A new poll by Public Policy Polling finds “record support for legalizing marijuana,” with 58% of Americans in favor and 39% opposed.

(Kudos to PPP for measuring intensity of opinion. People with strong opinions are evenly divided on the issue: 33% strongly support legalization, 34% strongly oppose it. Another 25% support the proposal, but not strongly; and 5% are opposed, but not strongly.)

But these results differ substantially from an evenly divided public reported in two other post-election polls:

• The ABC/Washington Post poll, conducted in the week after the election, reported 48% of Americans in favor, 50% opposed.
• The CBS poll, conducted a week after that, found an even split – 47% in favor and 47% opposed.

Is it possible that a divided public could, in the space of less than two weeks, be transformed into one with a solid majority in favor? Or are we seeing differences among polls that are loosely described as “house effects” (not real changes in public opinion, but differences due to the polling organizations’ methodologies)?

The PPP poll is instructive, because it shows that among people with intense opinions, views are evenly divided. The large majority support that PPP reports comes from people with weak opinions (the 25% who support legalization, but not strongly), people who are typically not too engaged in the issue and thus quite willing to take a contrary position with a shift in the breeze. Their opinions could, presumably, change quite  quickly.

Still, the “house effects” argument seems persuasive, when we examine past polls on the legalization of marijuana. Below are the results of polls conducted in the past three years.

Compare above ABC/Washington Post's, CBS', and Public Policy Polling's poll on public approval of legalizing marijuana.

Note the major differences among polls, even when they are conducted in a close time frame.

These differences make it difficult to fix on any given picture of the public. Still, there is a consistent trend among polls that have polled the issue over the years, all showing increasing support for legalizing marijuana. The graph below by Gallup, with the last entry in October 2011, is indicative of the trend, even if the actual point estimates may vary among pollsters:

(Credit: Gallup, screenshot)

I’m most inclined to accept the rough picture provided by PPP – that about two-thirds of the public is engaged on the issue and is evenly divided, while another third don’t really care very much one way or the other.