Types of Observation

Uses and Cautions


Use of Questionnaires

Sampling Procedures

Unobtrusive Measures


Interviews and Tests

Case History


Classical Experiment

Design of Experiments

Multifactor Studies


Research Ethics

Comparison of Methods



n naturalistic observation, the most basic of the four methods, the aim is to study behavior in its usual setting, without asking the subjects any questions or administering any tests. The investigator simply observes and records what happens in the natural environment. For this reason, naturalistic observation is often the first step in a research program.

Each morning for several years, Milgram awaited the commuter train to New York City, and there he observed the same people at the same time, standing in the same places every day. Yet they almost never spoke to each other. Fellow commuters, he decided, are like trees, posts, and billboards--regarded as scenery, not as people with whom to talk and exchange greetings. Milgram called these people familiar strangers, for they encountered each other daily but never introduced themselves. Instead, they stood in clusters, back to back, staring straight ahead. "I found a particular tension in this situation,'' he confessed (Milgram, 1992).


The basic technique in naturalistic observation is to be a very careful observer. As an eager tourist noted: "You can observe an awful lot just by watching.'' Yet there are important decisions to be made and subtle procedures to be followed. Among these, the most basic is whether to acknowledge or disguise the research purpose.

Overt Observation. In overt observation, the subjects are aware that they are being studied; the research purpose is acknowledged. Milgram sometimes studied familiar strangers by taking photographs of people at train stations, later showing them to commuters, and then asking them whom they recognized. From their responses, he determined that the typical commuter encountered four or five familiar strangers at the station compared with only one or two speaking acquaintances.

Milgram's finding gave a specific, quantified answer. The average New York City commuter had 4.5 familiar strangers in his or her life. As Milgram explained, city people must discourage many potential relationships. "If you live on a country road, you can say hello to each of the occasional persons who passes by; you can't do this on Fifth Avenue'' (Milgram, 1992).

But the focal point here is that these commuters knew they were being observed and photographed. Milgram and the other researchers even stated their purpose. Hence, this investigation was an instance of overt observation.

Overt observation may not influence subjects at some distance or subjects who are sleeping, for example, but in other instances this research procedure could be disruptive. Commuters might alter their behavior to impress the observer, or they might avoid the observer, taking a different train. To deal with this problem, the investigator might spend time helping them become accustomed to his or her presence and the research procedure. After the subjects seem to be behaving naturally once again, the actual research begins.

From these observations, Milgram drew a conclusion about familiar strangers. When making a small request, such as asking the time of day, a person is more likely to ask a complete stranger than a familiar stranger--someone never spoken to but seen regularly for years. "Each of you is aware that a history of noncommunication exists between you,'' he said, "and both of you have accepted this as the normal state.'' Requesting even a small favor would disrupt this well-established, tacit agreement (Milgram, 1992).

Covert Observation. To ensure that the subjects behave in a natural manner, the investigator sometimes uses covert observation, in which the individuals being studied do not know they are part of a research project. The investigator can mingle openly with the subjects and then make notes secretly or remain hidden in some way. Of course, this effort to hide one's work may restrict the range of observation.

A question of research ethics emerges immediately. To what extent is a researcher justified in secretly studying commuters, coworkers, or even neighbors? The answer is complex, but it depends on the way in which the unsuspecting individuals are involved, the extent to which they may be affected, whether recordings are made, and so forth. We shall return later to this issue, noting in passing, that even overt observation raises questions about subjects' rights (Pope & Vetter, 1992).

Both methods, overt and covert, are used with animals. Field studies have stimulated much interest in chimpanzees, owing partly to certain similarities to human beings (Figure 2­1). Of course, the most celebrated naturalistic studies are those of Charles Darwin, whose trip aboard the Beagle enabled him to make observations of plants and animals around the globe.


Naturalistic observation serves two purposes. It provides an excellent description of certain phenomena, and it can be a rich source of hypotheses. As one investigator commented: "I find that during the long hours of observation in the field, I not only learn about behavior patterns, but I get ideas, 'hunches,' for theories, which I later test by experiments whenever possible'' (Tinbergen, 1965).

Figure 2-1 Observation of Animals.
Dian Fossey conducted extensive studies of wild gorillas in the mountains of Rwanda, Africa. Spending months at the edge of their territory, until they were comfortable in her presence, she gained their confidence by imitating their vocalizations and adopting a submissive posture. Ultimately, she assembled a large mass of information on gorilla habits and characteristcs. She found, in fact, that individual adult gorillas possessed quite distinctive personalities (Fossey, 1983).

There is a drawback, however. Naturalistic observation is not notably useful as a source of explanation. It does not identify cause-and-effect relations with any certainty. These must be examined in a more controlled setting, typically with the experimental method. In the study of memory, for example, there has been recent debate over the most fruitful approach, some urging naturalistic observations, others advocating controlled laboratory conditions. This difference of opinion is partly resolved when naturalistic observation is regarded as a good starting point for research but not as a substitute for controlled experiments to reveal the underlying causal factors (Roediger, 1991).

Even as a starting point, the process of naturalistic observation is not as simple and straightforward as it may seem. It presents the investigator with some difficult questions and the constant problem of bias.

Question of Participation. Some years ago, a small religious group in Chicago believed that the world would be destroyed by a series of floods and earthquakes on December 21. They would be saved by flying saucers, they decided, if they followed appropriate rituals, such as removing all metal from their clothing, remaining indoors, and reading the sacred writings. A team of psychologists and sociologists wanted to study them, but the cult did not permit outsiders to observe its activities. Thus, the investigators used their only recourse: They became cult members. Their research method was participant observation, in which an investigator joins the people being studied and takes part in their activities, living with them for an extended period, if necessary.

These investigators used covert participant observation out of necessity. If they had used overt participant observation, they would have been banished from the premises as disbelievers. If they had not participated, they would not have gained access to the group's activities. They knew their mere presence in the group would tacitly support the members' convictions about world destruction, but there was no alternative.

Our interest here lies in research methods, not the findings, but in passing the reader will be relieved to learn that the world was not destroyed on that December day. And the faith of the cult members was not destroyed either. On December 22 and 23, after some doubt, delay, and debate, the members decided that their Creator had not destroyed the world precisely because they had maintained their faith in the face of skepticism from others. Their unwavering loyalty had saved the whole world from destruction (Festinger, Reicken, & Schachter, 1956).

As a rule, researchers do not engage in the daily activities of their subjects. They generally remain apart from the people they are observing, a research procedure called nonparticipant observation. Stanley Milgram once observed crowds of pedestrians from a sixth-floor window. These people did not know they were being studied, and Milgram did not participate in their activities. His method was covert nonparticipant observation. Later, he stood in the street and openly made notes about the pedestrians: overt nonparticipant observation.

Problem of Bias. Recognizing several difficulties in naturalistic observation, William James stated that exact procedures for observation could not be established in advance. Rather, he advised the observer: "Use as much sagacity as you possess.'' He also warned of the great sources of error in this method, especially the intrusions of personal bias (James, 1890).

A bias is a preference or inclination that inhibits objective observation. It results in an inaccurate judgment. A man who is suspicious of people may make biased judgments about strangers. One method for dealing with this problem is to use several observers, assuming their biases are randomly distributed. Another is to train observers carefully. Still another involves the use of remote recording equipment (Pepler & Craig, 1995).

Concerned about this problem, Stanley Milgram had his city observers work in pairs. Sometimes a newcomer and a long-term resident toured the area together, walking side by side down the street, but they made their recordings separately. This way, Milgram had a check or verification on what had taken place, for he believed that the long-term resident, while sensitive to nuances, might have the habit of tuning out many events noticed by the newcomer. City life perhaps required that sort of adaptation (Milgram, 1992). Using naturalistic observation, Milgram identified the phenomenon of familiar strangers and other habits of city dwellers. But he could not, through observation alone, determine the underlying causal factors. And he typically could not study significant moral issues in detail (Table 2­1).


Waiting with other communters, taking notes obviously

Waiting with other commuters, taking notes secretly
Standing aside from commuters, taking notes obviously

Hiding from commuters, taking notes secretly

TABLE 2-1 TYPES OF OBSERVATION. Two decisions, overt or covert, and participant or nonparticipant, yield four research strategies. The chief issues are case of observation and disruption of subjects' behavior.

© Prentice-Hall, Inc.
A Simon & Schuster Company
Upper Saddle River, New Jersey 07458

Legal Statement