Taylor Series Approximation of variance of ρ corrected for psychometric artifacts
Description
Functions to estimate the variance of ρ corrected for psychometric artifacts.
These functions use Taylor series approximations (i.e., the delta method) to estimate the variance in observed effect sizes predictable from the variance in artifact distributions based on the partial derivatives.
The available Taylor-series functions include:
estimate_var_rho_tsa_meas: Variance of ρ corrected for measurement error only
estimate_var_rho_tsa_uvdrr: Variance of ρ corrected for univariate direct range restriction (i.e., Case II) and measurement error
estimate_var_rho_tsa_bvdrr: Variance of ρ corrected for bivariate direct range restriction and measurement error
estimate_var_rho_tsa_uvirr: Variance of ρ corrected for univariate indirect range restriction (i.e., Case IV) and measurement error
estimate_var_rho_tsa_bvirr: Variance of ρ corrected for bivariate indirect range restriction (i.e., Case V) and measurement error
estimate_var_rho_tsa_rb1: Variance of ρ corrected using Raju and Burke's TSA1 correction for direct range restriction and measurement error
estimate_var_rho_tsa_rb2: Variance of ρ corrected using Raju and Burke's TSA2 correction for direct range restriction and measurement error. Note that a typographical error in Raju and Burke's article has been corrected in this function so as to compute appropriate partial derivatives.
Variance of square roots of reliability estimates for X.
mean_qy
Mean square root of reliability for Y.
var_qy
Variance of square roots of reliability estimates for Y.
...
Additional arguments.
mean_rtpa
Mean corrected correlation.
var_rxyi
Variance of observed correlations.
mean_ux
Mean observed-score u ratio for X.
var_ux
Variance of observed-score u ratios for X.
mean_qxa
Mean square root of unrestricted reliability for X.
var_qxa
Variance of square roots of unrestricted reliability estimates for X.
mean_qyi
Mean square root of restricted reliability for Y.
var_qyi
Variance of square roots of restricted reliability estimates for Y.
mean_uy
Mean observed-score u ratio for Y.
var_uy
Variance of observed-score u ratios for Y.
mean_qya
Mean square root of unrestricted reliability for Y.
var_qya
Variance of square roots of unrestricted reliability estimates for Y.
mean_ut
Mean true-score u ratio for X.
var_ut
Variance of true-score u ratios for X.
sign_rxz
Sign of the relationship between X and the selection mechanism.
sign_ryz
Sign of the relationship between Y and the selection mechanism.
mean_rxx
Mean reliability for X.
var_rxx
Variance of reliability estimates for X.
mean_ryy
Mean reliability for Y.
var_ryy
Variance of reliability estimates for Y.
Details
######## Measurement error only ########
The attenuation formula for measurement error is
ρXY=ρTPqXqY
where ρXY is an observed correlation, ρTP is a true-score correlation, and qX and qY are the square roots of reliability coefficients for X and Y, respectively.
The Taylor series approximation of the variance of ρTP can be computed using the following linear equation,
where b1, b2, and b3 are first-order partial derivatives of the attenuation formula with respect to qX, qY, and ρTP, respectively.
The first-order partial derivatives of the attenuation formula are:
b1=∂qX∂ρXY=ρTPqY
b2=∂qY∂ρXY=ρTPqX
b3=∂ρTP∂ρXY=qXqY
######## Univariate direct range restriction (UVDRR; i.e., Case II) ########
The UVDRR attenuation procedure may be represented as
where b1, b2, b3, and b4 are first-order partial derivatives of the attenuation formula with respect to qXa, qYi, uX, and ρTPa, respectively.
The first-order partial derivatives of the attenuation formula are:
where b1, b2, b3, and b4 are first-order partial derivatives of the attenuation formula with respect to qXa, qYi, uT, and ρTPa, respectively.
The first-order partial derivatives of the attenuation formula are:
where b1, b2, b3, b4, and b5 are first-order partial derivatives of the attenuation formula with respect to qXa, qYa, uX, uY, and ρTPa, respectively.
First, we define terms to simplify the computation of partial derivatives:
B=(ρTPa2qXa2qYa2+qXaqYaA−1)
C=2ρTPaqXa2qYa2uXuYA
The first-order partial derivatives of the attenuation formula are:
where b1, b2, b3, b4, and b5 are first-order partial derivatives of the attenuation formula with respect to qXa, qYa, uX, uY, and ρTPa, respectively.
First, we define terms to simplify the computation of partial derivatives:
where A, B, C, and D are first-order partial derivatives of the attenuation formula with respect to ρTPa, ρXXa, ρYYa, and uX, respectively.
The first-order partial derivatives of the attenuation formula are:
where E, F, G, and H are first-order partial derivatives of the attenuation formula with respect to ρTPa, qXa, qYa, and uX, respectively.
The first-order partial derivatives of the attenuation formula (with typographic errors in the original article corrected) are:
Vector of meta-analytic variances estimated via Taylor series approximation.
Notes
A typographical error in Raju and Burke's article has been corrected in estimate_var_rho_tsa_rb2 so as to compute appropriate partial derivatives.
References
Dahlke, J. A., & Wiernik, B. M. (2020). Not restricted to selection research:
Accounting for indirect range restriction in organizational research.
Organizational Research Methods, 23(4), 717–749. doi:10.1177/1094428119859398
Hunter, J. E., Schmidt, F. L., & Le, H. (2006).
Implications of direct and indirect range restriction for meta-analysis methods and findings.
Journal of Applied Psychology, 91(3), 594–612. doi:10.1037/0021-9010.91.3.594
Raju, N. S., & Burke, M. J. (1983). Two new procedures for studying validity generalization.
Journal of Applied Psychology, 68(3), 382–395. doi:10.1037/0021-9010.68.3.382