ERIC Number: ED164611
Record Type: RIE
Publication Date: 1978-Mar-28
Reference Count: 0
Toward a Quantitative Estimate of Internal Validity.
Popp, Jerome A.
Internal validity is described as a matter of how well a particular instance of data collection or generation can be described and explained. It is a property of the procedures used in the collection or generation of data. The notion of internal validity is examined in order to establish a method of quantitatively estimating it. A coefficient of internal validity is defined by the equation that it equals one divided by one more than the number of alternative independent plausible hypotheses. If experimented procedures rule out all alternative plausible hypotheses, then the internal validity coefficient equals one. As the number of alternative hypotheses approaches infinity, the internal validity coefficient approaches zero. Alternative hypotheses may be considered to be equivalent to alternative independent variables that are discrete and non-overlapping and cannot be ruled out and are independently definable. A suggestion is given for weighting independent variables according to their rank order of being plausible. This view is coherent with Harman's view of induction as inference to the best explanation. (Author/CTM)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Internal Validity
Note: Paper presented at the Annual Meeting of the American Educational Research Association (62nd, Toronto, Ontario, Canada, March 27-31, 1978)