Introduction
Internal validity is a concept in research design that refers to the extent to which a study can support a causal relationship between two variables. It is one of the most important concepts in research, as it helps to ensure that the results of a study are reliable and can be generalized to other settings.
History
The concept of internal validity was first introduced by Donald Campbell and Julian Stanley in their book, “Experimental and Quasi-Experimental Designs for Research.” Campbell and Stanley defined internal validity as “the approximate validity with which inferences can be drawn about causal relations from a comparison of the control group and experimental group.”
Characteristics
Internal validity is related to a number of other concepts in research design, including:
- External validity: External validity refers to the extent to which the results of a study can be generalized to other settings.
- Construct validity: Construct validity refers to the extent to which a study measures what it is intended to measure.
- Statistical conclusion validity: Statistical conclusion validity refers to the accuracy of the statistical inferences made in a study.
Threats to Internal Validity
There are a number of threats to internal validity, which can undermine the ability of a study to support a causal relationship. Some of the most common threats to internal validity include:
- History: Events that occur outside of the study can affect the outcome of the study. For example, if a study is conducted to test the effectiveness of a new drug, and there is a major news story about the drug’s side effects during the study period, this could bias the results of the study.
- Maturation: Participants in a study may change over time, even if they are not exposed to the independent variable. For example, if a study is conducted to test the effectiveness of a new educational program, and the participants in the study are all students who are about to graduate from high school, the results of the study may be affected by the students’ natural maturation process.
- Testing: Participants in a study may learn from the first test and perform better on the second test, even if they are not exposed to the independent variable. For example, if a study is conducted to test the effectiveness of a new study technique, and the participants are all students who take a pre-test and a post-test, the results of the study may be affected by the students’ familiarity with the test format.
- Instrumentation: Changes in the way that a variable is measured can affect the results of a study. For example, if a study is conducted to test the effectiveness of a new drug, and the participants’ blood pressure is measured with a different blood pressure cuff at the beginning and end of the study, the results of the study may be affected by the difference in the blood pressure cuffs.
- Selection bias: Participants in a study may be selected in a way that favors one group over another. For example, if a study is conducted to test the effectiveness of a new educational program, and the participants are all students who were selected by their teachers, the results of the study may be biased in favor of students who are already doing well in school.
- Regression to the mean: Participants who score very high or very low on a pre-test are more likely to score closer to the mean on a post-test, even if they are not exposed to the independent variable. For example, if a study is conducted to test the effectiveness of a new tutoring program, and the participants are all students who scored very low on a pre-test, the results of the study may be affected by regression to the mean.
- Attrition: Participants may drop out of a study for a variety of reasons, and this can affect the results of the study. For example, if a study is conducted to test the effectiveness of a new weight loss program, and participants who lose a lot of weight are more likely to drop out of the study, the results of the study may be biased in favor of participants who do not lose much weight.
Strategies for Increasing Internal Validity
Researchers can take a number of steps to increase the internal validity of their studies. Some of the most common strategies for increasing internal validity include:
- Random assignment: Participants should be randomly assigned to the experimental and control groups. This helps to ensure that the groups are equivalent at the beginning of the study.
- Blinding: The participants, the researchers, and the data collectors should be blinded to the group assignment. This helps to reduce the risk of bias.
- Placebo control: A placebo control group should be included in the study. This helps to control for the placebo effect.
- Multiple measures: Multiple measures of the dependent variable should be collected. This helps to reduce the risk of measurement error.
- Longitudinal design: A longitudinal design should be used. This allows the researchers to track changes in the participants over time.
Internal validity occurs when a researcher controls all extraneous variables and the only variable influencing the results of a study is the one being manipulated by the researcher. This means that the variable the researcher intended to study is indeed the one affecting the results and not some other, unwanted variables. (Take a look at the definition for “confound”) There are several “threats to internal validity” including: history, maturation of participants, testing, instrument decay, and statistical regression.