Of the four types of validity (see also internal validity, construct validity and external validity) conclusion validity is undoubtedly the least considered and most misunderstood. That's probably due to the fact that it was originally labeled 'statistical' conclusion validity and you know how even the mere mention of the word statistics will scare off most of the human race!
In many ways, conclusion validity is the most important of the four validity types because it is relevant whenever we are trying to decide if there is a relationship in our observations (and that's one of the most basic aspects of any analysis). Perhaps we should start with an attempt at a definition:
Conclusion validity is the degree to which conclusions we reach about relationships in our data are reasonable.
For instance, if we're doing a study that looks at the relationship between socioeconomic status (SES) and attitudes about capital punishment, we eventually want to reach some conclusion. Based on our data, we may conclude that there is a positive relationship, that persons with higher SES tend to have a more positive view of capital punishment while those with lower SES tend to be more opposed. Conclusion validity is the degree to which the conclusion we reach is credible or believable.
Although conclusion validity was originally thought to be a statistical inference issue, it has become more apparent that it is also relevant in qualitative research. For example, in an observational field study of homeless adolescents the researcher might, on the basis of field notes, see a pattern that suggests that teenagers on the street who use drugs are more likely to be involved in more complex social networks and to interact with a more varied group of people. Although this conclusion or inference may be based entirely on impressionistic data, we can ask whether it has conclusion validity, that is, whether it is a reasonable conclusion about a relationship in our observations.
Whenever you investigate a relationship, you essentially have two possible conclusions -- either there is a relationship in your data or there isn't. In either case, however, you could be wrong in your conclusion. You might conclude that there is a relationship when in fact there is not, or you might infer that there isn't a relationship when in fact there is (but you didn't detect it!). So, we have to consider all of these possibilities when we talk about conclusion validity.
It's important to realize that conclusion validity is an issue whenever you conclude there is a relationship, even when the relationship is between some program (or treatment) and some outcome. In other words, conclusion validity also pertains to causal relationships. How do we distinguish it from internal validity which is also involved with causal relationships? Conclusion validity is only concerned with whether there is a relationship. For instance, in a program evaluation, we might conclude that there is a positive relationship between our educational program and achievement test scores -- students in the program get higher scores and students not in the program get lower ones. Conclusion validity is essentially whether that relationship is a reasonable one or not, given the data. But it is possible that we will conclude that, while there is a relationship between the program and outcome , the program didn't cause the outcome. Perhaps some other factor, and not our program, was responsible for the outcome in this study. For instance, the observed differences in the outcome could be due to the fact that the program group was smarter than the comparison group to begin with. Our observed posttest differences between these groups could be due to this initial difference and not be the result of our program. This issue -- the possibility that some other factor than our program caused the outcome -- is what internal validity is all about. So, it is possible that in a study we can conclude that our program and outcome are related (conclusion validity) and also conclude that the outcome was caused by some factor other than the program (i.e., we don't have internal validity).
We'll begin this discussion by considering the major threats to conclusion validity, the different reasons you might be wrong in concluding that there is or isn't a relationship. You'll see that there are several key reasons why reaching conclusions about relationships is so difficult. One major problem is that it is often hard to see a relationship because our measures or observations have low reliability -- they are too weak relative to all of the 'noise' in the environment. Another issue is that the relationship we are looking for may be a weak one and seeing it is a bit like looking for a needle in the haystack. Sometimes the problem is that we just didn't collect enough information to see the relationship even if it is there. All of these problems are related to the idea of statistical power and so we'll spend some time trying to understand what 'power' is in this context. One of the most interesting introductions to the idea of statistical power is given in the 'OJ' Page which was created by Rob Becker to illustrate how the decision a jury has to reach (guilty vs. not guilty) is similar to the decision a researcher makes when assessing a relationship. The OJ Page uses the infamous OJ Simpson murder trial to introduce the idea of statistical power and illustrate how manipulating various factors (e.g., the amount of evidence, the "effect size", and the level of risk) affects the validity of the verdict. Finally, we need to recognize that we have some control over our ability to detect relationships, and we'll conclude with some suggestions for improving conclusion validity.
Copyright ©2006, William M.K. Trochim, All Rights Reserved
Purchase a printed copy of the Research Methods Knowledge Base
Last Revised: 10/20/2006