Those "Undemocratic" Party Conventions Were More Democratic Than You Think

News at Home

Mr. Nevin is a doctoral candidate in American history at the University of Virginia and is writing a dissertation about the role of public opinion polls in American politics in the 1960s.

With it looking more and more likely that the choice for the Democratic nominee for President this year will be made by a small group of party elites called superdelegates, and not by the party rank and file through popular primary elections, critics have been lining up to criticize the Democrats for the persistence of this patently antidemocratic element in their presidential selection process.

To many critics the superdelegates are anathema to the ideals of representative government and democratic choice. They are relics of the period before the adoption of the primary system in the 1970s when unaccountable and self serving party bosses met behind closed doors to bargain over the selection of presidential candidates without regard to the wishes of party members.

But was the previous presidential nomination system as unrepresentative as critics contend? Were the so-called closed conventions of the past really so dismissive of the popular will?

A consideration of the relationship between public opinion polls and the selection of presidential candidates under the former nomination system suggests that the closed convention was not so closed after all.

Before the current primary system enabled party members to vote for their preferred presidential candidate, the peoples’ preferences were transmitted to party elites through public opinion polls.

Since 1936, the Gallup Poll has conducted polls throughout the pre-nomination period of presidential elections to determine the presidential preferences of party members. These party preference polls ask self-identified party members to indicate their choice for the party’s presidential nominee among a list of likely nominees.

Between 1936 and 1968, the last year in which most of the delegates to the presidential nominating conventions were not bound by primaries, Democratic and Republican party leaders almost always nominated the most popular candidate among the party rank and file to be the party’s presidential candidate.

17 out of the 18 presidential nominees for the 9 presidential elections that took place between 1936 and 1968 were the preferred candidates of the party rank and file as indicated in the last Gallup party preference poll before the convention. Some of these were uncontested nominations. Presidents Franklin Roosevelt in 1936, 1940, 1944, Dwight Eisenhower in 1956, and Lyndon Johnson in 1964 ran for re-election with little or no active opposition. In one instance there was no clear favorite among the rank and file. In 1964, the final pre-convention Gallup Poll on the Republican side showed both Barry Goldwater and Richard Nixon, who was not an active candidate, to be the preferred nominees of 22% of Republicans, Henry Cabot Lodge, Jr., 21%, and William Scranton 20%. Conservatives, having seized control of the party machinery, pushed through their man Goldwater, who went on to lose big to Johnson in the general election.

Nevertheless, there was only one instance when party leaders from either party failed to nominate the most popular candidate during this period. In 1952, Democrats nominated Adlai Stevenson even though the final pre-convention Gallup Poll showed him running a distant third (12%) behind Estes Kefauver (45%) and Alben Barkley (18%) among Democratic voters. Stevenson lost to Dwight Eisenhower in the general election.

This record suggests that party preference polls served to inform and constrain the choices made by party elites during this period. Polls instructed the party leadership and membership alike about who the most popular candidate was. It simply made no sense for party leaders to nominate presidential candidates who were not popular among the party faithful. Party leaders want to win elections and they know that if they select an unpopular candidate they risk alienating party members, who then might sit out the general election--or worse still, vote for a candidate from another party.

Although the party rank and file had no direct role in the selection of their party’s presidential candidate under the old boss-controlled nomination system, the popular will, expressed through public opinion polls, snuck in to the convention hall through a backdoor and exerted a powerful, though subtle, influence over the decision making process of political leaders.

There is every reason to believe that the polls will regulate the choice of the superdelegates in a similar fashion. While Hillary Clinton and Barack Obama are busy wooing the superdelegates at the Democratic convention, you can bet that the polls will insure that the backdoor to the convention hall remains open.

comments powered by Disqus

More Comments:

sell mbt shoes cheap - 11/21/2009

For sell mbt shoes man is man and master of his mbt walking shoes fate. Have an aim in mbt anti shoes life, or your energies will all be wasted. What makes life dreary is the want of mbt uk motive. When the fight begins within himself, a man's mbt shoes worth something.

ugg boots on sale - 10/31/2009

An ugg bailey button better be unborn than untaught,for ignorance is the root of misfortune.Genius17 withoutugg knightsbridge education is like silver in the mine. ugg boots man is not made for defeat. an uggs boots man can be destroyed but not defeated.No rational man can die without ugg lo pro button? uneasy apprehension.

Lawrence Brooks Hughes - 4/1/2008

None of the cheating on polls would work if the stupid public did not in some measure desire more to vote for a winner than to vote for a candidate or position it approves. Sadly, that is human nature and will not change.

In addition to the wording of the
questions, there is arbitrary skewing of the samples taken, as for "likely voters." Since you must discount 18-25 year olds because a lower percentage of them actually vote than older people--the cheating area becomes by how much you discount such voters. A legitimate pollster will keep his adjustments the same from month to month, so his results are at least consistant with his own previous polling. But an unscrupulous pollster can make the results anything he likes by changing such adjustments right and left. Yet the legitimate pollster must change his adjustments from time to time also, to recognize modifications in public behavior--which makes it hard to prove cheating.

Unfortunately, most of the people who go into journalism are advocates of certain liberal positions, and few can resist the opportunity to help realize the outcomes they prefer.

Sometimes the scribes don't cheat on purpose. You can take TIME Magazine's pre-election forecasts every two years, adjust carefully for liberal bias, and be very close to the actual results every time. (I take their predicted net change in the Senate and swing it by plus two seats for the GOP.) This works because they have subconsciously included personal prejudice into each race they call. It would not work if they ever changed the composition of their staff, but there is no likelihood that will ever happen.

vaughn davis bornet - 4/1/2008

I instantly thought that doctoral candidate Mark D. Nevin had something with his essay on polls.

When researching my Johnson book I found a communication from Harris to an LBJ insider that gave rise to this sentence or two in my book: "Pollster Harris confided to the White House (improperly?) that it would crystalize the situation if Johnson would declare [candidacy in 1968] as soon as possible." Footnote was to Panzer to President, March 30, 1968.

Elsewhere, there is a long paragraph about the influence of a Harris poll on public opinion toward Vietnam escalation which the LBJ group was taking seriously. Harris admitted behind the scenes--to the White House--that the Chinese and Soviet embassies often checked with him on meaning of his polls.

I guess a pollster has freedom of speech, unrestricted by his occupation. But I can't but wonder if close and constant connection with an incumbent president and/or his staff can possibly have no influence on what is polled, and Exactly How.

When in very early 1968 a Harris poll showed the American public very hawkish on the war, favoring its expansion in dramatic ways, it could have had very important results. I wrote: "One aide told Johnson that the forthcoming State of the Union message shuld be influenced by this poll, and results of the poll should be distributed far and wide." [A Panzer letter in the Confidential File.] I had occasion to reuse this paragraph in a new essay on the War, recently, for the poll continues to resonate in my mind, even now.

By all means: research the polls and the pollsters. Though a poll was totally wrong in 1936 (that notorious telephone poll, at least), people like me use today's Pew results routinely, but a bit blindly, I fear.... Just HOW is timing decided? Wording is not neutral; how about that? Anybody can see that in today's climate the right poll can seriously influence Events in the Obama/Clinton race. It's all unregulated, isn't it?

Vaughn Davis Bornet

Robert Lee Gaston - 4/1/2008

In conclusion it won’t hurt (much), it will make your nails harder and it will clear up your complexion.

Lawrence Brooks Hughes - 4/1/2008

I think you could do a darn good thesis proving that the HARRIS poll was working for Hubert Humphrey in 1968, and contributed significantly to the tightening of the race at the end via unscrupulous questions, interpretations, headlines, and publishing dates. The Harris organization always seemed it was working hard for the Democrats in the 1960's and 1970's, behind a phoney mask of impartiality. Other polls followed suit over the years, of course, and usually for the Democrats, but we have seen more balance recently. I think they are still mostly run by liberals, (as, say, Pew), but are more afraid of being unmasked for advocacy now. Pew of course got caught for promoting McCain-Finegold. (I think the WSJ called it "The Stench from Pew.") But it always seemed to me that the Harris organization pioneered many of the dastardly tricks later used by many if not most of the pollsters.

I generally agree with your conclusions above, and also believe we would have had better candidates in recent years with the old system.