ERIC Number: ED046195
Record Type: Non-Journal
Publication Date: 1970-Jun
Reference Count: N/A
Correcting Correlations for Restriction in Range Due to Selection on an Unmeasured Variable. Interim Report.
Bryant, N. Dale
The paper points out that, in working with special groups, correlations are often distorted because the variability of the measures being correlated are restricted in the groups. Presented is a formula whereby a Pearson product-moment correlation can be corrected for restrictions in range in situations where the basis of selection is unmeasured, but where the extent of restriction for each of the two measures being correlated is known, and where the variables are assumed to be normally distributed in the population. Three examples of the use of the formula are given: in a case where a comparison is to be made between a value derived from an unrestricted sample and one derived from a restricted sample; a case when a correlation is obtained on a special restricted sample and must be generalized to the population; and in estimating the validity of a test, where the criterion and the test scores are available on the same individuals only in a restricted sample where the basis of the selection is not clear or not measured. (Author/KW)
Publication Type: N/A
Education Level: N/A
Sponsor: Bureau of Education for the Handicapped (DHEW/OE), Washington, DC.
Authoring Institution: Columbia Univ., New York, NY. Teachers College.