ERIC Number: ED109230
Record Type: Non-Journal
Publication Date: 1974
Reference Count: N/A
An Assessment of the Degree of Isomorphism among Alternative Interval Scaling Approaches.
Brunn, R. Beto; Fuller, Theron
It is hypothesized that the use of an empirical indicator of a theoretical construct implies certain epistemic assumptions about the nature of the construct. In this study, the implications of assuming interval level of measurement using various isomorphic monotonically increasing epistemic alternatives to a commonly used occupational status index was investigated in the following manner. (1) Simulated samples of 100 Duncan's Socio-Economic Index scores were drawn by generating random independent samples from normally distributed populations with different means. (2) The simulated SEI scores were transformed to correspond to six alternate assumptions about the nature of the pattern of intervalization. This resulted in seven isomorphic, monotonically increasing measures of occupational status differing only in the pattern of intervalization. (3) The degree of divergency among epistemic models was determined by analyzing matrices of coefficients of epistemic error corresponding to the "mean epistemic error" between two given conceptualizations under the assumption that one conceptualization is used and the other conceptualization perfectly describes the theoretical construct. (4) The nature and magnitude of the measurement error introduced into path models by using different assumptions of intervalization was investigated by comparing simulated causal models involving occupational status using the alternative interval conceptualizations as one of the two independent variables. (Author/BJG)
Publication Type: Reports - Research
Education Level: N/A
Sponsor: Texas A and M Univ., College Station. Texas Agricultural Experiment Station.
Authoring Institution: N/A