1994 U.S. Senate and Governor's races in Texas.
 Study made possible by the manner in which Bush and Hutchison ran their joint phone program.

1,101,195 Blind i.d. calls to households who met these conditions:

  1. In urban counties, they had not voted in the Democrat primary;
  2. They had not voted in any of the last four Republican primaries;
  3. They had no history from previous phone banks (stretching back to the early 1980's) identifying themselves as Republicans; and
  4. They could have indicated in a previous phone bank that they were a Democrat or Independent who supported at least one Republican candidate.
  • These calls generated over 438,876 households which were Bush and/or Hutchison favorable households and 252, 427 Bush and/or Hutchison undecided households.

  • The campaigns mailed Undecided and Favorable voters with targeted pieces of mail.

Bush:

  • Those phoned and mailed turned out at a rate of 52.09% while those not contacted by the campaign's phone bank and mail program turned out at a rate of 38.96%.

  • The more mail a household received, the more likely it was to turn out, ranging from a 15.45% for those receiving one piece of mail to 50.83% turnout for those receiving three.

  • Turnout peaked at 57.88% for those receiving seven pieces of mail and declined to 52.67% for those receiving eight pieces. 

Hutchison:

  • Those phoned and mailed turned out at a rate of 52.65%, compared to a 38.96% turnout among those not contacted by the campaign's phone bank and mail program.

  • As with Bush, the more mail a household received, the more likely it was to turn out. Turnout ranged from 47.46% for those receiving two pieces of mail to a 52.59% turnout for those receiving three.

  • Turnout peaked at 59.06% for those receiving seven pieces of mail and declined to 56.71% for those receiving eight pieces. By comparison, those phoned and identified by the Bush and Hutchison campaigns as Unfavorables in the five counties studied turned out at a rate of 20.34%.

  • The evidence is overwhelming that a well-run phone and mail program can have a major influence on turnout. While the effect is major (13 points equals a 33-36% improvement in this 1994 example), it is also likely to be significant in a presidential election year.

Karl Rove + Company 1994