strategy_multiattribute {multinomineq}

Strategy Predictions for Multiattribute Decisions

Description

Returns a list defining the predictions of different choice strategies (e.g., TTB, WADD)

Usage

strategy_multiattribute(cueA, cueB, v, strategy, c = 0.5, prior = c(1, 1))

Arguments

cueA

cue values of Option A (-1/+1 = negative/positive; 0 = missing). If a matrix is provided, each row defines one item type.

cueB

cue values of Option B (see cueA).

v

cue validities: probabilities that cues lead to correct decision. Must be of the same length as the number of cues.

strategy

strategy label, e.g., "TTB", "WADD", or "WADDprob". Can be a vector. See details.

c

defines the upper boundary for the error probabilities

prior

defines the prior distribution for the error probabilities (i.e., truncated independent beta distributions dbeta(prior[1], prior[2]) )

Value

a strategy object (a list) with the entries:

pattern:

a numeric vector encoding the predicted choice pattern by the sign (negative = Option A, positive = Option B, 0 = guessing). Identical error probabilities are encoded by using the same absolute number (e.g., c(-1,1,1) defines one error probability with A,B,B predictions).

c:

upper boundary of error probabilities

ordered:

whether error probabilities are linearly ordered by their absolute value in pattern (largest error: smallest absolute number)

prior:

a numeric vector with two positive values specifying the shape parameters of the beta prior distribution (truncated to the interval [0,c]

label:

strategy label

Examples

# single item type
v <- c(.9, .8, .7, .6)
ca <- c(1, -1, -1, 1)
cb <- c(-1, 1, -1, -1)
strategy_multiattribute(ca, cb, v, "TTB")
strategy_multiattribute(ca, cb, v, "WADDprob")

# multiple item types
data(heck2017_raw)
strategy_multiattribute(
  heck2017_raw[1:10, c("a1", "a2", "a3", "a4")],
  heck2017_raw[1:10, c("b1", "b2", "b3", "b4")],
  v, "WADDprob"
)