generateMNLrandomTab: Simple experimental design for choice-based conjoint analysis

In cnchapman/choicetools: Tools for Choice Modeling, Conjoint Analysis, and MaxDiff analysis of Best-Worst Surveys

Simple experimental design for choice-based conjoint analysis

Description

Create an experimental design for an exactly rectangular version of choice-based conjoint analysis (same number of concepts and trials for every respondent). This function is recommended primarily for simple surveys and didactic purposes; it does not do advanced optimization. It will attempt to find level balance through iterative selection across multiple randomized designs.

Usage

generateMNLrandomTab(
  attrLevels,
  cards = 3,
  respondents = 200,
  trials = 12,
  balanced.sample = TRUE,
  best.of = 50,
  verbose = TRUE,
  no.output = FALSE
)

Arguments

attrLevels	A vector of integers, where each integer is the number of levels for its respective attribute. For example, if a product has 3 brands, 4 performance levels, and 3 price points, this would be c(3, 4, 3).
cards	The number of concepts to show at one time.
respondents	The number of designs to generate.
trials	How many tasks will be given to each respondent
balanced.sample	Whether to attempt to do level balancing across attributes, i.e., to have fewer duplications of any level on a given task.
best.of	How many iterations to consider before selecting the optimized one.
verbose	Outputs more information as it works.
no.output	Whether to suppress all output.

Value

A data frame for the experimental design, with a row for each respondent * trial * concept, with columns for the attribute levels.

See Also

[generateRNDpws] to create a set of random utilities for a specific attribute list, [pickMNLwinningCards] to "answer" the survey design according to utilities.

cnchapman/choicetools documentation built on May 28, 2023, 9:14 a.m.