cjpwr: Simple Power Analysis and Sample Size Diagnostic for Conjoint...
In mbarnfield/cjpwr: Simple Conjoint Design Power Analysis (Orme, 2010)

Description Usage Arguments Details See Also Examples

Johnson's rule-of-thumb calculation for determining power of conjoint designs.

1	cjpwr(n = 100, t, a = 2, c)

`n`	The sample size (defaults to 100 for when you're only interested in finding out necessary value of n and don't know what to specify).
`t`	The number of choice-tasks per respondent.
`a`	The number of alternatives per choice task (defaults to two).
`c`	The number of analysis cells - equal to largest number of possible levels for any one feature, or the largest product of levels of any two attributes for power of two-way interaction estimates (Johnson and Orme, 2003).

cjpwr divides the product of n, t, and a by c, to give Johnson's rule-of-thumb estimation of conjoint design power. It returns a dataframe containing the inputs and result of this calculation, whether (yes/no) this exceeds the minimal minimum threshold (500) and ideal minimum threshold (1000), and the sample sizes (rounded up) necessary for minimum and ideal power thresholds.

pwr_n

#conjoint design with five choice tasks per respondent, two alternative profiles per task, and maximum six levels per feature
cjpwr(n = 1000, t = 5, a = 2, c = 6)
#same design but considering interactions, where largest product of levels of any two attributes is 18 (6*3)
cjpwr(n = 1000, t = 5, a = 2, c = 18)
#without argument labels, order only matters for c
cjpwr(5, 1000, 2, 18)
cjpwr(5, c = 18, 1000, 2)
cjpwr(c = 18, n = 1000, a = 2, t = 5)
#with defaults
cjpwr(t = 5, c = 18)