Equality Advances, Pie Charts Decline

Posted by jns on 28 April 2009

Via Joe.My.God I learn that a new CBS News / New York Times poll shows an amazing 42% of those polled in favor of legal marriage for same-sex couples. That’s amazing because the previous poll by the same group only one month ago found only 33% in favor of legal marriage. This increase of 9 points lies well outside the sampling error of the poll and must represent some significant fall-out from the addition of Iowa and Vermont to the growing list of states with marriage equality.

That’s interesting enough, but this graphic from CBS (source for graphic and poll results) is also interesting:

One thing to note is that fully two-thirds of those polled support some sort of legal recognition for same-sex couples.

The other thing to note is that there is a 5% wedge missing from the pie chart. Add up the percentages and you get 95%. Now, while it is not at all unusual to have some 5% of those asked who are undecided or who wish to give no opinion, it is not kosher to leave them out of the pie chart and sort of fudge the other wedges around to fill in the space.

Does this chart make it clear that there is a 5% wedge unaccounted for? Not at all. Somehow the wedges are made to look as though they fill 100% of the circle and yet they should not. You might think that 5% is small and wouldn’t be visible, but it amounts to one-fifth of the “civil unions” wedge, an amount that would be quite noticeable. So, the chart is inaccurate and misleading. I wonder which wedge, or wedges, got the undecideds?

Shame on CBS. I hate it when big corporations who could have science and math consultants without even noticing the cost can’t be bothered.
—–
[Updated a few minutes later:] In thinking about the poll results and the remarkable shift since the previous poll, Timothy Kincaid at Box Turtle Bulletin (“Americans Shift Sharply in Favor of Marriage“) gives a long list of significant steps that have been taken towards marriage equality in the interval between the polls.