Threats to Internal Validity and External Validity in Quantitative Research

Threats to internal validity. The term internal validity is used to describe the effectiveness and appropriateness of a study in examining what was intended to examine (Gabay, 2016). It is aimed at addressing the issues of how well the research was designed and carried out and the confidence the researchers and the readers can be in the findings derived from such studies. One of the crucial questions in internal validity is whether the results or the findings of the study can be linked to the interventions and not to other extraneous factors or confounding variables (Trochim, Donnelly, & Arora, 2016). A researcher ensures the internal validity of a study by using a robust research design capable of eliminating alternative explanations of research results (Trochim et al., 2016).

Many threats to the internal validity exist. One of these threats is selection bias (Elwood, 2017). Every time a researcher selects participants for a research study using a non-probabilistic way or a non-random technique, the internal validity of such a study is reduced. Elwood (2017) further noted that the selection of the participants affects the internal validity of research if the process of picking the participants for the study has differential impacts on the groups under comparison.

Hawthorne effect also affects internal validity (Schmidt & Brown, 2017). This refers to the participants' awareness that they are in experimental groups making them not to react normally to experimental conditions (Riazi, 2016). It is also known as the observer effect, the likelihood of individuals changing their behavior when they are under observation, when enrolled in a study, or when they are receiving additional attention.

Hawthorne effect is manifested in many ways. For instance, the mere act of being in a study can alter the participants' behaviors. Consequently, any changes observed should not be solely linked to the stimulus unless there is a control group which enables the investigator to make comparisons (Boyle & Schmierbach, 2015). Additionally, the research subjects may alter their behaviors to suit the researchers' expectations (Boyle & Schmierbach, 2015). Mortality or attrition has also been reported to negatively affect internal validity (Schmidt & Brown, 2017). This is because the participants who drop out adversely affect the equalization of the group or other features of the research subjects (Schmidt & Brown, 2017).

Threats to external validity. External validity can be described as the extent to which researchers believe the findings of a particular study will be generalizable to a broader population (Lankoski & Bjork, 2015). One of the threats to external validity is sampling error. In this case, the confounding factor that affects the generalizability of the findings is a failure of the sample to represent the population (Schmidt & Brown, 2017). Poor ecological validity, where the research conditions do not reflect real-world practice, is another threat to external validity (Schmidt & Brown, 2017).

Mitigating Threats to Internal and External Validity

Mitigating Threats to Internal Validity

Selection bias threat to internal validity is mitigated by using a robust research design, such as a randomized controlled trial. In this type of design, the participants are randomly allocated to a comparison group and a treatment group. This research design protects against measured and unmeasured confounders attributable to treatment selection bias. Additionally, in experimental research designs, manipulation of the independent or predictor variable(s) and random assignment of the subjects to conditions supports the internal validity of a study. Conversely, in quasi-experimental and correlational research designs, the researchers ensure that the internal validity of achieved through theoretical evidence or measurement of control variables (Lankoski & Bjork, 2015).

Mitigating Threats to External Validity

A threat to the external validity of a study attributable to non-random selection of participants to a study can be eliminated through a random selection of participants. Some of the examples of random selection techniques include probability sampling techniques such as cluster sampling, systematic sampling, stratified random sampling, and simple random sampling (Neelankavil, 2015). Sampling error threat can also be mitigated in a study through replication of the research using new samples, use of a considerable number of participants, selecting participants from many sites, and random assignment (Schmidt & Brown, 2017). Moreover, poor ecological validity can be mitigated through replication with effectiveness studies and by ensuring that the investigator maintains sensitivity to issues related to real-life practice (Schmidt & Brown, 2017).

Hawthorne effect can be mitigated through the use of a special research design that utilizes what is referred to as Hawthorne control. Hawthorne control is comprised of three groups of research subjects: the Hawthorne control that receives a treatment that is not relevant to the outcome of interest of the study, the control group which receives a treatment of interest to the investigators, and the control group which is not given any treatment. Moreover, the Hawthorne effect can be controlled by statistically estimating missing data, by employing strategies aimed at encouraging participation, and carrying out intent to perform analysis (Schmidt & Brown, 2017). Thirdly, attrition or mortality can be mitigated by making sure that there is equal attention of both the treatment and the control groups and by carrying out the blind intervention of the participants (Schmidt & Brown, 2017).

Ethical Issue in Quantitative Research and Impact on Design Decisions

One of the potential ethical issues in quantitative research is informed consent (Fan, 2015). When a participant gives informed consent in a research study, it means that he or she knowingly, voluntarily and intelligently agreed to participate in a study. This is intended to ensure that the participant's autonomy to take part in a study is maintained. This ethical issue affects design decisions because the researcher is required to issue every prospective participant an informed consent form detailing the title of the research, its purpose, study procedures, risks, benefits, confidentiality, compensation, and researchers' contact information. Consequently, the researcher is required to prepare an informed consent document.

Amenability of Research Topic to Scientific Study

Amenability of the research topic to scientific study in quantitative methodology refers to the appropriateness of the research design to the research topic. For instance, if the study is aimed at examining the causal association between two variables, randomized, double-blind experimental research design (Merrill, 2017). On the other hand, if a research topic is aimed at examining the relationship exist between two variables, correlation research design is used (Gravetter & Forzano, 2016).

References

Boyle, M. P., & Schmierbach, M. (2015). Applied communication research methods: getting started as a researcher. New York, NY: Routledge.

Elwood, M. (2017). Critical appraisal of epidemiological studies and clinical trials. Oxford, OX: Oxford University Press.

Fan, R. (2015). Family-oriented informed consent: East Asian and American perspectives. Retrieved from http://public.eblib.com/choice/publicfullrecord.aspx?p=1998113

Gabay, M. (2016). The clinical practice of drug information. Retrieved from http://public.eblib.com/choice/publicfullrecord.aspx?p=4441213

Gravetter, F. J., & Forzano, L.-A. B. (2016). Research methods for the behavioral sciences. Stamford, CT: Cengage Learning.

Lankoski, P., & Bjork, S. (2015). Game research methods: an overview. Pittsburgh, PA: ETC Press.

Longo, D. L., Fauci, A. S., Kasper, D. L., Hauser, S. L., Jameson, J. L., & Loscalzo, J. (2015). Harrison's principles of internal medicine. New York, NY: Mc Grow-Hill.

Merrill, R. M. (2017). Introduction to epidemiology. Burlington, MA: Jones & Bartlett Learning.

Neelankavil, J. P. (2015). International business research. New York, NY: Routledge Publishers.

Riazi, A. M. (2016). The Routledge encyclopedia of research methods in applied linguistics. New York, NY: Routledge Publishers.

Schmidt, N. A., & Brown, J. M. (Eds.). (2017). Evidence-based practice for nurses: appraisal and application of research. Burlington, Massachusetts: Jones & Bartlett Learning.

Trochim, W. M. K., Donnelly, J. P., & Arora, K. (2016). Research methods: the essential knowledge base. Boston, MA: Cengage Learning.