JNSiena.2way: Calculates Johnson-Neyman significance regions and MOM...

In johninpdx/JMLUtils: Wrappers for R administration, general statistics, data management

Calculates Johnson-Neyman significance regions and MOM signficance values (raw and Bonferroni-Holm adjusted) for SAOM 2-way interactions.

Description

Calculates Johnson-Neyman significance regions and MOM signficance values (raw and Bonferroni-Holm adjusted) for SAOM 2-way interactions.

Usage

JNSiena.2way(
  siena07out,
  theta1,
  theta2,
  thetaInt,
  theta1vals,
  theta2vals,
  sigRegion = TRUE,
  alpha = 0.05
)

Arguments

siena07out	An object of class 'sienaFit'
theta1	Number (from RSiena output) of 1st effect involved in a 2-way interaction effect included in 'siena07out'. This effect will be the primary predictor for the first table of pick-a-point results in the output, and the moderator in the second such table.
theta2	Number (from RSiena output) of 2st effect involved in a 2-way interaction effect included in 'siena07out'. This effect will be the moderator for the first table of pick-a-point results in the output, and the primary predictor in the second such table.
thetaInt	Number (from RSiena output) of the multiplicative interaction of the two effects identified by theta1 and theta2.
theta1vals	A vector of numerical values of effect 1 for which raw and B-H-adjusted significance should be obtained, applying both to effect 1 as a predictor (in the 1st output table) and as a moderator (in the 2nd output table).
theta2vals	A vector of numerical values of effect 2 for which raw and B-H-adjusted significance should be obtained, applying both to effect 2 as a predictor (in the 2nd output table) and as a moderator (in the 1st output table).
sigRegion	Logical; default TRUE. If TRUE, Johnson-Neyman (raw) significance limits are calculated for the case of effect 1 as primary and effect 2 as moderator, and vice versa.
alpha	Number (default = 0.05) corresponding the 2-tailed significance level used to calculate whether a moderated effect is significant for a given value (for points in theta1vals and theta2vals) or range (for JN significance region computation).

Value

A list of 5 elements.

1. 1. *thetas* Three matrices of the predicted theta values of each of predictors 1, 2, and 3, calculated on each pair of values in the range (or discrete values) provided for each of the remaining two moderators.

2. 2. *standard_errors* SEs calculated conditionally in the same way as for the thetas.

3. 3. *p_values* P-values calculated conditionally in the same way as for the thetas.

4. 4. *significance* TRUE if the p-value falls within the Region Of Significance, as specified by alpha and, possibly also any adjustment for multiple tests; otherwise FALSE. Calculated in a matrix with the same conditions as for thetas, etc.

5. 5. *plots* A list with three class gg and ggplot elements as its members, corresponding to each of the three predictors being taken as the primary predictor, with the same ordering as for the previous list elements (theta, etc.).

johninpdx/JMLUtils documentation built on Dec. 29, 2024, 1:27 p.m.