backtr: Back-transform normal score transformed data.

In truemoid/mstats: Some Miscellaneous Statistical Functions

Description

Given a vector of normal scores and a normal score object (from nscore), the function returns a vector of back-transformed values. One major issue is how to extrapolate to the tails. Options other than none may result in dramatically incorrect tail estimates!

Usage

backtr(scores, model, tails = "none", draw = FALSE)

Arguments

scores	Numeric vector of normal score data to be transformed.
model	Dataframe with column of raw data matched to column of normal score data. Is output of 'nscore' function.
tails	Treatment of distribution tails: "none", "separate", or "equal".
draw	Plot the distributions.

Details

Tails options: 'none': No extrapolation; more extreme score values will revert to the original min and max values. 'equal': Calculate magnitude in std deviations of the scores about initial data mean. Extrapolation is linear to these deviations. Will be based upon deviations from the mean of the original hard data - possibly quite dangerous! 'separate': This calculates a separate sd for values above and below the mean.

Value

Numeric vector of back-transformed values.

Author(s)

Alex M Trueman Modified from Ashton Shortridge, May/June, 2008. https://msu.edu/~ashton/research/code/nscore.R Originally based on the GSLIB code for backtr function.

truemoid/mstats documentation built on May 24, 2019, 7:17 a.m.