adjust.by.pre: Adjusting Learning Threshold

In haberdashPI/wright.lab.util: A set of utilitiy functions used in our lab.

Description

Used for adjusting learning thresholds by a pre-test. In learning studies usually the amount of learning is influenced by where a subject's performance began at the start of learning. To account for this one can either introduce the pre test as a covariate in statistical anlaysis, or more informally for graphs and the like, one can adjust the performance measure so that it isas if all participants started with the same pre-test, assuming a linear relationship between the pre-test and any other test. This function returns a modified data frame that allows both approaches to be easilly utilized, by adding column to every row with the adjusted performance measure that row and the the pre test performance.

Usage

adjust.by.pre(formula,data=list(),agg.slope.by = NULL,agg.pre.by = NULL)

Arguments

formula	A formula describing the data frame. The dependent variable indicates a numeric performance measure to be adjusted. The first independent variable should be numeric, and is the indication of time (e.g. testing days). The value 0 is assumed to be the pre-test. All other variables should be ordinal or nominal grouping variables. Each unique combination of grouping variables is assumed to be a unique subject.
data	The data environment for the aforementioned formula.
agg.slope.by	By default (when this is NULL) the slope used to adjuste the pre-test to a given day is found across all conditions. If you would like to calculate the slope across some subset of the data, specify which columns the slope should be aggregated by using this variable.
agg.pre.by	When adjusting pre-tests there must be a mean pre-test to adjust to. This is normally done across the entire data set (when agg.pre.by is NULL). If you would like to calculate the mean pre-test across some subset of the data, specify which columsn the mean pre-test should be aggregated by using this variable.

Author(s)

David Little

References

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (2nd ed.). Hillsdale, NJ: Erlbaum.

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

haberdashPI/wright.lab.util documentation built on Nov. 4, 2019, 1:25 p.m.