External, Internal and Construct Validity

External, Internal and Construct Validity

External Validity

The extent to which an issue in a study can be said to be true in other cases is referred to as external validity. Eternal validity is concerned with whether a study's findings can be generalized validly. Example, if the same research is conducted at different times, places or to different people, would the same results be obtained. Thus, external validity tries to determine whether a study's result, which came from a sample can be functional to the population as a whole. One crucial determinant to external validity is the sample representation of a population one the pertinent scopes.

Internal Validity

Presence of sufficient evidence to support that an experiment makes a difference is what concerns internal validity. Internal validity estimates the extent of making a conclusion on the cause and effect relationship. This is subject to the measures deployed, setting and designing of the research as a whole. Creswell (2008)

identifies internal validity has threats obtained from researcher's data results which, is insufficient and experiences from participants' opinions. The variable used in a study ought to be accurate and valid if internal validity is to be satisfied.

Construct Validity

Construct validity works around operationalization of a test that is practical derived from theory. This is to say it assess to what extent a survey to test an existing theory actually, measures actually measures the variables of the theory. It involves theoretical and empirical backing to how the construct is interpreted. The assessment involves analyses of the relationship revealed in responses from different test items and includes test's internal structure. Creswell (2008)

emphasizes that, the content factor is not only the determining aspect. The results ought also to be correlated with other result to show internal evidence inclusiveness and demonstrate reliability.

Comparisons and Contrasts

It has been a controversial subject which validity is more valuable. Campbell and Stanley (1963)

said that a good study ought to have all three validities. Internal validity is essential and not dispensable. External validity however, is involved with the same result of a study and can be observed in other situations. A Study's external validity can easily be weakened by one single counter-example despite the number of previous findings that concur, Campbell & Stanley, 1963.

This statement seems to support the argument that internal validity bears more weight that external validity.

A lack of generalizability in a study means that cause and effect relationship (causal effect) that stems from internal validity is useless for predictions and decision making. Further Briggs (2008)

adds weight to the argument saying although internal validity and the statistical conclusion adds weight to the causal effect, external validity and construct validity are essential for the causal effects and generalization purposes to many settings. To ensure concrete construct validity level on a study's dependent variable, re-examination for reliability is noteworthy not just the causal effect. Details on the sample and sampling procedures should be analyzed thoroughly for rigorous data Creswell, 2008()

Threats to external validity

Non-representative sampling

It implies a situation where sample picked to act as a representative of the population is insufficient to cover all facets of the population. In this case, the sample is inadequate to act as a representative and is expected to obscure generalization of research result.

Non-representative research context

It describes a situation where the research was set out of context. Example if the survey conducted tried to evaluate the reaction of persons with high intelligence quotient while generalization is subjected to the average person. This question obscures the research results from generalization to a population of persons with an average intelligence quotient.

Threats to construct validity