Internal Validity In Research
Neil J Salkind defines internal validity as, “the quality of an experimental design such that the results obtained are attributed to the manipulation of the independent variable. In other words, if what you see is a function of what you did, then the experiment has internal validity.” (p. 235). When conducting research with experimental designs, researchers usually have two groups to test their experiment. These two groups are called “the experimental group” and the control group”. Experimental designs can provide a tremendous amount of power and control over understanding relationships between variables. This essay will how the presence of a control group accounts for threats to internal validity and summarize the types of threats to internal
…show more content…
If pretests are administered the researchers might not be able to be completely confident that the treatment given during the experiment had an effect on the results or if the pretest made the participants “test-wise”. For example, a researcher pretests a group of math students, and then teaches them how to solve the problems. After the post test is given and the results of the test show an increase in correct answers, it may be difficult to confirm the students understand what they were taught or they became familiar with the way the test was administered. To avoid this threat, researchers can use a control group that is given the pretest, treatment, and post-test, then use a different group and only administer the treatment and post-test. After the results are compiled, the researchers can analyses the results and compare the two groups. Mortality is another threat to internal validity that can be avoided. Mortality is a threat to the experiment when subjects in the experiment drop-out and it changes the nature of the group. By using random assignments, researchers can ensure that the drop-outs from both the control group and non-control group do not affect the end result. Maturation is another threat to internal validity much like mortality. As the subjects in the experiment mature over the duration of the project, they can affect the results of the experiment. For example, young children can mature tremendously in the time frame of a year. So this maturation can affect the end results of the experiment.
Randomization to control and experimental groups can help researchers avoid threats such as testing, mortality, and maturation. If the researcher continually uses random control and experimental groups they will decrease the threats to the internal validity of the project. When researchers control all the variables they can possibly control in an experiment
If pretests are administered the researchers might not be able to be completely confident that the treatment given during the experiment had an effect on the results or if the pretest made the participants “test-wise”. For example, a researcher pretests a group of math students, and then teaches them how to solve the problems. After the post test is given and the results of the test show an increase in correct answers, it may be difficult to confirm the students understand what they were taught or they became familiar with the way the test was administered. To avoid this threat, researchers can use a control group that is given the pretest, treatment, and post-test, then use a different group and only administer the treatment and post-test. After the results are compiled, the researchers can analyses the results and compare the two groups. Mortality is another threat to internal validity that can be avoided. Mortality is a threat to the experiment when subjects in the experiment drop-out and it changes the nature of the group. By using random assignments, researchers can ensure that the drop-outs from both the control group and non-control group do not affect the end result. Maturation is another threat to internal validity much like mortality. As the subjects in the experiment mature over the duration of the project, they can affect the results of the experiment. For example, young children can mature tremendously in the time frame of a year. So this maturation can affect the end results of the experiment.
Randomization to control and experimental groups can help researchers avoid threats such as testing, mortality, and maturation. If the researcher continually uses random control and experimental groups they will decrease the threats to the internal validity of the project. When researchers control all the variables they can possibly control in an experiment