Tuesday, August 26, 2008

funny numbers, faulty logic and the "liberal media"

being an econ major as well as somewhat of a stats/math geek, i enjoy the way statistics are portrayed in arguments. take for example the following post the weekly standard arguing an apparent increase in the liberal media bias.

"THE ARGUMENT over whether the national press is dominated by liberals is over. Since 1962, there have been 11 surveys of the media that sought the political views of hundreds of journalists. In 1971, they were 53 percent liberal, 17 percent conservative. In a 1976 survey of the Washington press corps, it was 59 percent liberal, 18 percent conservative. A 1985 poll of 3,200 reporters found them to be self-identified as 55 percent liberal, 17 percent conservative. In 1996, another survey of Washington journalists pegged the breakdown as 61 percent liberal, 9 percent conservative. Now, the new study by the Pew Research Center for the People and the Press found the national media to be 34 percent liberal and 7 percent conservative.

Does this affect coverage? Is there really liberal bias? The answers are, of course, yes and yes. It couldn't be any other way. Think for a moment if the numbers were reversed and conservatives had outnumbered liberals in the media for the past four decades. Would President Bush be getting kinder coverage? For sure,
and I'll bet any liberal would agree with that. Would President Reagan have been treated with less hostility if the national press was conservative-dominated? Yes, again. And I could go on."

so according to his data, the ratio of liberals to cons has increased from just above 3 to 1 in 1971 to almost 5 to 1. however, lost in his numbers argument is the fact that the percentage of those polled who declared neither liberal or con has increased from 30% in 1971 to 59% at the time of the poll. that 59% outnumbers the current "liberal media" by a ratio above 2 to 1. if, to simplifying this argument, we to assume that each of the parties stick up for their beliefs in equal proportions. i.e. that all have different views exert equal influence in using their media power to propagate their political agendas, then it would seem that what we see is not an increase in left-winged media but a substantial increase in media which shares neither a left nor right wing bias.

on a personal note, the economist in me that attempts to measure what people do more than what they say they do finds some potential flaws in this form of data. at times there seems to be trends where it is not correct to admit to certain beliefs which are strongly held, for example witness the change in the discussion of faith in politics over the last 15-20 years. but i didn't write this to criticize the accuracy of polls in portraying true public opinion, but rather to show how numbers were being manipulated to portray a political point.

regardless of you political persuasion, i hope you see the errors in what was argued in the post.

NR


the full article

No comments: