Summary
This chapter classifies the Equating Designs in order to show the similarities and differences between them. For example, the NEAT Design can be viewed as containing the EG Design as a special case when P=Q and A has only a single score value. Similarly the CB Design contains both EG and SG Designs in it.
Another approach to classifying the designs is based on the estimation of the parameters in the pre-smoothing step, i.e., the numbers and type of the distributions to be estimated—univariate (EG) and bivariate (SG, CB, and NEAT).
We can also classify the designs by the number of populations and samples that are involved. Table 2.5 identifies some of the various ways Equating Designs may be classified.
Some designs are simple, i.e., involve a single population of examinees, have less assumptions, but need either larger sample sizes (EG), or require the same people to take two test forms at the same time (SG and CB). Other designs are more complicated, i.e., involve two populations of test takers, make use of an anchor test, and require that additional assumptions be fulfilled (i.e., NEAT). These complexities are often compensated by their increased versatility.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Rights and permissions
Copyright information
© 2004 Springer-Verlag New York, Inc.
About this chapter
Cite this chapter
(2004). Data Collection Designs. In: The Kernel Method of Test Equating. Statistics for Social Science and Behavorial Sciences. Springer, New York, NY. https://doi.org/10.1007/0-387-21719-3_2
Download citation
DOI: https://doi.org/10.1007/0-387-21719-3_2
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-01985-7
Online ISBN: 978-0-387-21719-2
eBook Packages: Springer Book Archive