Chapter 6 Summary

Chapter 6 Summary
Field Research:
Naturalistic and Case-Study Research

The Challenge of Low-Constraint Research

Field research is a term that applies to a variety of research methods, ranging from low- to high-constraint. These methods focus on observing naturally occurring behavior under largely natural conditions. They include naturalistic research and case studies.

Naturalistic and case-study methods are at the lowest level of constraint in scientific research. Researchers observe the behavior of participants in a flexible way, which allows them to take advantage of sudden occurrences and new ideas that develop in the course of the study. The focus of low-constraint research is on the natural flow of behavior. Therefore, the investigator intervenes as little as possible, imposing few constraints on the behavior of participants.

Case-study research is slightly higher in constraint than naturalistic research, with the researcher intervening to some degree. Case-study research is usually carried out with one participant at a time, often in an interactive, face-to-face situation. In naturalistic and case-study research, the constraints are primarily on the observer, with little control exerted over the participant.

Examples of Naturalistic Observation

The defining characteristic of naturalistic observation is that the researcher observes and records naturally-occurring events and later develops hypotheses about why they occurred. Classic examples of naturalistic research include the work of Charles Darwin and Alfred Russel Wallace on biological diversity and natural selection, Jane Goodall’s classic study of chimpanzees, Adeline Levine's study of the Love Canal disaster in Niagara Falls, and Phillip Davis’s study of parents disciplining children in public settings. An example from psychology is the work of David Rosenhan, who investigated the use of psychiatric diagnoses and the experiences of mental patients in hospitals.

Examples of Case-study Research

Case-study research focuses on behavior with little constraint placed on the participant by either the researcher or the setting. Case-study research is like naturalistic research in that it imposes very little constraint on participants. However, it is not typically carried out in completely natural environments. Case-study research is typically focused on individuals, and it usually looks at limited classes of behavior, rather than at the total context and natural flow of behavior.

Classic examples of case-study research in psychology are the work of Sigmund Freud, who observed individual adult clients in psychotherapy sessions, and E. L. Witmer and Jean Piaget, who observed individual children. Examples of contemporary case studies can be found in the investigations of puzzling clinical disorders. For example, some people have severely distorted perceptions of their bodies, leading them to bizarre behavior, such as demanding the amputation of healthy body parts. Research is just beginning in this area, and low-constraint case studies are appropriate approaches at this time.

The Value of Low-Constraint Methods

Low-constraint methods describe phenomena and lead to hypotheses and subsequent research to understand them.

Conditions for Using Low-constraint Research

Low-constraint research is used: (1) to study the natural flow of behavior in natural settings; (2) at the beginning of a research project or in a new area, as in exploratory research; (3) to demonstrate a new research or treatment procedure; (4) to check the generalizability of findings from the research laboratory; and (5) to ask questions about a specific person who is the object of the study.

Information Gained from Low-constraint Research

Low-constraint research is not necessarily poor research. It is useful, and sometimes it is the only way to proceed. It can yield important information when used appropriately. It is the only way to study the natural flow of behavior in natural settings. Low-constraint research can provide information to negate a general proposition, although it cannot establish a general proposition. It can also establish contingencies among variables, which can be studied in more detail with higher-constraint research.

Using Low-Constraint Methods

In recent years, low-constraint research methods, such as used in ethological research, have been extended to other disciplines and have come to be known collectively as qualitative research methods. These include naturalistic and participant observation, use of questionnaires, and analyses of conversations.

Problem Statements and Research Hypotheses

Although problem statements and their ultimate development into hypotheses are most fully developed at the experimental level of constraint, they are important at all levels of constraint. They help us to organize our thinking, to specify how we will carry out the research, and to focus on the inferences that we can confidently make based on the data obtained in the research.

The inferences vary among constraint levels. At the experimental level, the problem statements and the research hypotheses are focused on questions of causality. In differential studies, they are focused on determining differences between groups. At the correlational level, they are focused on the direction and strength of relationships among variables. At the naturalistic and case-study levels, they are focused on identifying contingencies. We cannot confidently draw causal inferences from low-constraint research.

Problem statements in low-constraint research are often general and they can change as the researcher grasps new issues. The strength of low-constraint research is being able to move flexibly from one issue to another depending on what is found.

Making Observations

The central phase of any research project is the observation or data-gathering phase. In low-constraint research, the planning for those observations is less formal and more fluid than it is in experimental research.

There are two general ways of making observations in naturalistic and case-study research: as an unobtrusive observer or as a participant observer. Both procedures tend to conceal from the participants the fact that they are being observed. Measurement reactivity can result when participants know that they are being observed. One way to reduce such reactivity is by use of unobtrusive measures (i.e., measures that are not obvious to the person being observed and thus are less likely to influence the person's behavior). Information can also be gathered from already existing records (archival records) and numerous other non-reactive measures.

Sampling of Participants

Participants should be selected so that they are representative of a population, which is a process called sampling. Representative samples allow us to make generalizations from the data to the population. When we generalize findings, we assume that what was observed in the sample of participants is characteristic of the population as a whole. In most low-constraint research, we usually do not have the opportunity to select our own samples. Therefore, we must make careful judgments about how well the sample actually represents the population.

Sampling of Situations

The sampling of situations also affects generalizability. The broader the sample of situations studied, the more confidence we can have in the generalizability of the findings.

Sampling of Behaviors

The term sampling generally refers to the selection of participants. Representativeness of the sample refers to how closely the sample resembles the population to be studied. A related issue is the importance of adequately sampling behaviors within any given situation. Organisms may behave in many different ways. Therefore, a single or small number of observations may not give us an accurate picture of the organism's functioning. By sampling behaviors repeatedly in a given situation, it is possible to recognize behavioral variability when it exists. Also sampling behavior in several different situations will help generalizability.

Evaluating and Interpreting Data

By its very nature, low-constraint research uses few controls. Therefore, we must be cautious in analyzing and interpreting data obtained in low-constraint research. The purpose of controls is to eliminate alternative explanations of the results, thus making it easier to draw a strong conclusion. Since those controls are absent from this level of constraint, we are seldom able to draw strong conclusions from a low-constraint study.

Limitations of Low-Constraint Methods

In low-constraint research, representativeness can always be a problem and, therefore, conclusions must be tentative. The temptation to go beyond the data must be resisted.

Poor Representativeness

Because representative samples are not used in most low-constraint research, one should not generalize from the specific participants studied to a larger population.

Poor Replicability

Because the procedures are so flexible in low-constraint research, replicability is often impossible to obtain. Replication requires a thorough knowledge of the details of a study, which are often unrecorded in low-constraint research.

Causal Inference and Low-Constraint Research

Drawing causal inferences from case studies is virtually impossible. The ex post facto fallacy involves drawing a conclusion that variables are causally related based on research in which the independent variable was never manipulated and in which, therefore, alternative explanations cannot be ruled out. Such conclusions are formed "after the fact." We might observe, for example, that most adult alcoholics report that they drank beer as teenagers, long before becoming alcoholics. If we concluded that beer-drinking leads to alcoholism, we would be committing an ex post facto fallacy.

Limitations of the Observer

In a highly interactive setting, such as in Freud's therapy sessions with his clients, it would be extremely difficult to control experimenter reactivity or experimenter bias.

Going Beyond the Data

One of the problems with low-constraint research is that we may be tempted to go beyond the data--that is, to draw conclusions that are not clearly supported by the research or by ignoring alternative interpretations.

It is important for researchers to understand the limitations of these low-constraint methods. Knowing the limitations will allow us to use the methods appropriately. When used appropriately, these methods are extremely useful. The danger is in the temptation to infer causality erroneously, to generalize beyond the participants studied, to consider the findings certain rather than very tentative, and to fail to recognize how the observer might be reactive, thus influencing the outcome of the study.

Ethical Principles

Although a major ethical principle in human research is informed consent, many low-constraint studies do not get informed consent because they are simply observing he natural flow of behavior without causing undue stress. Nevertheless, an IRB should review the proposal and agree that consent is not needed before the research begins a naturalistic study.