Spatial vote probabilities

Binary probabilities F((U_1 - U_2) / scale) with normal, logistic, Laplace, Student-t, Gompertz or complementary log-log F; multinomial (conditional logit) or exponential-distance (Luce) choice probabilities; and, for quadratic utility with a normal link, the exact integral over an uncertain ideal point with covariance ideal_cov.

Usage

VoteProbability(
  voters,
  alternatives,
  model = "quadratic",
  link = c("normal", "logistic", "laplace", "student_t", "gompertz", "cloglog"),
  scale = 1,
  df = 5,
  rule = c("binary", "multinomial", "exponential"),
  ideal_cov = NULL,
  ...
)

Arguments

voters, alternatives

Points (matrix rows, list of vectors, or a vector for one dimension).

model

Utility model.

link

Binary link.

scale

Error scale.

df

Student-t degrees of freedom.

rule

"binary", "multinomial" or "exponential".

ideal_cov

Optional covariance of the voter ideal point.

...

Passed to VoterUtility.

Value

list(probability, utility).

References

Poole, K. T. and Rosenthal, H. (1985). A spatial model for legislative roll call analysis. American Journal of Political Science 29, 357-384. McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior. In Frontiers in Econometrics, 105-142.

Examples

VoteProbability(0, c(1, -1))$probability
#>      [,1] [,2]
#> [1,]  0.5  0.5