Meyer and Miller (2015) nicheness of the parties in one election

For every policy dimension a party's squared deviation from the (weighted) mean of its rivals is taken; its nicheness is the root mean square over the dimensions (dimensions nobody mentions are dropped when only_non_zero), minus, with normalize, the weighted rival mean of that quantity. transform = "bischof" applies \(\log(x + 1)\) first. This is manifestoR's single-election Meyer-Miller computation.

Usage

PartyNicheness(
  emphases,
  weights = NULL,
  normalize = TRUE,
  only_non_zero = TRUE,
  transform = NULL
)

Arguments

emphases: Parties x dimensions attention matrix.
weights: Party weights (e.g. vote shares); default 1.
normalize: Subtract the weighted rival mean.
only_non_zero: Drop dimensions with zero total attention.
transform: NULL or "bischof".

Value

Numeric vector of nicheness scores.

References

Meyer, T. M. and Miller, B. (2015). The niche party concept and its measurement. Party Politics 21, 259-271.

Bischof, D. (2017). Towards a renewal of the niche party concept. Party Politics 23, 220-235.

Examples

PartyNicheness(rbind(c(10, 0), c(4, 6), c(2, 8)), normalize = FALSE)
#> [1] 7 2 5