| Anthony F. Gregorc | |
| Nationality: | American |
| Field: | Phenomenology |
| Work Institution: | Gregorc Associates Inc.[1] |
| Alma Mater: | Miami University (Ohio), Kent State University (Ohio). |
| Known For: | Mind Styles Model |
Anthony F. Gregorc is an American who has taught educational administration. He is best known for his disputed theory of aMind Styles Model and its associatedStyle Delineator.[2] The model tries to match education to particular learning styles, as identified by Gregorc.
Gregorc obtained a B.S. degree from Miami University and an M.S. degree and a Ph.D. degree from Kent State University. He has taught mathematics and biology and has been principal of a laboratory school for gifted youth. He was an associate professor of education administration at the University of Illinois at Urbana-Champaign and associate professor of curriculum and instruction at the University of Connecticut.[1] He is president of Gregorc Associates, Inc., in Columbia, Connecticut.
The Gregorc Style Delineator is a self-scoring written instrumentthat elicits responses to a set of 40 specificwords.[3] Scoring the responses will give values for a model with two axes: a"perceptual space duality," concrete vs. abstract, andan "ordering duality," sequential vs.random[4] The resulting quadrants are the "styles":
Descriptions of the characteristics of these styles can be found in thematerials available from Gregorc Associates.
A similarly structured (two-axis, four-style) learning style modelwith rather different axes and interpretation can be seen in theKolb LSI.
The design, conduct, and results of Gregorc's original testing of thevalidity of his instrument and model are presented in hisDevelopment, Technical, and Administration Manual,[5] self-published and sold by Gregorc Associates. Some peer review has sinceappeared in conventional channels:
With the exception of Joniak and Isakson (1988) and O'Brien (1990), the onlyother psychometric analysis of the GSD has been limited to Gregorc's (1979)initial assessments made during the instrument's early development in whichGregorc interviewed several hundred participants. He compared the agreement ofGSD scores with an untested self-assessment scale to establish the instrument'sface validity for each individual (i.e., the instrument's results versus anindividual's subjective agreement that their learning style profile tends tofit them). The correlations of the instrument's general results and thesubjectively rated agreement attributes were reported to be between .55 and.76. This problematic method was adopted again in a subsequent comparativeanalysis by the author (Gregorc, 1982c) and also yielded what Gregorcconsidered positive results--29% strongly agreeing, 57% agreeing, 14% unsure,and nonedisagreeing.[6]
Timothy Sewall, in a comparison of four learning style assessments(Gregorc's, Myers Briggs, Kolb LSI, and an LSI byCanfield) by review of their published supporting studies (i.e., withoutnew experimental work) concluded of Gregorc's design, "the most appropriateuse of this instrument would be to provide an example of how not toconstruct [an] assessment tool."[7]
Reio and Wiswell (2006) report on their own independent study andon those done earlier by O'Brien (1990) and Joniak and Isakson(1988).[8]
Internal consistency or reliability concernswhether evidence can show that an instrument is repeatably measuring something(which may be, but is "not necessarily what it is supposed to bemeasuring"[9]).
Gregorc (1982c) reported test-retest correlation coefficients of .85 to .88(measured twice with intervals ranging from 6 hours to 8 weeks) and alphacoefficients of .89 to .93 on all four scales. In this study, theCronbach's alpha coefficients on all scales or channels ranged from .54 to.68 (CS = .64, CR = .68, AR = .58, AS = .54). This study's alpha coefficientsare consistent with those reported by O'Brien (1990) and Joniak and Isakson(1988), which ranged from .51 to .64 and .23 to .66, respectively, on allscales.[10]
For internal consistency reliability estimates, although an alpha level of .70can be considered "adequate," for the purposes of this study we considered astricter alpha level of .80 as a "good" cutoff value for our psychometricexamination of the GSD(Henson, 2001).[11]
Construct validity concernswhether evidence can show that what the instrument is measuring is at all whatthe offered theory claims it is (whether each construct in the model "adequatelyrepresents what is intended by theoretical account of the construct beingmeasured"[12]).
The data disconfirmedboth the two- and four-factor confirmatory models. In the post hoc exploratoryfactor analyses, many of the factor pattern/structure coefficients wereambiguously associated with two or more of the four theoretical channels aswell. Overall, there was little support for the GSD's theoretical basis ordesign and the concomitant accurate portrayal of one's cognitive learningstyle.[13]
[F]ar more work is needed on the GSD if indeed two bipolardimensions and Gregorc's mediational or channel theory are to be empiricallysupported and if it is to be appropriately used with samples ofadults.[14]
Consistent with Joniak and Isaksen (1988) and O'Brien (1990), the GSD did notdisplay sufficient empirical evidence to validate the instrument's scores or toconfirm Gregorc's theoretical interpretation of four learning style channels ortwo bipolardimensions.[15]
A report from the UK think-tank Demos reported thatthe evidence for a variety of learning style models is "highly variable",that "authors are not by any means always frank about the evidence for theirwork, and secondary sources ... may ignore the question of evidence altogether,leaving the impression that there is no problem here."[16]