Psychologists conduct researches to build knowledge, to test theories and to discover new phenomena in varied fields of specialization (Beins, 2004). Research is a complex process that starts with the identification of a problem which the researcher wishes to answer and to explore. The second step is to determine what kind of research design the problem would be best answered or demonstrated. In psychology, there are many research designs that could be employed by the researcher, however many prefer to use experiments.
Now, what is an experiment?
Take for example that one wants to determine children’s reactions to a lost kitten. If you just simply leave a kitten in the middle of the room and observe how children will respond to the kitten, does it mean that one has conducted an experiment? You might say yes, because you have tried to answer your question, but essentially, this example is not a true experiment. An experiment is characterized by the manipulation of variables in order to effect the change in behavior that the researcher wants to study (Shaughnessy, Zechmeister, & Zechmeister, 2000).
The factors present in the research environment that is not related to the behavior under study must be controlled or held constant in order to avoid the possibility that these factors might influence the variables in the study
The most basic element of an experiment is the variables of the study. An experiment must have an independent variable and a dependent variable (Shaughnessy, Zechmeister, & Zechmeister, 2000). The dependent variable is the response measure of the effect of the manipulation of the independent variable while the independent variable refers to the manipulation of the environment controlled by the experimenter. Hence, going back to our original example, the reaction of children is the dependent variable and the presence of the kitten is the independent variable. However, the independent variable must at least have two levels either qualitatively (cute, clean and fat kitten and a scrawny, smelly kitten) or quantitatively (1 and 2 kittens).
The most important thing to note is that two conditions must be compared with each other to determine if a change in behavior has occurred due to the independent variable. Such as the appearance or the number of kittens affects children’s reactions to the kitten. In simpler forms, an experiment can be designed wherein a group of participants will experience the manipulated independent variable while the control group does not experience it, hence any difference in behavior can be attributed to the manipulation of the independent variable.
The advantage of conducting an experimental research is that it offers a better way of controlling extraneous variables that can confound the results of the experiment. Another advantage is that experiments allow the researcher to establish causality between the independent and dependent variable, hence any behavioral change is caused by the manipulation of the conditions that affect the behavior (Huitt, 2003). Lastly, experiments have the advantage of economy, it allows the experimenter to conduct the research in a controlled environment and does not need to wait for the right conditions, and data gathering is done immediately as opposed to surveys and interviews.
However, one has to be aware that experiments are not always 100% fool proof nor it is always correct. It has been established that experiment subjects are influenced by the knowledge of the experiment’s purpose and or objectives or the expectations of the experimenter which is called the Hawthorne effect (Beins, 2004). Another pitfall of experiments is the effect of the experimenter on the subjects. Like when the experimenter exhibit behaviors that cue the subjects on what is expected of them such as an experimenter smiles when a child gives the correct response.
Beins, B. (2004). Research Methods: A Tool for Life. Boston: Allyn & Bacon.
Huitt, W. (2003, June). Assessment, measurement, evaluation, and research: Types of studies in
scientific research. Educational Psychology Interactive. Valdosta, GA: Valdosta State University. Retrieved December 13, 2006 from
Shaughnessy, J., Zechmeister, E. & Zechmeister, J. (2000). Research Methods in Psychology,
5th ed. Boston: McGraw Hill.
Discussion Topic 3
An important consideration that any experimenter must think about is whether experimental conditions are ethically sound. It is a reality that some of the things we want to study are ethically impossible (Boniface, 1995), for example, we want to measure the group behavior of college students who use drugs, it is practically and ethically impossible to form a group of college students and have them take drugs and then compare their behavior to a group that does not use drugs. Another example is when you want to study the coping behavior of survivors of a plane crash, you cannot have subjects experience an actual crash and leave them stranded and then observe their behavior. Although information from these kinds of studies is valuable it is not possible to conduct actual or true experiments with the said variables. On the other hand, one can conduct quasi-experiments.
Quasi-experimental design is a research design wherein the independent variables like age, sex, status, experience, intelligence etc. are tested in relation to dependent variables (Dawson, 1997). Like when one wants to compare the behavior of male and female drug users after rehabilitation, or to compare the reading comprehension of high IQ and low IQ students. Quasi-experiment is different from an experimental design in the sense that the independent variable in the study is usually a personal attribute or characteristic of the subject which occurs through nature or time (Boniface, 1995).
Quasi-experiments are popular in psychological researches because it offers the observational and correlational procedures of research and combine this with experimentation (Breakwell, Hammond & Fife-Schaw, 1995). The need for specific independent variables such as subject characteristics means that one can select the subject instead of varying the independent variables; hence if you want to compare high IQ and low IQ you just need to identify subjects that fall within these two conditions.
However, quasi-experiments also carries with it elements that may affect the validity of the research results, for example, because the researcher has to take only subjects that fulfill the criteria of the research problem, then the possibility that extraneous variables like religion, race, motivation and personality can confound the effect of the variables (Dawson, 1997). Hence, difference in comprehension between high and low IQ can also be affected by other factors like kind of education, exposure to reading materials and even motivational attitudes of the subjects.
Quasi-experiments can be conducted in research studies that would like to measure subject variables in relation to specific behaviors (Dawson, 1997) such as reading ability of male and females, play behavior of early childhood and late childhood, work satisfaction of managers and rank and file, problem behavior of white and African American juvenile delinquents, superstitious beliefs of Catholics and Protestants, buying behavior of the rich and poor, kinds of jobs of college graduates and drop-outs and others. From the said examples, it is obvious that quasi-experimental design is best applicable in situations wherein one wants to know the behavior of certain types or classes of individuals in specific aspects or situations rather than identifying what factors or situations affect their behavior.
Dawson, T. (1997). A primer on experimental and quasi-experimental design. Paper presented at
the annual meeting of the Southwest Educational Research Association, Austin, January. Retrieved December 13, 2006, from http://ericae.net/ft/tamu/Expdes.HTM
Boniface, D. (1995). Experimental design and statistical methods: For behavioural and social research. London: Chapman & Hall.
Breakwell, G., Hammond, S., & Fife-Schaw, C. (1995). Research methods in psychology. London: Sage.