Discuss the importance of external and internal validity and some of the threats to each of these.
Validity means a measurement quality indicating the extent to which the measure replicates the underlying construct, to be precise, whether it measures what it claims to measure. Internal validity may be important in determining the accuracy of the results. Given that samples may not be selected randomly, the results may perhaps be inaccurate. External validity maybe important in determining the “generalizability” of the population findings (Bernard, 2000). For instance, if the selected sample is only African-Americans under the age of 35, then it may be difficult to generalize the results to the entire U.S. population.
Internal validity threats include:
- Testing: this relates to the potential effects of a pre-test on the participants performance in a study on the post-test. This may alert participants to the fact that they are being intentionally studied (Beyer-Westendorf and Bu¨ ller, 2011).
- Statistical regression: this refers to the affinity of extreme scores to regress (or move) in the direction of the mean score on succeeding re-testing. For instance, students scoring on an IQ test, below twenty five percent (lowest extreme) may be given a post-test. A high post-test score may be expected.
- Mortality: this refers to loss of subjects from a study owing to the initial non-availability or consequent withdrawal from the study For instance, most high-scoring people may drop-out from the “experimental group” than from the “control group”
External validity threats include:
- Experimenter effects: refers to the likelihood of the experimenter un-intentionally influencing the performance of participants in a study
- Multiple-treatment interference: this pertains to the situation in which subjects in a study may get more than one treatment. In this case, the results of the multiple-treatments may interrelate.
- Experimental arrangement Effects: this pertains to situations where subjects become conscious that they may be involved in a study, as a effect of that awareness, their performance or response may be different.
What is the impact of having one type of validity but not the other?
If you could only have one type of validity in your experiment, which would it be and why?
The two dissimilar characterizations of the connection involve the applicability and generalizability. Internal Validity relies much on applicability while external validity may be more on generalizability. In addition, with internal validity one may have more control than with external validity (Walwyn and Roberts, 2010). However, I would prefer to use external for the reason that: even through it may be harder to maintain the control; at least the entire facts may be present when you setting-up an experiment and one doesn't have to presuppose something that may lead to bias .
References
Bernard H. (2000). Social Research Methods. Thousands Oaks, Calif: Sage Publications. Print.
Beyer-Westendorf J., and Bu¨ ller H. (2011).External and internal validity of open label or
double-blind trials in oral anticoagulation: better, worse or just different? J Thromb
Haemost; 9: 2153–8.
Calder B., Phillips L., and Tybout A. (1983). The concept of external validity. J Consum Res.
10 (1).
Green L. (1977). Evaluation and measurement: some dilemmas for health education. Am J
Public Health. 1977; 67: 155–161. [PMC free article] [PubMed] Retrieved March 18, 2013; from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1653552/pdf/amjph00477-0045.pdf
Walwyn R., and Roberts C., (2010). Therapist variation within randomised trials of
psychotherapy: implications for precision, internal and external validity. Statistical
Methods in Medical Research 2010; 19: 291–315.
