Threats to Internal and External Validity - Essay Example

THREATS TO THE VALIDITY OF AN EXPERIMENT Firstly, we begin by clarifying what we mean by threats to validity. Maxfield and Babble (2005) put it in very simple terms: these are, according to them, “possible ways that a researcher might be wrong in inferring causation. (page 191)” Burns (1997:1) defines validity as, “ the best available approximation to the truth of a given proposition, inference or conclusion. In other words, when we make some claim, does the evidence support our conclusions.

”  Factors affecting internal validity ask whether or not the research conditions warrant the conclusions. On the other hand, factors affecting external validity ask whether or not the conclusions of the research are replicable over diverse populations. Here are some examples of threats to validity of an experiment. Threats to Internal Validity The first one is history. Events may occur during the course of the experiment that affect the experimental results. This contemplates an outside or external event that impacts on the dependent variable.

Campbell defines it as “the specific events occurring between the first and second measurement in addition to the experimental variable.” (Campbell and Stanley, 5). If Dr. Williams wants to ensure that there is no threat pertaining to history affecting the outcomes, then there must be nothing coming in between the first and second measurement, ie., an unforeseeable humanitarian crisis that would congest the hospitals. A second example is maturation. People are continually growing and changing, whether in an experiment or not, and these changes affect the result of the experiment.

To distinguish from history, maturation is the processes within the subjects, not in their external environments. Campbell and Stanley (5) gave some examples: “growing older, growing hungrier, growing more tired, and the like.” A possible threat is that since the doctors know that they are being tested, they might treat their patients in a manner different from how they normally treated previous patients, always being on alert that they are being tested and the person before them is an actor.

For example, they might be kinder, more patient, more tolerant than they normally are – especially when they see that the patient is of a different race. A third example is differential selection, which takes place when the samples are not identical, and the research outcomes are different, not for any reason pertaining to the object of the experiment (for example, not because of race bias) but because of exposure to different interventions and conditions. This might be a big threat in Dr. William’s research.

Situations and conditions confronting doctors on a daily basis change and are highly variant. Stress, number of other patients, work load – these are things that vary from doctor to doctor and affect their behavior. Also, the doctors might be of different ages. Younger people tend to be more open-minded, and older people, more predisposed to racist thinking. Threats to External Validity A first example would be generalization to other kinds of people. This asks the question: Can the study’s results be generalized to groups not included in the design: e.g., to people of a different sex, age, race, social group, culture, educational background, socio-economic class, etc.? It might be difficult to generalize that there is no racial bias in health care (if indeed that is Dr.

William’s intent) on the basis of doctors, because doctors tend to be more educated and exposed, and there are numerous findings that suggest that those who are more educated and more exposed to other settings, people and world-views are less racist. A positive finding that doctors do not discriminate against patients of another race may not translate to a race-blind health care system. Another example is generalization to other kinds of setting. This asks the question: Do the effects observed generalize to settings not examined in the experiment?

Which state is the experiment to be performed in? Does Doctor Williams intend to make conclusions for the whole country on the basis of that one state? An experiment on race and race perceptions and behavior set in New York will yield results that are going to be different from results that an experiment performed in Kansas or Georgia would yield. These are things that the Doctor must take into consideration as well – attitudes on race are highly variant over space. Lastly, we go to generalization to other times.

This asks the question: Does the effect of a particular treatment generalize to other times? It is hoped that Dr. Williams does not intend this experiment to be the basis of a conclusion that racial attitudes of doctors were like this in the past, or will be in a certain way in the future. The experiment should only measure the conditions for this particular time period. References Burns, P. (1997). “Securing Internal Validity – How do we avoid the threats?” Internet. Available at http://www.

socialresearchmethods.net/tutorial/Burns/int.html Campbell, D. and Stanley, J. (1963). Experimental and Quasi-Experimental Designs for Research. Boston: Houghton Mifflin Company. Maxfield, M. and Babbie, E. (2005). Basics of Research Methods for Criminal Justice and Criminology. Belmont, CA: Wadsworth.