help linequate
-------------------------------------------------------------------------------

Title
    linequate -- Calculates linear equating constants A & B using Tucker and
        Levine's equal and unequal reliability methods.


Syntax
        linequate anchortest newtest oldtest

  -----------------------------------------------------------------------------
    Note: Error variances used in the Levine Methods are estimated using
      Angoff's method (see Peterson, Cook, & Stocking, 1983).


Description
    Calculates linear equating constants A & B needed to equate a new test
    (X) to an old test (Y) using an anchor test.  The linear equating
    constants are to be applied to transform scores using the following
    equation: T(x)=Ax+B, where x represents the scores to be transformed and
    T(x) represents the transformed scores.  This command uses three methods:
    Tucker, Levine's equal reliability method, and Levine's unequal
    reliability method (see Peterson, Cook, & Stocking, 1983).


Options
    none no options are currently available.


Results
    Results are printed in the results window. The equating constants are
    also stored in the following scalars r(tae), r(tb), r(La), r(Lb), r(LUa),
    and r(Lub). The number of observations for each linking group is also
    printed as Nya (the number of students who took the anchor and old form
    of the test), Nxa (the number of students who took the anchor and the new
    form of the test, and Na (the number of students who took the anchor
    test).

Remarks
    If the application of the formulas for the linear equating constancts
    involve imaginary numbers, such as the square root of a negative number,
    a blank result will print.


Examples
. example
linequate anchor testX testY

. example score transformation using results from Tucker's method.
gen testx_on_y=r(ta)*testX+r(tb)

Author
    L.W. McGuire, University of Minnesota-Twin Cities, Educational
    Psychology, contact: lwmcguir@umn.edu



References
    Petersen, N., Cook, L. & Stocking, M. (1983).  IRT versus Conventional
    Equating Methods: A Comparative Study of Scale Stability. Journal of
    Educational Statistics, 8, 137-156.