output.design: Summary of Factor Designs

Description

These functions return useful summaries of a factor design, including the design matrix itself as well as other parameters and a list of random factors as experimental units.

Usage

output.design(
  design,
  group_by = NULL,
  order_by = NULL,
  randomize = FALSE,
  rename_random = TRUE
)
design.formula(
  design,
  contrasts = NULL,
  expand_contrasts = !missing(contrasts),
  interactions = TRUE,
  intercepts = TRUE,
  response = "dv",
  env = parent.frame()
)
design.units(design, rename_random = TRUE, include_interactions = FALSE)
design.codes(
  design,
  group_by = NULL,
  order_by = names(random.factors(design, include_interactions = FALSE)),
  randomize = FALSE,
  rename_random = TRUE
)

Value

output.design returns a list containing all output summaries, including the following named entities:

codes: Either a tibble with all experimental codes or a list of tibbles of experimental codes. The list entries are matched to the rows of $groups.
groups: If grouped, contains a tibble in which each row represents an output group, matched to the entries in $codes. If not grouped, this is NULL.
ordered: If ordered, contains a vector of order criteria. If not ordered, this is NULL.
randomized: Value of randomized.
units: A list of random factors and their levels for this design as tibbles. Empty list if no random factors in the design.
formulas: A list of possible model formulas for use with functions such as lm() and lmer().

The functions design.codes, design.formula and design.units only return the values of the fields codes (a tibble or list or tibbles of experimental codes), formulas (a list of model formulas), and units (a list of random factors and their levels), respectively.

Arguments

design: The factorDesign object to summarize.
group_by: If not NULL, the design matrix is grouped by these factors. Factors must be valid columns of the design matrix. If used, $codes will be a list matched to the entries in $groups.
order_by: If not NULL, output within each output group is ordered by these columns.
randomize: After ordering, remaining rows in the same order rank are randomly shuffled.
rename_random: Should random factor levels be renamed? If TRUE, levels are renamed as strings composed of the factor name and factor level (e.g., Subj01, Subj02, ...). FALSE disables renaming of random factor levels. Alternatively, you may provide a function which should accept the vectorized ID (integer) as a first argument and the name (single character value) of the random factor as second argument or ignore it. Functions such as as.double or as.integer *are* possible because they ignore the second argument and only convert the ID.
contrasts: The contrasts to override (NULL if none to override)
expand_contrasts: If TRUE, factors with more than one contrast are replaced by so many contrasts, i.e. the result contains the names of the individual contrasts, not of the factors.
interactions: Should fixed effects be additive or interactive?
intercepts: Should an intercept be included?
response: The left-hand side of the equation. Typically, this is just the response/dependent variable.
env: The environment in which to embed the formula
include_interactions: Whether to include random factor interactions (i.e., counterbalancing factors) in the output

Functions

design.formula(): Retrieve only the model formulas suitable for the design
design.units(): Retrieve only the experimental units of a design
design.codes(): Retrieve only the codes of planned observations of an experimental design

Details

The function design.units returns the experimental units of the design. Those are defined by random factors and their levels. See units return value below.

design.codes returns a dataframe or tibble of all planned observations including each observation's experimental codes, i.e. fixed and random factor levels. If you group the output, a list is returned. See codes return value below.

design.formula returns a list of formulas suitable for regression analysis. Currently, formulas for lm and lme4 are returned. See formulas entry,

Examples

Run this code


des <- fixed.factor("Factor1", c("1A","1B")) +
       fixed.factor("Factor2", c("2A","2B")) +
       random.factor("Subject", c("Factor1"))
       
output.design(des)
design.codes(des)
design.units(des)
design.formula(des)

Run the code above in your browser using DataLab