mc.get.initial.design: mc.get.initial.design
In MConjoint: Conjoint Analysis through Averaging of Multiple Analyses

Description Usage Arguments Details Value Author(s) See Also Examples

Given a set of cards, determine an initial design including extra cards.

1 2	mc.get.initial.design(full.design, cards = NULL, extra.cards = 3, slack = 1, tol = 0.2, max.trials = 100)

`full.design`	All cards (combinations of levels) that are deemed possible
`cards`	The number of cards in the base design. If this is null then the minimum possible number of cards + 3 (to allow for model fit issues) is used.
`extra.cards`	The number of extra cards added to the base design
`slack`	How much the number of each factor can vary in a "good" design
`tol`	The largest cross correlation in a "good" design
`max.trials`	The maximum number of samples of extra.cards cards to look at.

This can be slow for big designs (lots of samples, a long time testing each one). You can set max.trials to limit the length of time the function runs. It will use the best sample it has seen.

For finer control call mc.get.one.design and mc.add.to.design separately.

A design with cards + extra.cards rows

William Hughes

mc.get.one.design, mc.add.to.design

data(hire.candidates)

#default except max.trials=10 for speed

hire.questionaire = mc.get.initial.design(hire.candidates,max.trials=10)

#A base design of 10 cards with 5 extra cards and good cross correlations less than .17
#takes about 10 seconds

#hire.questionaire = mc.get.initial.design(hire.candidates,cards=10,extra.cards=5,tol=.17)