the best source of unbiased information that the society has. Yet there is no requirement, or assumption, that scientists are personally any more objective than any other group of people. Objective evidence refers to evidence gathered in a specific fashion, and does not refer to evidence gathered by mythical, unbiased people.

SUMMARY

There are circumstances where even careful observations can lead to incorrect conclusions. Some of these situations have been identified, and procedures to reduce the probability of reaching erroneous conclusions have been developed. These procedures constitute the methods used by researchers. Use of these procedures is synonymous with good research design.

An important component of good research design is the use of (at least) two similar groups, the only difference between the two groups being the specific condition or treatment being tested for effectiveness. The group with the treatment is called the experimental group, and the group without the treatment is called the control group. The factors in the situation that are deliberately kept identical in the two conditions are called control variables. When it is time to draw conclusions from a research project, to avoid distortions in memory, all of the evidence is simultaneously examined.

Although maximal effort is devoted to avoiding any systematic differences between the experimental and control conditions, this attempt is not always successful. Factors that different...