Question:
I am going to read match-ups between some possible Democratic and Republicans in the November general election for U.S. Senate. For each pairing, please tell me which candidate you would vote for if the election for U.S. Senate were being held today. What if the candidates were (SEE RELEASE FOR NAMES AND PARTIES)? If the election for U.S. Senate were being held today, for whom would you vote?

Boxer - 43%
Campbell - 44%
Undecided - 13%

Boxer - 45%
Fiorina - 44%
Undecided - 11%

Boxer - 45%
DeVore - 41%
Undecided - 14%
Poll Demographics

About this Poll
The findings in this report are based on a Field Poll survey completed March 9-15, 2010 among a random sample of 748 likely voters in the November general election. Additional interviews were conducted among likely voters in the GOP primary to increase this portion of the sample to 353 voters. Interviewing was conducted by telephone in English and Spanish with live interviewers working from Field Research’s central location telephone interviewing facility.
Up to eight attempts were made to reach, screen and interview each randomly selected voter on different days and times of day during the interviewing period. Interviewing was completed on either a voter’s landline phone or a cell phone depending on the source of the telephone listing from the voter file. After the completion of interviewing, the overall sample was weighted to align it to its proper statewide proportions by party registration, region, and other demographic characteristics of the registered voter population.
Sampling error estimates applicable to the results of any probability-based survey depend on sample size as well as the percentage distribution being examined. The maximum sampling error estimates for results based on likely voters in the general election is +/- 3.7 percentage points at the 95% confidence level, while the maximum sampling error of likely voters in the GOP primary election is +/- 5.5 percentage points.
The maximum sampling error is based on results in the middle of the sampling distribution (i.e., percentages at or near 50%). Percentages at either end of the distribution (those closer to 10% or 90%) have a smaller margin of error. Findings from subgroups of the overall sample have somewhat larger sampling error levels. There are other potential sources of error in surveys besides sampling error. However, the overall design and execution of the survey sought to minimize these other sources of error.
#