Discussion: Designing Quantitative Research
Researchers consider validity and reliability with each new study they design. This is because validity and reliability are not fixed but rather reflect a particular study’s unique variables, research design, instruments, and participants.
In the context of research design, two types of validity, which speak to the quality of different features of the research process, are considered: internal validity and external validity. Assuming that the findings of a research study are internally valid—i.e., the researcher has used controls to determine that the outcome is indeed due to manipulation of the independent variable or the treatment—external validity refers to the extent to which the findings can be generalized from the sample to the population or to other settings and groups. Reliability refers to the replicability of the findings.
For this Discussion, you will consider threats to internal and external validity in quantitative research and the strategies used to mitigate these threats. You will also consider the ethical implications of designing quantitative research.
With these thoughts in mind:
An explanation of a threat to internal validity and a threat to external validity in quantitative research. Next, explain a strategy to mitigate each of these threats. Then, identify a potential ethical issue in quantitative research and explain how it might influence design decisions. Finally, explain what it means for a research topic to be amenable to scientific study using a quantitative approach.
Be sure to support your Main Issue Post and Response Post with reference to the week’s Learning Resources and other scholarly evidence in APA Style.
Learning Resources Required Readings
Babbie, E. (2017) Basics of social research (7th ed.). Boston, MA: Cengage Learning.
· Chapter 3, “The Ethics and Politics of Social Research”
Burkholder, G. J., Cox, K. A., & Crawford, L. M. (2016). The scholar-practitioner’s guide to research design. Baltimore, MD: Laureate Publishing.
· Chapter 7, “Quality Considerations”
Document: Threats to Internal Validity (PDF)
Threats to Internal Validity
(Shadish, Cook, & Campbell, 2002)
1. Ambiguous temporal precedence. Based on the design, unable to determine with certainty which variable occurred first or which variable caused the other. Thus, unable to conclude with certainty cause-effect relationship. Correlation of two variables does not prove causation.
2. Selection. The procedures for selecting participants (e.g., self-selection or researcher sampling and assignment procedures) result in systematic differences across conditions (e.g., experimental-control). Thus, unable to conclude with certainty that the “intervention” caused the effect; could be due to way in which participants are selected.
3. History. Other events occur during the course of treatment that can interfere with treatment effects and could account for outcomes. Thus, unable to conclude with certainty that the “intervention” caused the effect; could be due to some other event to which the participants were exposed.
4. Maturation. Natural changes that participants experience (e.g., grow older, get tired) during the course of the intervention could account for the outcomes. Thus, unable to conclude with certainty that the “intervention” caused the effect; could be due to the natural change/maturation of the participants.
5. Regression artifacts. Participants who are at extreme ends of the measure (score higher or lower than average) are likely to “regress” toward the mean (scores get lower or higher, respectively) on other measures or retest on same measure. Thus, regression can be confused with treatment effect.
6. Attrition (mortality). Refers to dropout or failure to complete the treatment/study activities. If differential dropout across groups (e.g., experimental-control) occurs, could confound the results. Thus, effects may be due to dropout rather than treatment.
7. Testing. Experience with test/measure influences scores on retest. For example, familiarity with testing procedures, practice effects, or reactivity can influence subsequent performance on the same test.
8. Instrumentation. The measure changes over time (e.g., from pretest to posttest), thus making it difficult to determine if effects or outcomes are due to instrument vs. treatment. For example, observers change definitions of behaviors they are tracking, or the researcher alters administration of test items from pretest to posttest.
9. Additive and interactive effects of threats to validity. Single threats interact, such that the occurrence of multiple threats has an additive effect. For example, selection can interact with history, maturation, or instrumentation.
Reference
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasiexperimental designs for generalized causal inference. Boston, MA: Houghton-Mifflin.
Download the “Litmus Test” document.
Required Media