Wednesday, September 28, 2011

1949 WISC: What date(s) should be used to calculate the Flynn effect for historical IQ scores?

What date should be used when calculating the effect of norm obsolescence (aka, Flynn Effect) for old 1949 WISC scores?

I have seen psychological reports for three different Atkins MR/ID death penalty case reports and one social security report that referenced old WISC scores. The WISC was published in 1949 and was used well into the mid 1970's (when replaced by the WISC-R in 1974). The Flynn Effect (click here for posts linking to Flynn Effect Series of reports) was not yet a documented phenomenon, so examiners using the test late it it's life cycle were often not aware of the potential for a massive Flynn effect. At the end of it's life cycle WISC results were often interpreted on the basis of norms that were 20-25 years out of date! This is the worst case scenario for the Flynn Effect I have seen in old psychological reports.

There is some debate on which date for any IQ test should be used to estimate the Flynn Effect. The date of publication or the median date of the years during which the norm data was collected. I believe the consensus is the later.

But, when one turns to the WISC manual, it makes no mention of the years spanned during the norming. Thus, I made a request for information to a number of professional listservs and a number of people directed me to Flynn's (2006) "Tethering the Elephant" article. There is a footnote in that table that provides a lead.

Note. All dates assigned to tests refer to the date at which the test was normed. This is what is relevant, of course, not the date when the test was published. Another date that practitioners might like to have is that for the norming of the WISC: from 1947 to 1948.

Alan Kaufman also communicated with me privately (personal communication, 9-23-11) regarding my question. He worked with David Wechsler on the the WISC-R. He said that during that time the date of 1947 was almost always mentioned during their work, but he never did see concrete proof.

Finally, a member of a neuropsych. listserv sent me a 1950 smoking gun journal article authored by staff from Psychological Corporation. The article was a more in-depth description of the WISC standardization. Although only mentioned briefly in one sentence on p. 102, the dates 1947-1948 are mentioned as the period of the data collection.

Thus, when faced with historical records with old WISC scores, especially those where the effect of norm obsolescence on the scores is dramatic, the best estimate to use for calculating the Flynn Effect is 1947 to 1948. I would tend to think that given the publication and production processes at the time that most of the data was collected during 1947 as many 1948 activities were likely involved in printing and production. But, to be safe, I would suggest examiners use both 1947 and 1948 and then round to the average estimate. One year typically does not make a huge difference.

Thanks to all who sent me tips and the article. The internet, and many professional listservs, are awesome sources of information.

- iPost using BlogPress from Kevin McGrew's iPad

Generated by: Tag Generator