median_voter {manifestoR}

Median Voter position

Description

The position of the median voter, calculated after Kim and Fording (1998; 2003), with possible adjustment after McDonald 2002.

Usage

median_voter(
  positions,
  voteshares = "pervote",
  scale = "rile",
  groups = c("country", "edate"),
  na.rm.voteshares = FALSE,
  na.rm.positions = FALSE,
  ...
)

median_voter_single(
  positions,
  voteshares,
  adjusted = FALSE,
  scalemin = -100,
  scalemax = 100,
  na.rm.voteshares = FALSE,
  na.rm.positions = FALSE
)

Arguments

positions

either a vector of values or (possible only for median_voter) a data.frame containing a column as named in argument scale (default: rile) and one as named in argument voteshares (default: pervote);

voteshares

either a vector of values or (possible only for median_voter) the name of a column in the data.frame positions that contains the vote shares

scale

variable of which to compute the median voter position (default: rile)

groups

names of grouping variables to use for aggregation, default results in one median voter position per election

na.rm.voteshares

remove observations where voteshares is NA (default: FALSE)

na.rm.positions

remove observations where positions is NA (default: FALSE)

...

further arguments passed to median_voter_single

adjusted

flag for adjustment after McDonald 2002

scalemin

The minimum of the scale of the positions, used for computing the voter position intervals

scalemax

The maximum of the scale of the positions, used for computing the voter position intervals

Details

median_voter is able to compute the median voter positions for multiple elections at once, while median_voter_single treats data as coming from a single election.

calculated according to the formula by Kim and Fording (1998; 2003)

m = L + \frac{K-C}{F} W

Where m is the median voter position, L is lower end of the interval containing the median, K is 0.5*sum(voteshare), C is the cumulative vote share up to but not including the interval containing the median, F is the vote share in the interval containing the median and W is the width of the interval containing the median.

Different parties with the same left-right position (e.g. alliances) are treated as one party with the cumulative vote share.

In the adjusted formula the midpoint is "mirrored" from the midpoint of the other side: "Rather than assuming the party's voters are so widely dispersed, this variable assumes they are spread in a symmetrical interval around the party's position. For example, for a leftmost party at -15 and a 0 midpoint between it and an adjacent party on the right, we assume the left boundary of that party's voters is -30." (McDonald 2002)

References

Kim, Heemin and Richard C. Fording (1998). "Voter ideology in western democracies, 1946-1989". In: European Journal of Political Research 33.1, 73-97. doi: 10.1111/1475-6765.00376.

Kim, Heemin and Richard C. Fording (2003). "Voter ideology in Western democracies: An update". In: European Journal of Political Research 42.1, 95-105.

McDonald, Michael D. (2002). Median Voters: 1950-1995. url: www2.binghamton.edu/political-science/research/MedianVoter.doc