SensMixed (version 2.1-0)

sensmixedControl: Control of sensmixed function

Description

Construct control structures for sensmixed function.. All arguments have defaults, and can be grouped into

• mixed assessor model (MAM) parameters (are only used if MAM = TRUE in the sensmixed function)

• model analysis specifications

Usage

sensmixedControl(MAM_mult_scaling = FALSE, MAM_oneway_rand = FALSE,
                              MAM_balanced = FALSE, MAM_adjusted = FALSE, 
                              MAM_alpha_conditional = 1,
                              calc_post_hoc = FALSE, parallel = FALSE, 
                              reduce.random=TRUE, alpha.random = 0.1, 
                              alpha.fixed = 0.05, interact.symbol = ":", 
                              keep.effs = NULL)

Arguments

MAM_mult_scaling

logical. Whether multiple scaling should be used. This option is not fully investigated (recommened is FALSE)

MAM_oneway_rand

logical. Whether there should be just one prod_effect (in a multi-factorial product structure setting this means the highest order interaction) as part of the random part in MAM

MAM_balanced

logical. if MAManalysis function should be called (scaling correction). Can only be used for one product effect, no replications and balanced data.

MAM_adjusted

logical. should MAM be adjusted for the scaling (is only used if MAM_balanced = TRUE)

MAM_alpha_conditional

logical. scaling should be part of the model in case its p-value is less than MAM_alpha_conditional (is only used if MAM_balanced = TRUE)

calc_post_hoc

logical. Should the post hoc analysis be performed for the attributes

reduce.random

logical. Eliminate non-significant random effects according to alpha.random or not. The default is TRUE

keep.effs

a vector with the names which effects should be kept in a model (even if non-significant). By default the assessor effect and interaction between assessor and highest order product effects are always kept.

alpha.random

significance level for elimination of the random part (for LRT test)

alpha.fixed

significance level for elimination of the fixed part (for F test)

interact.symbol

symbol for the indication of the interaction between effects (for presenting in tables and plots). the default is ":".

parallel

logical. Should the computation be done in parallel for the attributes (saves computational time). the default is FALSE

Value

The sensmixedControl function returns a list containing

1. mixed assessor model fitting related control parameters (are used only if MAM=TRUE in the sensmixed function).

2. control parameters for the analysis.

See Also

sensmixed

Examples

# NOT RUN {
str(sensmixedControl())
# }