abconv: Derive linear conversion coefficients from a set of indeterminate coefficients

Description

If a method comparison model is defined as $y_{mi} = \alpha_m + \beta_m \mu_i, m=1,2$ y_mi = alpha_m + beta_m*mu_i, m=1,2 the coefficients of the linear conversion from method 1 to 2 are computed as: $\alpha_{2|1} = -\alpha_2-\alpha_1\beta_2/\beta_1$ alpha_(2|1) = -alpha_2-alpha_1*beta_2/beta_1 $beta_(2|1) = beta_2/beta_1$ Morover the the point where the linear conversion function intersects the identity line is computed too.. The function is designed to work on numerical vectors of posterior samples from BUGS output.

Usage

abconv( a1, b1 = 1:4, a2 = NULL, b2 = NULL, col.names = c("alpha.2.1", "beta.2.1", "id.2.1") )

Arguments

Numerical vector of intercepts for first method. Alternatively a dataframe where the vectors are selected from.

Numerical vector of slopes for first method. If a1 is a dataframe, b1 is assumed to be a numerical vector of length 4 pointing to the columns of a1 with the intercepts and slopes.

Numerical vector of intercepts for second method.

Numerical vector of slopes for second method.

col.names

Names for the resulting three vectors.

Value

A dataframe with three columns: intercept and slope for the conversion from method 1 to method 2, and the value where the conversion is the identity.

References

B Carstensen: Comparing and predicting between several methods of measurement, Biostatistics, 5, pp 399-413, 2004

Examples

Run this code

abconv( 0.3, 0.9, 0.8, 0.8 )