Monday, June 18, 2007

WJ III NU (normative update) special pub

Last week a number of questions were asked on the NASP listserv about the WJ III Normative Update (WJ III NU). A number of members provided good responses.

At this time I'd like to announce that a special WJ III NU Assessment Service Bulletin (ASB) will be available (from the Riverside web page) in approximately two weeks. The ASB will provide more in-depth information that addresses the major questions users may have about the WJ III NU. Below is the title of the ASB and the submitted draft (note - the text in the final publication will likely have some minor edits, changes, etc.) of the introduction, inclusive of a list of major questions that are answered in the document.

Once the final ASB is available for download, I will make a post at this blog and on a couple of listservs.

[Conflict of interest disclosure - I'm a coauthor of the WJ III]


WJ III/NU score differences: What the user can expect and why

The Woodcock-Johnson® III Normative Update (WJ III NU®) (Woodcock, McGrew, Schrank, & Mather, 2001, 2007) is a recalculation of the normative data for the Woodcock-Johnson III (WJ III ®) (Woodcock, McGrew, & Mather, 2001), based on the final 2000 U.S.census statistics (U.S.Census Bureau, 2005). The final 2000 census data are reflected in the norms provided by the WJ III NU Compuscore and Profiles Program (Compuscore®) (Schrank & Woodcock, 2007) and in the documentation provided in the WJ III NU Technical Manual (McGrew, Schrank, & Woodcock, 2007).The WJ III NU norms replace the original WJ III norms that were based on the U.S.Census Bureau’s 2000 census projections, which were issued in 1996 (Day, 1996).

The census bureau’s Population Projections Program issues projections of the United States resident population based on assumptions about future births, deaths, and international migration. Census projections are estimates of the population for future dates, and are subsequently replaced by census statistics. The 2000 census statistics produced a somewhat different description of the U.S. population than was assumed from the last projections (issued in 1996). For example, according to the bureau’s Greg Spencer, “when we took the 2000 census, we found about 6.8 million more people than we were expecting. When we went in and looked at the sources of that growth, we found that during the late 1990s, there was more migration than we had been measuring.” (Landphair, 2004, p. 1). Other unanticipated changes in the population were documented (U.S. Census Bureau, 2005), including shifts in age, sex, race, Hispanic origin, and residence. Some states grew at three times the national rate, and people had tended to cluster in locations were jobs were available and climate was preferred. For users of the WJ III, the normative update (NU) provides the most current comparisons to the U.S. population.

In addition, the WJ III NU utilized innovative statistical advancements to calculate the new norms. The use of bootstrap resampling procedures (Efron & Tibshirani, 1993) allowed for estimates of uncertainty and potential bias (in the sample data) to be incorporated into the calculation of the WJ III NU norms. The bootstrap-based norm development procedures used to recalculate WJ III NU norms results in more precise estimates of an individual’s tested performance.

Collectively, the refinement of the demographic characteristics in the WJ III NU norm data, based on updated U.S.Census statistics, and the utilization of “state-of-the-art” statistical methods for estimating the sample statistics used to calculate norms, allows users of the WJ III to have greater confidence in the accuracy of the WJ III NU-based scores.

Purposes

The following four general categories of questions are presented and discussed:
  • What score differences can I expect between test scores based on the original WJ III and WJ III NU norms? If there are differences, are the differences more noticeable with certain age groups?
  • Did the year 2000 U.S.Census population demographic final statistics change enough (since the year 2000 census projections) to make a real difference in the scores subjects will receive on the WJ III? That is, are the NU norms needed to provide better estimates of scores for comparison to the current U.S. population?
  • What is “bootstrap sampling” and, more importantly, how does this innovative statistical method provide more accurate norms than those published with the WJ III data in 2000?
  • What WJ III/NU scores should be used to track and compare test performance for individuals across time. What is the suggested best practice for tracking student growth from the WJ III to WJ III NU?

Technorati Tags: , , , , , , , , , , , , ,

Powered by ScribeFire.

No comments: