External and Internal Validity

there are many forms of validity and research in general validity refers to how sound your research design and method are we are going to talk about two types of validity internal validity and external validity external validity refers to how well the findings of our study can be generalized to other populations and settings after all the purpose of research is to learn more about the world at large and not just about our specific group of participants or what goes on in our lab alone it would be great if we could just study everyone in all possible contexts but unfortunately our access is typically limited to a small population of participants in a single context or situation however we still want to be able to take our findings and say with a degree of confidence that they represent what goes on outside of our study that's what external validity gives us a lack of external validity means that our findings cannot be applied to other people or other contexts outside of the study we performed or that application to other people and contexts is limited for example if we only had white males participate in our study our findings might not apply to females or non-white races and ethnicities so while external validity refers to factors outside of our study internal validity on the other hand relates to the validity of measures within our study it depends on factors such as did we capture the construct that we intended to with our measure was the independent variable really the cause of the changes in the dependent variable or did we control for all other possible factors that could have affected our results a lack of internal liddie means that our findings may not necessarily reflect what we think they do for example our study may have looked at the effect of pet ownership on mental health but let's say that unbeknownst to us at the time the type of pet has a significant effect on whether or not the owner will experience mental health benefits if we did not control for the type of pet in our study that our findings do not necessarily reflect what we think they do and our study lacks internal validity let's review some common threats to external and internal validity this is by no means an exhaustive list rather here we just have three examples of threats to external validity remember anything that reduces the generalizability of the study results to other populations and settings will reduce external validity let's take a look at participant characteristics the question we want to ask is do they represent the larger population we want to generalize to for example a large majority of studies in psychology recruit undergraduates looking to get extra credit in their courses however these participants may differ from the larger population in a number of factors such as age IQ and socioeconomic status setting the research environment is typically a very controlled and artificial one very much unlike the everyday world that we want to generalize our findings to it is important for studies to be able to control all possible extraneous variables to be sure of the relationship between the independent and dependent variables but this does come at a cost of external validity to a world in which the variables of interest will almost always occur in the presence of other extraneous variables and timing the world changes with time and so studies conducted in the past may not have external validity to the world today for example a study conducted sixty years ago on sexism in the workplace may have produced very different results than a similar study conducted today and what about threats to internal validity we'll go over a few more here since internal validity can involve such a wide variety of factors within a study so the thing to remember here is that anything that reduces the studies ability to establish a relationship between the dependent and independent variable reduces its internal validity so starting with selection bias selection bias can be a problem for the same reasons that participant characteristics can be a problem for external validity if we are comparing two groups of participants we want to make sure the two groups are as similar to each other as possible at the beginning of the study so that differences in the characteristics of participants don't affect our results for example if one group consists of college undergrads volunteering for extra credit and the other group was recruited from flyers at a local arcade we may find that the two groups differ in their level of education motivation and possibly IQ not saying that those who hang out at the arcade have low IQs just that there may be a larger proportion of individuals with higher IQs in a group that is pursuing an education and putting forth the outside effort to earn extra credit next we have sample size having too few participants in the study can mean that the statistics don't have enough power to produce results we can trust in other words our findings are more likely to be the result of chance if we are basing them on only a few individuals confounding this occurs when an extraneous variable we did not control for has an effect on our results in other words something other than the independent variable is affecting the dependent variable in history history refers to events outside of the study or between a pretest and a post-test having an effect on how participants respond this could be something as significant as a natural disaster or as simple as a participant reading a book on memorization skills during a study on memory maturation maturation is a natural process in which participants change between a pretest or post-test or during the course of a study in which multiple measurements are made it may simply be that they got older in a study that spans years or that they became fatigued or hungry during testing maturation is similar to history and that both affect how a participant responds in a study but they differ in that maturation is internal and is a natural course of things while history has to do with an external event of some kind instrumentation this occurs when the instrument or measure we are using in a study changes in some way over the course of a study if the measure is different at the beginning than it is at the end it could possibly be producing different results instrumentation can occur for example if an experimenter decides to tweak a test to make it better or if the springs on a scale start to wear out it can also occur in observational measures if the observer gets better with practice at detecting certain behaviors over the course of a study next we have testing also sometimes referred to as practice effects in studies that use the same or similar measure for a pretest and a post-test run the risk of practice effects if a participant is given a math test for a pretest and a post-test she is likely to perform better the second time that she has seen the test before it could be a matter of the participant getting better with repeated practice or it could simply be familiarity with a test that leads to better performance and finally we have attrition or mortality attrition refers to participants dropping out during the course of a study not only does this become a problem for sample size but there may also very likely be a reason that some participants drop out and others continue on with the study for example participants who drop out may have a more difficult time with transportation to and from the location of a study and therefore may be of lower socioeconomic Attis this factor that differs between those who stick around and those who don't may have an impact on the sorts of findings we end up with if the study involves a pretest and a post-test the group who took the pretest at the beginning of the study may differ from the group who took the post-test at the end also the factor determining which participants are still in the study may be exerting a sort of selection bias on the characteristics of the participants our results are based on which is a threat to the internal validity of our results but this can also create a threat to external validity if the participant sample our results are based on differs in a significant way from the general population with careful consideration of the potential threats to internal and external validity when designing your study you should be confident that your results reflect a true relationship between your dependent and independent variables and that those results are generalizable to most other populations and contexts

Channel: Alyson Froehlich

Views: 44,014

Rating: 4.884058 out of 5

Keywords:

Id: V3StgmV1LsM

Length: 11min 55sec (715 seconds)

Published: Tue Sep 22 2015