Fundamentals of Social Research
Adam J. McKee, Ph.D.
This content is released as a draft version for comment by the scholarly community. Please do not distribute.
Much like survey research, observations can be structured or unstructured. The major difference between survey research and observational research is that survey research seeks understanding through asking; observational research seeks understanding through observation. A potential advantage of observational research over survey research is objectivity. Asking people about criminal behaviors, for example, may not be the best data gathering strategy. Observation is much more likely to yield accurate data in these sorts of situations.
Structured observations are based on a set of predefined behaviors that are known before the observations begin. Observers record the frequency and magnitude of these predefined behaviors. Because numerical measurements are often taken (frequency and magnitude), this type of data collection method is often used in quantitative studies. When little is known about a phenomenon or the researcher is interested in aspects of behavior that are not quantifiable, then unstructured observations are used. Anthropologists, for example, often use unstructured observations in studying unfamiliar cultures.
The Natural Setting
While some observational research is conducted in a laboratory setting, the vast majority of observational research is conducted in a natural setting. In conducting research in a natural setting, the researcher is essentially relying on natural processes to manipulate the independent variable rather than the researcher manipulating the independent variable directly. This is a key difference between observational research and experimental research. Recall that experimental research requires that the researcher manipulate the independent variable—a treatment must be administered. While scientists would always prefer to conduct a true experiment, it is not always possible to do so.
A primary reason that experimental designs are not always available is that social scientists, especially those in applied fields such as social work and criminal justice, are keenly interested in bad things happening to people. It would be unethical in the extreme to do harmful things to people in the name of science. To study harmful social phenomena, social scientists must find individuals that have already experienced the harmful independent variable in a natural (non-laboratory) setting and make observations. Thus, social scientists often rely on observational research because the idea of conducting true experiments would be unethical.
A potential problem with empirical observational research is that these designs do a very poor job of supporting causal statements. Recall that the hallmark of an experiment is the idea of control. Extraneous variables are ideally eliminated by the research design. Because observational designs lack these elements of control, it is impossible to rule out the spurious effects of these uncontrolled variables. Because it does a poor job of making causal statements, observational research is often limited to the classification of descriptive research.
A Note on Scales
Most of the variables social scientists are interested in defy easy definition. They are abstract concepts. Because of this, a single question will rarely be useful in measuring the variable. Take intelligence for example. Most people have some idea of what intelligence is, but operationalizing the concept can be quite difficult, and experts disagree on the finer points. However it is defined, there is no single question that can be asked to reveal if a person’s intelligence. Psychologists have developed batteries of questions (IQ tests) that, when taken together, give a much better idea of a person’s intelligence than a single question ever could. Composite measures consisting of several items (questions) are known as scales. Scales usually result in a numerical score and are thus best suited for quantitative research.
Sometimes, researchers will devise scales, test them extensively, and publish them. Well tested, published scales are often referred to as standardized scales. Psychologists use many scales, such as the MMPI and the IQ test.
The most commonly encountered type of scale in social science research is known as the Likert scale. For this type of scale, the respondent is given a series of statements, and asked to indicate whether they Strongly Agree, Agree, Have No Opinion, Disagree, or Strongly Disagree. These five category Likert scales are common, but others are commonly used as well. All of the variants are still referred to as Likert scales. Note that each category indicates a different level of agreement. If we assign numbers to each point along the scale, an “agreement” score can be easily computed.
In this form of observational research, the researcher is not directly involved in the situation being observed. The idea is for the researcher to look in from the outside, and have no impact on the social situation under observation. The most common form of nonparticipant observation is known as naturalistic observation. In this type of researcher, the researcher purposely controls and manipulates nothing. The overarching idea is to record social phenomena as they naturally occur. Let us say, for example, that a researcher is interested in interactions between citizens and police officers. A method of naturalistic observation would be for the researcher to participate in a “ride along” program, accompanying police officers as they perform their daily duties. The researcher would make careful observations about the interactions, but would never become involved. Naturalistic observation studies often provide the framework for more controlled types of research later on.
Participant observation is perhaps the most difficult type of research to conduct. This is because the researcher must be a full participant in the social phenomenon under study as it is occurring, as well as maintaining objectivity and accurately recording events as they unfold. Researchers trying to describe the characteristics of the police subculture have used this method extensively, for example. The key advantages of this method include a depth of understanding that cannot be paralleled by more objective quantitative methods. The tradeoff is the potential lack of objectivity that may result from a lack of detachment and neutrality on the part of the participant observer.
A focus group is a gathering of people that is moderated by a researcher. Usually, the group is composed of seven to ten individuals who are not familiar with each other. The big idea is to use the group interaction to obtain information about a specific, focused issue. For example, a group of citizens could be brought together to examine attitudes toward a new police policy regarding a city’s juvenile curfew. Perhaps the most difficult aspect of conducting focus groups is preventing researcher bias from influencing how participants respond. It must be remembered that the ideas and opinions of the focus group members are what are important, and the researcher should make every effort to eliminate bias. Another critical function of the moderator is to ensure that every member has a voice, and that a few very vocal participants do not dominate the discussion.
Careful researchers will seek out diverse opinions among the group members, and not seek to build a consensus. It is important to find out what participants’ opinions are, not to change them. Researchers will pay close attention to the dialog, carefully noting any trends and patterns in the groups perceptions of the research topic.
One of the major problems with focus groups is the generalizability of findings. Keep in mind that focus groups are not random samples. They were purposely chosen because of some specific characteristic. Other groups with different characteristics may not hold the same views and opinions. A major advantage of focus groups is that data can be gathered much more quickly (and at less cost) than would be the case with individual interviews. Natural group interactions make the process more “natural,” ideally improving the face validity of the findings. In applied research settings, the words used by focus groups are easier for non-researchers to grasp than statistical analyses of survey data. The success of a focus group depends largely on the preparation and skill of the researcher. Such groups only produce useful results when participants are made to feel comfortable and free to speak their mind.
Modification History File Created: 07/25/2018 Last Modified: 07/25/2018
This work is licensed under an Open Educational Resource-Quality Master Source (OER-QMS) License.