CMA's nationwide scientific poll: Plurality of Americans "pro-life"

The Christian Medical Association (CMA) and the CMA-managed Freedom2Care coalition on conscience rights recently conducted a nationwide poll on conscience rights. The poll also included a question on respondent's self-identification as "pro-life" or "pro-choice" regarding abortion.
On conscience rights, the poll found:
•77 percent of American adults surveyed said it is either “very” or “somewhat” important to them "that healthcare professionals in the U.S. are not forced to participate in procedures or practices to which they have moral objections." 16 percent said it is not important.
•50 percent of American adults surveyed "strongly" or "somewhat" support "a law under which federal agencies and other government bodies that receive federal funds could not discriminate against hospitals and healthcare professionals who decline to participate in abortions." 35 percent opposed.
On abortion, the survey asked, "On the topic of abortion, do you consider yourself pro-life or pro-choice?"
Kellyanne Conway, CEO, the polling company, inc., explains the results: "Americans were closely divided on the abortion issue, as a 46%-plurality was either firmly pro-life or leaned toward that viewpoint, and 43% self-identifying with a pro-choice affiliation. These data confirm years’ worth of other polling, including for CMA, that shows neither viewpoint claiming a majority of Americans, and slightly more calling themselves 'pro-life' than 'pro-choice.'”
Technical notes: This nationwide survey of 1,000 Americans ages 18 and older was conducted April 29-May 2, 2011, at a Computer Assisted Telephone Interviewing (CATI) phone facility using live callers. The sample was drawn utilizing a Random Digit Dial (RDD) in which phone numbers were generated by a computer to ensure that every household in the nation had an equal chance to be surveyed. Sampling controls were used to ensure that a proportional and representative number of adults were interviewed from such demographic groups as age, gender, race and ethnicity, and geographic region. The margin of error is calculated at +/- 3.1 percent at the 95 percent confidence level, meaning that in 19 out of 20 cases, the results obtained would differ by no more than 3.1 percentage points in either direction if the entire adult population nationwide were to be surveyed.