Median Voters, 1950-1995 Codebook August 2002 Version Michael D. McDonald Department of Political Science Binghamton University (SUNY) Binghamton, NY 13902-6000 UNITED STATES mdmcd@binghamton.edu Acknowledgements. Many of the data entries in this file are derived from data collected and input in a collaborative effort with Silvia M. Mendes and Aida Paskeviciute. 1 Data Set Description The data report estimated left-right locations of median voters of 21 countries at each constitutionallybased election from 1950 through 1995. For each country and election, there are six estimated median voter positions. The six variations arise from differences with respect to two assumptions: (1) the left-right position of parties at the time of an election, and (2) the distribution of voters who are to the left of the left-most party or to the right of the right-most party. Cases are ordered by (1) country, (2) basis of median voter assumption, (3) election. In addition to reporting the estimated median voter positions, the data report information used to create the estimates. The formula can be found in most introductory statistics textbooks under a heading such as ‘computing a median from grouped data.’ HeeMin Kim and Richard Fording have applied the grouped median formula to party position data and party vote percentage data and, in that context, present it as: M = L + {(50 - C)/F} * W. where M = Median voter position L = The lower end (left-right score) of the interval containing the median C = The cumulative vote share up to but not including the interval containing the median F = The vote share in the interval containing the median W = The width of the interval containing the median—i.e., the range of midpoints between the party of the median voter and adjacent parties to its left and right See: Kim, HeeMin and Richard Fording. 1998. “Voter Ideology in Western Democracies, 1946-1989.” European Journal of Political Research 33:73-97. Number of Cases N = 1614 There were 269 elections in these 21 countries. Given six estimates of a median voter’s position per election, the number of cases is 1620 (i.e., 269 x 6). The three elections reported here but not included in the associated Governments data file are two from the French Fourth Republic and the December 1995 Austrian election that did not produce a government until March 1996. Accuracy Status as of August 2002 The data have been checked only once after input. Contact McDonald by e-mail, if you suspect that an entry is in error. 2 Sources We rely on five sources. These are identified here. In the Variable List below, we identify the source of each variable by a short title provided in parentheses following the source citation. Variables that are either created or computed as logical or arithmetic derivations of other data are also identified as such. Budge, Ian Hans-Dieter Klingemann, Andrea Volkens, Eric Tannenbaum, and Judith Bara. 2001. Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945-1998. Oxford: Oxford University Press. (source = CMP98) European Journal of Political Research, annual country political updates (source = EJPR). Mackie, Thomas T. and Richard Rose. 1991. International Almanac of Electoral History, 3rd ed. Washington: Congressional Quarterly, Inc. (source = MR) Volkens, Andrea, Kai-Uwe Schnapp, and Jürgen Lass. 1992. Data Handbook on Election Results and Seats in the National Parliaments of 26 Contemporary Democracies, 1945-1990. Berlin: WZB. (source = Volkens et al.) Woldendorp, Jaap, Hans Keman, and Ian Budge. 2000. Party Government in 48 Democracies (19451998). Amsterdam: Kluwer Academic. (source = WKB) 3 VARIABLE LISTING Var #1 NATION: Country name (source = created) Var #2 NATIONID: Country numeric code (source = CMP) Var #3 MVBASIS: Basis of median voter calculation (source = created) Var #4 ELECDATE: Election date (source = Volkens et al, MR, & EJPR) Var #5 TOTALPCT: Vote percentages of parties for which there are CMP data (source = Volkens et al, MR, & EJPR) Var #6 MDNLEVEL: Vote percentage to identify the median voter, given that only parties for which CMP data are available are used in the calculation (source: computed) Var #7 VOTELEFT: Cumulative vote percentage of the parties to the left of the party supported by the median voter (sources: alignment = CMP; vote% = Volkens et al, MR, & EJPR) Var #8 MPVPCT: Vote percentage of the party supported by the median voter (source = Volkens et al, MR, & EJPR) Var #9 LEFT: Position of the party adjacent and to the left of the median (source = CMP) Var #10 MEDNPTY: Position of the party supported by the median voter (source = CMP) Var #11 RIGHT: Position of the party adjacent and to the right of the median (source =CMP) Var #12 C: The cumulative vote share up to but not including the interval containing the median (source = computed) Var #13 F: Vote share in the interval containing the median (source = computed) Var #14 MIDLEFT: Midpoint between party in median position and party in the adjacent position to the left (source = computed) Var #15 MIDRIGHT: Midpoint between party in median position and party in the adjacent position to the right (source = computed) Var #16 W: Width of interval containing median (source = computed) Var #14 ADD: Term added to lower end of interval (source = computed) Var #15 MDNVOTR: Median voter left-right position (source = computed) 4 VARIABLE DESCRIPTIONS Var #1 NATION: Country name Australia Austria Belgium Canada Denmark Finland France Var #2 NATION: 11 12 13 14 15 21 22 Var #3 2 3 4 5 6 Norway Portugal Spain Sweden Switzerland United Kingdom United States Country numeric code Sweden Norway Denmark Finland Iceland Belgium Netherlands MVBASIS: 1 Germany Iceland Ireland Italy Luxembourg Netherlands New Zealand 23 31 32 33 35 41 42 Luxembourg France Italy Spain Portugal Germany Austria 43 51 53 61 62 63 64 Switzerland United Kingdom Ireland United States Canada Australia New Zealand Median voter basis Uses election-specific party position and Kim-Fording assumption about the distribution of voters Uses three-election moving average party position and Kim-Fording assumption about the distribution of voters Uses post-war average party position and Kim-Fording assumption about the distribution of voters Uses election-specific party position and adjusted Kim-Fording assumption about the distribution of voters Uses three-election moving average party position and adjusted Kim-Fording assumption about the distribution of voters Uses post-war average party position and adjusted Kim-Fording assumption about the distribution of voters Note: the election-specific, three-election moving average, and post-war average indicators of party positions are described in the Parliaments and Governments data set under the variables ELECL_R, MA3L_R, and PTYXL_R. Voter distributions under codes 4, 5, & 6 are assumed to be different from those assumed by Kim and Fording. When the farthest left or farthest right party in a system is involved in the formula, Kim and Fording allow the extreme score of –100 or +100 to mark the endpoint where voters of that party are located. 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. 5 Var #4 ELECDATE: Election date Year and month of election (e.g., 5001 is January 1950) Var #5 TOTALPCT: Vote percentages of parties for which there are CMP data The median voter calculations reported here require that parties be aligned from left to right. Thus, only parties that have CMP left-right scores are involved in the calculations. This variable is used to re-proportionalize vote percentages with respect to parties with left-right scores. Var #6 MDNLEVEL: Vote percentage to identify the median voter, given that only parties for which CMP data are available are used in the calculation Computed as TOTALPCT divided by 2 Var #7 VOTELEFT: Cumulative vote percentage of the parties to the left of the party supported by the median voter Vote percentage of all parties to the left of the party supported by the median voter Var #8 MPVPCT: Vote percentage of the party supported by the median voter Vote percentage of the party supported by the median voter Var #9 LEFT: Position of the party adjacent and to the left of the median Left-right position of the party adjacent and to the left of the party supported by the median voter Var #10 MEDNPTY: Position of the party supported by the median voter Left-right position of the party supported by the median voter Var #11 RIGHT: Position of the party adjacent and to the right of the median Left-right position of the party adjacent and to the right of the party supported by the median voter Var #12 C: The cumulative vote share up to but not including the interval containing the median Computed as [(VOTELEFT / TOTALPCT) * 100) 6 Var #13 F: Vote share in the interval containing the median Computed as [(MPVPCT / TOTALPCT) * 100) Var #14 MIDLEFT: Midpoint between party in median position and party in the adjacent position to the left Computed as [(LEFT + MEDIAN) / 2] Var #15 MIDRIGHT: Midpoint between party in median position and party in the adjacent position to the right Computed as [(RIGHT + MEDIAN) / 2] Var #16 W: Width of interval containing median Computed as MIDRIGHT - MIDLEFT Var #17 ADD: Term added to lower end of interval Computed as {(50 - C)/F} * W Var #18 MDNVOTR: Median voter left-right position Computed as LEFT plus ADD