(CN) – As Donald Trump prepares to accept the Republican party’s presidential nomination Thursday night, a new study purports to show how his candidacy gained momentum over time, even as GOP voters changed their mind about him time and again.
The Pew Research Center first asked Republican voters their preferences for the GOP presidential nomination in March 2015.
At the time, barely 1 percent said Donald Trump, billionaire real estate developer and realty TV show star, as their first choice.
Thirteen months later, in April 2016, Trump was the first choice of 44 percent of Republican and Republican-leaning registered voters, more than any of his rivals.
Today, 88 percent of these voters back him in the general election against the Democratic nominee, Hillary Clinton.
Looked at broadly, the Pew survey numbers appear to suggest that once the Trump train got rolling, it consistently gained momentum.
However, while his support increased in the aggregate with each survey the Pew Research Center did over the course of the primary, researchers found that individual voters’ preferences over this period were actually quite fluid.
These are the conclusions of an analysis of the primary and general election preferences of Republican and Republican-leaning registered voters in Pew Research Center’s nationally representative American Trends Panel.
From March 2015 to April 2016, nearly all Republican voters changed their minds at least once about who they were supporting for the GOP nomination, the researchers said, and this volatility was not confined to the early stages of the GOP nomination contest.
Sixty-one percent of GOP voters changed their minds at least once across the three surveys conducted from December 2015 to April 2016, as key primary contests were unfolding.
And a quarter of Republicans changed their minds twice in this period, the researchers said.
What’s more, only 34 percent of Republican voters supported the same candidate at all three points, including 23 percent who consistently backed Trump over this period — which was by far the largest share “sticking” with any GOP candidate.
The result is a portrait of the fluidity in voters’ decision-making throughout the primary process, the Pew Research Center said.
Some voters changed their allegiance by necessity, when their preferred candidate dropped out of the race. But others shifted their preferences, from survey to survey, among the active candidates.
In the case of Trump, his overall share of support grew throughout the primary process as he gained support from voters who had backed other candidates or had previously been undecided.
But he also lost ground.
Trump held on to 70 percent of his supporters between August and December of 2015, while 30 percent of those who had supported Trump in August went on to support other candidates or to say were undecided in December.
This “churn” in candidate preference was not limited to Trump; all candidates gained, and lost, support during this period.
This volatility continued into the primary season.
Aside from the 23 percent of voters who backed Trump consistently across three surveys between December and April, and the 7 percent of GOP voters who consistently backed Ted Cruz, no single pattern stands out.
By the April survey (conducted April 5-May 2), when only Trump, Cruz and Ohio Gov. John Kasich remained in the race for the Republican nomination, Trump was supported by 44 percent of GOP voters the 23 percent who had consistently supported him since December and 21 percent who were, effectively, newcomers to his cause.
Of the remaining 56 percent of Republican voters, 44 percent had not named Trump as their first choice in any of the three surveys (December, March or April), including 39 percent who had named another candidate in at least one survey (just 5 percent of Republican voters were consistently undecided during this period).
An additional 12 percent of GOP voters had named Trump their preferred nominee in either December or March, but then did not support him in April.
For all this, an overwhelming majority of Republican voters in each of these groups now prefer Trump to Clinton in a general election matchup.
However, the intensity of that support varies.
Nearly all of those who were loyal supporters of Trump since December (98 percent) say they will vote for him against Clinton — and 91 percent are certain they will do so.
By contrast, a smaller share of voters who did not support Trump in any of these three surveys back him in the general election: 79 percent do so, and only 53 percent of those in this group say they are “certain” they will vote for Trump over Clinton.
Throughout the primaries, surveys regularly showed that Trump garnered more support from GOP voters who had not attended college, those who were less religiously observant and men.
The patterns of support over the course of the primaries show that those in these groups were more likely than others to back Trump early on and stay with him.
While 30 percent of Republican voters who have not attended college backed Trump across all three surveys between December and April, just 15 percent of Republican college graduates were early — and consistent — backers of the presumptive GOP nominee.
Republicans who frequently attend religious services were less likely to be steady supporters of Trump in this period.
Just 15 percent of Republican voters who attend services weekly backed Trump consistently, compared with 28 percent of GOP voters who attended services less often.
Conversely, college educated and more religiously observant Republicans are more likely to have not backed Trump in any of these three surveys.
Fully 57 percent of those who attend services regularly fall into this skeptical group, compared with just 36 percent of those who attend less regularly.
Among Republicans with a college degree, about half (53 percent) also did not back Trump in any of these three surveys, compared with 39 percent of those without a college degree.
The American Trends Panel, created by the Pew Research Center, is a nationally representative panel of randomly selected U.S. adults living in households.
Respondents who self-identify as internet users and who provided an email address participate in the panel via monthly self-administered Web surveys, and those who do not use the internet or decline to provide an email address participate via the mail.
Members of the American Trends Panel were recruited from two large, national landline and cellphone random-digit-dial surveys conducted in English and Spanish. At the end of each survey, respondents were invited to join the panel.
The analyses in this report depend upon six separate surveys (fielded in March, August and December 2015 and March, April and June 2016). The data for 5,544 panelists who completed any of these six waves were weighted to be nationally representative of U.S. adults.
For the report, results for December 2015 and later are based on all 2,079 Republican and Republican-leaning registered voters who responded to any of these six waves.
Results for March and August 2015 are based on the 1,345 Republican and Republican-leaning registered voters who were members of the panel at the time.
- Updated Brain Map|Reveals 97 New Regions
- Insurer Must Defend W.Va. Pill-Mill Lawsuit