Sim and Wright, (2000:125) stated that, Validity relates to the truthfulness of data and requires independent knowledge of the true nature or magnitude of the entity. Clinical research scientists are much aware of the need for validity in their research. Validating the treatment is important to determine the efficacy of a treatment as well as effectiveness and the treatment provided and its outcome. It is learnt that in 1957, American Social Scientist Donald Campbell introduced the concepts of internal validity and external validity. The study has an internal validity, when the result of a study demonstrates the existence of a causal relationship between the independent and dependent variables. Campbell and Stanley, (1963) noted that the internal validity is the basic required accuracy for any experiment. On the other hand, the external validity refers to the extent to which the results of a study or the experiments are made applicable to the population from which the participant in the experiment is drawn. Similarly, as Thomas and Nelson, (1990) noted that the external validity is about the generalizability of the findings. The first part of this essay will discuss the concepts of internal validity and external validity, the causal effect of variables and extraneous variables with examples. In the second part, the essay will critically evaluate the implication of internal and external validity and their relationship. Due to word constraint this essay will focus only on the quantitative research.
The experimental (explanatory) research concerned with cause and effect by its nature, internal validity can be discussed within the frame work of quantitative research. Furthermore, Thomas and Nelson, (1990) stated that the cause and effect can be determined by the use of good research knowledge to the experiment design. According to Judd, et al., (1991, cited in French, et al., 2001:16) the internal validity concerns the ”extent to which conclusions can be drawn about the casual effects of one variable on another”. For example, the treatment or a cause (independent variable) such as moist heat which effectively relieves the symptoms of low back pain or an effect (dependent variable). Before establishing a cause-effect relationship, it is important to find the relationship between the treatment and the symptom. The argument is that when treatment is given, the relief in pain (outcome) is observed, but when no treatment is given no outcome is observed. This provides the evidence that, the treatment and the outcome are related. However, it does not provide the evidence that the particular treatment caused the outcome. Perhaps, it may be some other factor caused the outcome rather than the treatment. For example, if the subjects were on pain relieving medication and medication may be the confounding factor in relieving the symptoms of low back pain. If the pain relieving medication factors are not controlled, they will have a confounding influence on the independent variable which threatens the internal validity. However, sometimes it is important to emphasize that the confounding variables cannot be fully controlled. The researcher assumes that there is a casual relationship in the study and claims that, research findings have implication for other groups and individuals in other settings and time. These claims examine the external validity. Polit and Beck, (2008) stated that the external validity is a major concern in quantitative research where there is a difficulty in generalizing the result from controlled research settings to real clinical practice settings. For example, findings about a pain relieving treatment in a study of African women cannot be generalized to women in Australia. Finally, it is a researcher decision to design a study to enhance the external validity.
The potential threats to the internal validity are history, maturation, testing, instrumentation, statistical regression, selection error and mortality. Threats to the external validity are, selection, setting and time. A few threats are discussed in the following example. A randomised control trial study was conducted by Deyle, et al., (2000), to evaluate the effectiveness of physical therapy for osteoarthritis of the knee. The study concluded that a combination of manual physical therapy and supervised exercise yields functional benefits for patients with osteoarthritis of the knee and may delay or prevent the need for surgical intervention. Even though confounding variables were controlled in this study by using randomisation, treatment and placebo groups, the researcher could not overcome with the threats like maturation and other extraneous variables. During the intervention patients were also on medication and exercises. They were asked to contact the researcher after one year of the intervention for the feedback. The researcher failed to explain whether intervention caused the effect or the medication or due to result of the passage of time. Simultaneously this study lacks the generalizability due to the fact that patients had to have sufficient English language skills and were required to live within a one hour drive from the clinics. Portney and Watkins, (2009:162) stated that, ”extraneous variables can be controlled by the experimenter by manipulating the independent variable, by random assignment of subjects to the groups and by using a control or comparison group within the design”. Masking or blinding is also be used in many quantitative research studies to prevent the bias and to strengthen the internal validity. The external validity can be strengthened by selecting the sample which represents the population and there by applying the findings to a broader group. Similarly in the representative of clinical settings where the study takes place and the findings can be applied to strengthen the other settings. It was stated by Polit and Beck, (2008) that multisite study results are confident enough in attaining the generalizability, if the results have been replicated in several sites where the study involves a heterogeneous sample.
The relationship between the internal validity and external validity refers to the famous researchers Campbell and Stanley, (1963) statement that, often external validity sacrificed if the researcher concentrates on the internal validity. It is learnt from the literatures that, many researchers have given importance to the internal validity in their studies. However, in an applied area of clinical research, the purpose is to improve the health of the public and it is also important that the external validity be emphasized and strengthened. It is true that, the researcher implements more control measures to increase the internal validity, the experiment will become more artificial and thereby the external validity suffers or decreases. Cronbach, (1982) argued that, if a treatment is expected to be relevant to a broader context, the causal inference must go beyond the specific conditions. If the study lacks generalizability, then the so-called internally valid causal effect is useless to decision makers. Polit and Beck, (2008) noted that there are many differences between the countries in methods of diagnosis and management, as well as important racial differences in susceptibility to disease and natural history of the disease. Differences between health-care systems can affect the external validity. In an article, ”Efficacy and Effectiveness Trials (and Other Phases of Research) in the Development of Health Promotion Programs” Flay, (1985) proposes a model that emphasizes the internal and external validity at different stages of the research process and that would lead to the translation of research to practice. In a controlled experiment like randomized control trials of public health interventions, efficacy trials have a high internal validity but often have the liability of low external validity. But, effectiveness trials have a high external validity. Historically, researchers have tended to focus on maximizing the internal validity. The idea is that the interventions should work under highly controlled conditions than its implication to the different population groups, organizations, or settings. Similarly, Polgar and Thomas, (2008) noted that funding organizations and journals have tended to be more concerned with the scientific rigor of intervention studies than with the generalizability of results. The consequence of this emphasis on internal validity led to lack of attention to and information about external validity, which has contributed to the researcher’s failure to translate the study into the general population. Thus, the progress from efficacy trials to effectiveness trials has not become a reality because of the time and cost involved in the process of research to practice. As a result of the failure of this model, practitioners are often unable to determine whether the given study’s findings can apply to their local setting, population or resources. However, there are several reasons for the lack of information on external validity being an important contributor to the failure to translate research into public health practice. Policy and administrative decision-makers are unable to determine the generalizability or breadth of applicability of research findings. In addition Hamersley, (1991) criticised that the distinction between the internal and external validity is fundamentally misleading because it reflects a failure to distinguish relations between events and relations between variables. Polit and Beck, (2008) noted that in the recent studies researchers are interested in promoting designs that aim to achieve a balance between internal and external validity in a single intervention study. Such practical (or pragmatic) clinical trials or studies attempt to maximize external validity with the smallest possible negative effect of internal validity. For example, Thomas, et al., (2006) conducted a study on ”randomised control trial of short course of traditional acupuncture compared with usual care of persistent non-specific low back pain”. The study concluded that referral to a qualified traditional acupuncturist for a short course of treatment seems safe and acceptable to patients with low back pain. However, the generalizability of this study finding rests on the assumption that participating acupuncturists, general practitioners, and patients in York are similar to those found elsewhere. According to Polit and Beck (2008), in health care research RE-AIM framework has been formed by Russell. E. Glasgow in 2006 to design and to evaluate the intervention research. It involves a scrutiny of five aspects of study like, reach, efficacy, adoption, implementation and maintenance. It may help in improving the generalizability of the research study or enhance the external validity. An Importance of an internal validity or the external validity is a controversial topic in the research community. Campbell and Stanley, (1963) stated that, a good study should be strong in both types of validity. The internal validity is indispensable and essential while the question of external validity is never completely answerable. In other words, Campbell and Stanley’s statement implies that the internal validity is more important than the external validity. However a study with no external validity still found true relationship for the sample that was studied. For example, if the researcher conducts a study on Bangladeshi farmers in the Afghan region, the findings of the study cannot be generalize to the Chinese horticulturist in the Malaya region. But researcher still knows more about the Bangladeshi farmers.
In conclusion, it is understood that, the internal validity and the external validity plays an important role in the scientific research. The internal validity establishes the truth about inferences regarding cause-effect or causal relationship. The external validity establishes the truth of conclusion that involve generalisation. As internal and external validity are fundamental to any experimental research, the researcher should aware of threatening factors. In quantitative research, the use of randomisation and the control groups reduces the threats to internal validity. Meanwhile, sample selection and settings in the research helps in controlling the threats to external validity. The strength of internal and the external validity of a study can help researchers to evaluate the relative importance of that study in an overall program of research. To balance the validity, researcher should conduct a study that emphasizes the internal validity (efficacy studies) and undertake effectiveness studies that emphasize the external validity. The internal validity and the external validity are important to building an evidence based study in the clinical research. A clear definition of concepts involved in the experiment not only ensures its validity, but also increases the chances that it could be used by other researchers wishing to carry out a follow-up or similar investigation.