Experiments

Abstract

Experimental research is one of the primary ways by which science advances. In experimental paradigms, independent variables (stimuli) are manipulated and the effect of the manipulation on the value of the dependent variable (response) is measured and analyzed. In addition, good experimental research is designed to control as much as possible the effects of extraneous and intervening variables, often by the use of control groups. Experimental research can be performed in the laboratory, through simulations, or through the manipulation of variables in field experiments. These approaches to experimental research allow the design of increasingly more realistic experimental paradigms but give the researcher decreasing control over the experimental situation. In addition to being concerned over designing an experiment that will collect uncontaminated data that can be meaningfully analyzed using inferential statistics, researchers using human subjects also need to be concerned about ethical considerations of the effects of their experiment on its subjects.

Overview

In order to be considered a true behavioral scientist, it is necessary for sociologists not only to observe and describe human behavior, but to perform research so that they can better understand and predict behavior. In some situations, it is only possible to collect data using surveys—data collection instruments used to acquire information on the opinions, attitudes, or reactions of people. In other situations, however, it is actually possible to control the research situation and manipulate variables in order to obtain better data and a clearer picture of the processes that underlie human behavior. These experiments are situations that are under the control of a researcher in which an experimental condition (independent variable) is manipulated and the effect on the experimental subject (dependent variable) is measured. Most experiments are designed using the principles of the scientific method and are analyzed using inferential statistics to determine whether or not the results are statistically significant.

Experimental Variable. In the simplest experimental design, a stimulus is presented to the research subjects and a response is observed and recorded. For example, one might present subjects with pictures of various situations and ask them to describe what they think they would do if they were in that situation. However, what one thinks one would do and what one would actually do can be very different and may depend on a number of factors. Behavior in the real world tends to be complicated, and several types of variables need to be considered when designing an experiment. Of most concern in most research paradigms are the independent variable (i.e., the stimulus or experimental condition that is hypothesized to affect behavior) and the dependent variable (i.e., the observed effect on behavior caused by the independent variable). However, there are typically other variables that need to be considered and controlled as much as possible, particularly in real-world research. As shown in Figure 1, extraneous variables that affect the outcome of the experiment but that have nothing to do with the independent variable may also need to be considered. For example, a person who responded to a picture of an automobile accident by saying that he or she would not stop to help might respond that way because he or she was feeling tired or ill that day and wanted nothing more than to get home. On another day, he or she might give a different response. In most real-world situations, there are innumerable variables that are extraneous to the research question being asked but that still affect the outcome of the research. However, a well-designed experiment is created so that it controls for as many of the extraneous variables as possible. It is, of course, impossible in most cases to anticipate and control for every possible extraneous variable. However, the more of these that are accounted for and controlled in the experimental design, the more meaningful the results will be.

Another type of variable not directly related to the experiment but that might affect the results is the intervening variable. These are things that occur between the manipulation of the independent variable and the measurement of the dependent variable. For example, if a person in the experimental situation responding to a picture of an automobile accident responded that he or she would call 911 but would not stop to help later took a course in CPR, he or she might actually stop to help if encountered with the situation in the real world because of the intervening training. Like extraneous variables, intervening variables need to be controlled as much as possible in the experimental situation so that the effect of manipulation of the independent variable on the dependent variable can be determined and statistically analyzed. Extraneous and intervening variables are often controlled in an experiment by the inclusion of a control group comprising subjects that do not receive the experimental condition in order to level the effects of these variables.

Types of Experiments. There are a number of ways to collect data that can be used by behavioral scientists. These range from laboratory experiments that allow the researcher great control over the conditions and variables in the study to secondary analysis methods that allow researchers to examine the results of studies done in the past but give them no control over the research design or data collection whatsoever. Laboratory research allows the most control over variables. However, it often is far-removed from real life. Laboratory methods tend to be more appropriate to basic research questions where the influences of the real world are not as important as in applied studies. However, as the research situation becomes more realistic, the research loses a greater degree of control over the situation. To be able to extrapolate research results to the real world and have them be meaningful, it is important to design an experiment that not only controls extraneous variables, but also is as realistic as possible.

There are several general types of experimental studies that can be used to explore the behavior of people in the real world. As discussed above, laboratory experiments allow researchers the most control over extraneous variables. However, laboratories tend to be far removed from the reality of how most people live their lives. Simulation is an approach to experimental research that allows a more realistic setting for the experiment while still allowing researchers a great degree of control. Using the example above, research subjects could be placed in a situation where they encounter a simulated automobile accident so that the researcher could determine how they actually would respond in such a condition. Or, the researcher could set up a field experiment in which the experimental condition is introduced into the real world and the researchers observe how people react. For example, the researcher could set up a situation where it looks like someone has had an automobile accident and then determine what percentage of people actually stop. In an effort to gauge whether disability affects an individual's chances of being considered for a job, one study involved sending out a certain number of applications detailing physical disabilities for open accounting positions that would not be limited by such disabilities to employers to see how many selected the candidates for potential hiring (Ameri et al., 2018). Both simulations and field research have the advantages of being increasingly more realistic and, therefore, more likely to elicit real responses from people than is possible in the artificial setting of the laboratory. In the twenty-first century, sociological researchers have noted the increased usage of online field experiments as another, convenient method for those leading a study to take advantage of the benefits of gathering results from a real-world, contextual social setting. These types of experiments also have drawbacks, though, such as concerns over the potential invasion of privacy to conduct online experiments, particularly those using social media sites (Keuschnigg & Wolbring, 2016). Field approaches to experimental design also decrease the researcher's control over extraneous variables.

Designing an Experiment. The specifics of how one designs a study depend on the goals of the research and the practical constraints placed on the design by the statistical tools needed to analyze it. In general, however, experiments are designed to test hypotheses as part of the theory building process. As shown in Figure 2, research design starts with the development of a tentative theory that is based on real-world observation. For example, from personal experience, the researcher may have observed that people with certain personality traits are more prone to help others than are people without those traits. Based on these observations, the researcher might form an empirically testable hypothesis concerning the relationship between personality traits and willingness to help others. To find out if this hypothesis is true, the researcher would then operationally define the various terms in the hypothesis (i.e., personality traits, helping others). These definitions would include ways to measure the personality traits (e.g., through an existing personality test, the development of a new test, or ratings of friends) and specific criteria for what constitutes "helping others." The researcher would then conduct the experiment, statistically analyze the resulting data using inferential statistics, and—based on the statistical significance of the answer—determine whether it was likely for those possessing the studied personality trait to help others.

In many cases, behavioral research can be more complicated than research in the physical sciences because of the complex nature of the real-world situation and the difficulty in operationally defining and measuring mental constructs. In addition, behavioral research needs to take into account ethical considerations for the treatment of research subjects. Although a compound in a chemistry experiment does not care if it is immersed in boiling water or flash frozen, a human subject would certainly object.

In scientific research, ethics refers to a code of moral conduct regarding the treatment of research subjects that is subscribed to by the members of a professional community. Care must be taken in the design and conduct of experiments so that research subjects are not harmed.

Applications

The Stanford Prison Experiment. The ethical treatment of prisoners has been of concern for as long as there have been prisoners. In the aftermath of World War II, for example, the Fourth Geneva Convention specified standards for the ethical treatment of prisoners of war. During the Iraq and Afghan Wars, the ethical treatment of prisoners made the headlines with controversies over the treatment of prisoners at Abu Ghraib and the use of torture to gain information from prisoners in the War on Terror. One of the landmark studies on the treatment of prisoners was performed at Stanford University by a team lead by Philip Zimbardo in 1971. This study is an outstanding example of how behavioral research can be conducted in the real world through simulation, as well as of some of the practical and ethical problems that can be encountered when performing such experiments.

Zimbardo advertised in a local newspaper for male volunteers to participate in a study of the psychological effects of prison life. The applicants were given diagnostic interviews and personality tests to eliminate candidates with psychological problems, medical problems, or a history of drug or alcohol abuse. The application process reduced the number of applicants from seventy to twenty-four. The resultant group was considered to be healthy, intelligent, middle-class individuals who tested within normal parameters on all measures administered. These individuals were randomly assigned to one of two groups: prison guards and prisoners. The experiment began with nine guards and nine prisoners. Guards worked in teams of three for eight-hour shifts, providing around the clock coverage of the "jail." Nine subjects were assigned to the prisoner group. The remaining subjects were placed on call in case they were needed. At the beginning of the experiment, there was no difference on any discernible factor between the individuals in the two groups. All subjects were required to sign an informed consent agreement, which told the subjects to expect some harassment, violation of privacy and other civil rights, and a minimally adequate diet during the course of the two-week experiment.

Without further notice to the subjects of the experiment, the local police—who were confederates in the study—went to the homes of those who had been assigned to the prisoner group and arrested them. Subjects were subjected to the full range of procedures associated with an arrest, including being charged, read their Miranda rights, spread-eagled on the police cruiser, searched, and handcuffed, often in front of their neighbors. The subjects were taken to the local police station, formally booked, read their rights again, and fingerprinted. Subjects were then left blindfolded in a holding cell. The prisoners (still blindfolded) were then transported to the "Stanford County Jail."

The jail simulation had been designed based on the inputs of several consultants, including one who had been incarcerated for nearly seventeen years. A corridor in the basement of the Stanford University building was boarded up at each end and some of the doors to laboratory rooms were replaced with doors with steel bars and cell numbers. A closet approximately two feet square (large enough for a prisoner to stand but not to sit) was transformed into a cell for solitary confinement. No windows or clocks were included in the simulation environment that would allow prisoners to note the passage of time. Listening devices were concealed in the cells to record the prisoners' conversations and a hole in the wall at one end of the corridor allowed videotaping of activities in the hall.

Upon arrival at the "jail," the "prisoners" were lectured by the "warden" (a confederate of the experimenters) about the fact that they would be receiving different treatment as prisoners. They were then systematically searched and stripped naked and subjected to a spray for the purported purpose of delousing. These were some of the procedures to which the prisoners were subjected in order to humiliate them as is often done in real prison situations. Subjects then were given a smock with their prison identification number on the front and back as their sole piece of clothing. A chain was also padlocked around the foot of each prisoner. Many of these activities in the simulated prison (e.g., the chain around the foot) are not done in real prisons but were included in the simulation in order to increase the feeling of helplessness and oppression. Prisoners spent most of their time in the small cells that were barely large enough to hold three cots each. Subjects randomly assigned to the "guards" group were not given any training on how to treat the prisoners. Within broad limits, they were allowed to do what they thought appropriate to maintain order and discipline in the "jail." Just as the prisoners were informed of the differences between being in jail and living in the outside world, the guards were warned of the dangerousness of their job. The guards made up their own rules for the running of the jail under the supervision of the warden. Guards were also given uniforms and wore mirrored sunglasses. They carried Billy clubs and whistles.

Following the arrests and indoctrination, subjects began the roles of prison guards and prisoners. At first, the subjects did not take the assigned roles very seriously: They were conscious of the fact that this was a simulation and that they were merely playing roles. At 2:30 a.m. the first morning, prisoners were awakened by blasting whistles and were required to gather in the hall to be counted. This was the first in a series of confrontations between guards and prisoners that resulted in increasingly polarized roles between the two groups. By the second day, the prisoners rebelled. The on-call guards were called in and the prisoners were forced back, stripped, and deprived of what little comfort they had. This began an increasing escalation of harassment and intimidation against the prisoners. After this incident, the guards developed a series of psychological tactics to maintain order and discipline among the prisoners. The guards' behavior was often arbitrary and humiliating to the prisoners. Within thirty-six hours of the start of the experiment, one of the prisoners began to "act crazy" and eventually was released from participation in the study. A little later in the experiment, another prisoner, who had been feeling ill and had refused to eat, broke down and became hysterical while talking to a priest (who had been brought in as a confederate) and the principal experimenter (in the role of prison "superintendent"). The researcher eventually removed the prisoner's chain and cap and reminded him that this was merely an experiment and not a real prison. The subject stopped crying and asked to go back to his cell to prove he was not a bad prisoner.

Eventually, the guards each developed one of three types of coping mechanisms. One group was tough but fair and followed the prison rules. The second group comprised "good guys" who did not punish the prisoners and would do small favors for them. The final group of guards was hostile, arbitrary, and very inventive in developing new ways to humiliate prisoners. Such behaviors, however, were not predicted by the initial personality tests that had been administered to each of the subjects.

Prisoners coped with their situation and concomitant feelings of powerless and frustration in a number of ways. At first, some of the prisoners attempted to rebel against the authority of the guards. Four of the prisoners broke down emotionally. One prisoner developed a psychosomatic rash over his entire body after learning that the "parole board" had denied his request for parole. Other subjects tried to be model prisoners and do everything required of them by the guards. However, by the end of the experiment, the prisoners had disintegrated both individually and as a group, and the guards were in total control of the prison.

Although the experiment was originally planned to last two weeks, the changes in behavior in the participants and the negative impact of the study made the researchers stop the study after only six days. By that time, the "prisoners" had become withdrawn and depressed and had started to behave in pathological ways. Some of the guards had started to behave sadistically, and none of the other guards attempted to intervene. The study was terminated after late-night videotapes showed the guards engaging in degrading and pornographic abuse of the prisoners when they thought that the experimenters were not watching. Another precipitating event was when a behavioral scientist brought in to interview the prisoners objected strongly to the behaviors she observed. However, she was the only one who ever questioned the morality of the experiment.

Conclusion

Experimental research is one of the foundations of science. In the simplest experimental design, a stimulus (independent variable) is presented to the research subjects and a response (dependent variable) is observed and recorded. Behavioral scientists also need to be concerned with the control of extraneous variables that can affect the outcome of the study without being directly related to the research question and intervening variables that occur after the manipulation of the independent variable but before the measurement of the dependent variable. As much as possible, these need to be controlled or eliminated through the experimental design. Laboratory experiments allow researchers great control over the variables in a study but have the drawback of not realistically emulating the real-world situation. Simulations and field experiments allow the design of an increasingly realistic experiment but also give the experimenter decreased control over the experimental situation. In addition, when working with human subjects, care must be taken in the design and conduct of experiments so that research subjects are not harmed.

Terms & Concepts

Confederate: A person who assists a researcher by pretending to be part of the experimental situation while actually only playing a rehearsed part meant to stimulate a response from the research subject.

Control Group: A subset of participants in an experiment that does not receive the experimental condition (i.e., does not experience the manipulation of the independent variable). Control groups help researchers determine whether the observed results of a research study were due to the manipulation of the independent variable or some other factor.

Data: (sing. datum). In statistics, data are quantifiable observations or measurements that are used as the basis of scientific research.

Ethics: In scientific research, a code of moral conduct regarding the treatment of research subjects that is subscribed to by the members of a professional community. Many professional groups have a specific written code of ethics that sets standards and principles for professional conduct and the treatment of research subjects.

Experiment: A situation under the control of a researcher in which an experimental condition (independent variable) is manipulated and the effect on the experimental subject (dependent variable) is measured. Most experiments are designed using the principles of the scientific method and are analyzed to determined whether or not the results are statistically significant.

Hypothesis: An empirically testable declaration that certain variables and their corresponding measure are related in a specific way proposed by a theory. A null hypothesis (H0) is a statement that the findings of the experiment will show no statistical difference between the current condition (control condition) and the experimental condition.

Inferential Statistics: A subset of mathematical statistics used in the analysis and interpretation of data. Inferential statistics are used to make inferences such as drawing conclusions about a population from a sample and in decision making.

Operational Definition: A definition that is stated in terms that can be observed and measured.

Statistical Significance: The degree to which an observed outcome is unlikely to have occurred due to chance.

Subject: A participant in a research study or experiment whose responses are observed, recorded, and analyzed.

Variable: An object in a research study that can have more than one value. Independent variables are stimuli that are manipulated in order to determine their effect on the dependent variables (response). Extraneous variables are variables that affect the response but that are not related to the question under investigation in the study. Intervening variables occur between the manipulation of the independent variable and the measurement of the dependent variable.

Bibliography

Ameri, M., Schur, L., Adya, M., Bentley, F. S., McKay, P., & Kruse, D. (2018). The disability employment puzzle: A field experiment on employer hiring behavior. ILR Review, 71(2), 329–364. Retrieved October 23, 2018, from EBSCO Online Database Sociology Source Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=sxi&AN=127895962&site=ehost-live&scope=site

Black, K. (2006). Business statistics for contemporary decision making (4th ed.). New York, NY: John Wiley & Sons.

Keuschnigg, M., & Wolbring, T. (2016). The use of field experiments to study mechanisms of discrimination. Analyse & Kritik, 38(1), 179–202. Retrieved October 23, 2018, from EBSCO Online Database Sociology Source Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=sxi&AN=114787797&site=ehost-live&scope=site

Mikėnė, S., Gaižauskaitė, I., & Valavičienė, N. (2013). Qualitative interviewing: Field-work realities. Socialinis Darbas, 12, 49–61. Retrieved November 5, 2013 from EBSCO online database SocINDEX with Full Text. http://search.ebscohost.com/login.aspx?direct=true&db=snh&AN=88008107&site=ehost-live&scope=site

The Stanford Prison Experiment . Retrieved 18 March 2002 from: www.prisonexp.org.

Wendt, O., & Miller, B. (2012). Quality appraisal of single-subject experimental designs: An overview and comparison of different appraisal tools. Education & Treatment of Children (West Virginia University Press), 35, 235–268. Retrieved November 5, 2013 from EBSCO online database SocINDEX with Full Text. http://search.ebscohost.com/login.aspx?direct=true&db=snh&AN=74231752&site=ehost-live&scope=site

Witte, R. S. (1980). Statistics. New York, NY: Holt, Rinehart and Winston.

Suggested Reading

Calfee, R. C. (1975). Human experimental psychology. New York, NY: Holt, Rinehart and Winston.

Lê, Q., & Lê, T. (2013). Conducting research in a changing and challenging world. Hauppauge, NY: Nova Science Publishers.

Ruane, J. M. (2016). Introducing social research methods: Essentials for getting the edge. Malden, MA: Wiley Blackwell.

Weisburd, D., Morris, N. A., & Ready, J. (2008, Mar). Risk-focused policing at places: An experimental evaluation. Justice Quarterly, 25, 163–200. Retrieved April 2, 2008 from EBSCO Online Database SocINDEX with Full Text: http://search.ebscohost.com/login.aspx?direct=true&db=snh&AN=31334370&site=ehost-live&scope=site

Zimbardo, P. G., Maslach, C., & Haney, C. (2000). Reflections on the Stanford Prison Experiment: Genesis, transformations, consequences. In T. Blass (Ed.), Obedience to authority: Current perspectives on the Milgram paradigm (pp. 193–237). Mahwah, NJ: Erlbaum.