CDB said:
As was said, that is a study of candidates, not labels. The person who might label them might not feel a need to do so if the interviewee's party affiliation is widely known. This study is a perfect way to lead to irrelevant and/or skewed results. If you want to study labels you need to study the labels, not specific candidates where a lot of other variables are in play.
Let's recap what the study showed: Nunberg searched a database with 20 major US dailies for 10 well-known politicians.
-5 liberals with ADA scores over 90: Senators Barbara Boxer, Paul Wellstone, Tom Harkin, and Ted Kennedy, and Representative Barney Frank.
-5 conservatives: Senators Trent Lott and Jesse Helms, Attorney General John Ashcroft, and Representatives **** Armey and Tom DeLay.
The time period started between 1982 and 1992 depending on when the particular paper became available, and ended in 2002.
Details
Invalid Link Removed.
He then searched for "conservative" and "liberal" within 7 words of the the politician.
Print (and other) media generally always includes the party affiliation for Senators and Representatives so that should not be an issue.
Nunberg later did the same experiment for some actors and judges with similar results. He also checked for possible noise and some other factors, but found that it was about equal on both sides.
So the study showed how often those people were labeled during the available 10-20 year time period in mainstream print media, and looking at the results individually by politician I don't see any liberal bias. It's certainly not definitive proof of no liberal bias, but it shows that Goldberg's thesis doesn't hold up in this seemingly reasonable example and therefore it may not hold up in general. Thus Goldberg's claims are highly suspect and he should offer some empirical proof for them rather than just opinions and isolated incidents.
Which only gives you information, once more, about the candidates in the context of the articles they were mentioned. By nature this study can say nothing about media tendencies in general, only tendencies towards certain candidates in certain newspapers. To find whether or not the media is using a conservative or liberal label you have to look at the labels first and then the context they are used in which will divorce the findings from any variables that looking for specific candidates might introduce.
The sample included 20 major newspapers including all the big ones and they were analyzed for all available content, which was between 10-20 years depending when the newspaper became electronically available. Nunberg also replicated the study for "liberal" papers and the basic result was the same.
Now you could search for the label first, but checking the context for over 100,000 hits in Nunberg's sample would not have been feasible for Nunberg, nor informative without adjusting the results for the ADA or other ratings which would not have been feasible.
He did, however, check several hundred of the hits for context and noticed that the noise (labels not referring to the person) was roughly equivalent on each side, liberal papers vs. all paper did not change basic result, and neither did including synonyms for conservative and liberal make a big difference to the basic finding.
Nunberg is also full of it on another issue in his response to Goldberg's response to him; just because a group like NOW "[rejects] explicitly partisan labels" does not change the fact that their political position is to the left of Marx by a decent shot. The idea that a person or organization cannot be identified as liberal unless they voluntarily take on the label is pure BS.
Yes and no. The country is split about 20-47-32 between liberal-moderate-conservative according to the 2006 midterm exit polls, so as a result most Dems don't want to be identified as liberals or even take fully liberal stands, and instead prefer to reject these labels (like Kerry) or just call themselves moderate, whereas Republicans like to talk up their conservative credentials. McCain comes to mind, last Sunday he was once again making the case of how he has been a good conservative, whereas I don't see that from the Dems when it comes to the liberal label. If you watch the Sunday shows you'll know that's just how it is.
Therefore I'm guessing that "conservative" is used a fair bit to label Repubs because that's how they want to be introduced, whereas Dems not not want to be introduced as liberals. And I think this is the major reason you get the 4:1 ratio in the MRC study. On a side note, I noticed that some in the media even tried to labels some of the newly elected Democrats such as Jim Webb (D-VA) and Jon Tester (D-MT) as conservatives even though they are both to the left of center on economics.
I recall he also criticizes the comparisson of Limbaugh to Rosie O'Donnell when pointing out how people are labeled. All I can say is it's very convenient to make claims when you propose to limit the sample upon which you can base your evidence. The one thing he correctly points out is that some groups and people are labelled more than others, which is rather nitwitish of him since that's exactly why looking for people or groups and how they are labeled can lead to skewed results.
Rosie O'Donnell is a liberal, no disagreement there. I just think Nunberg's point was that she appears in a lot of time in non-political shows during which labeling is not necessary.
But bottom line is if every time Rush Limbaugh is talked about he's labeled as conservative, and everytime rosie O'Donnell is talked about she isn't labeled as liberal much if at all, and this trend sticks across all people and groups, not cherry picked ones or ones that are redefined as middle of the road for sake of convenience despite their political stance, then there is a media bias to the liberal side.
I'd expect O'Donnell to be labeled "liberal" in political discussions as often as conservative actors are labeled "conservative." I don't think it's fair to expect her to be labeled as often as Limbaugh because being a conservative is Limbaugh's identity and job description whereas the same is not true, or at least not to the same extent, with O'Donnell. I'm also guessing that Limbaugh is labeled conservative as often as someone like Al Franken is labeled liberal or progressive since his gig on Air America Radio.
The studies you point to seem to need to do a lot of redefining and evidence picking/slanting before they can get the results they want. But then again this is indicative of the liberal echo chamber they live in. Only a liberal could even contemplate defining NOW as middle of the road politically.
I don't know NOW, but I have no reason to believe that Nunberg cherry-picked his sample. However, if he did then the MRC or some other organization could easily prove him wrong by changing the sample time period, set of mainstream newspapers, set of news channels, and/or set of high profile politicians with similar ADA ratings as in the original study. But at this point the burden is on Goldberg to verify his claims.
And MRC's way of responding to Nunberg was too roundabout to be credible IMO as a response, though independent of that it's an interesting finding.
I seriously doubt that. You take three 'studies' from blatant left wing sites and make excuses with no evidence as to why the one conservative study can't be trusted. Meanwhile the Nunberg study has a blatant flaw in that it concentrates on people and if they are labeled and not on labels and how they are applied when they are used. Nunberg's methodology is so open to variances from so many other factors it's ridiculous. Everything from media time to the paper in which the candidate's name shows up in most often can affect the results.
The MRC study doesn't look at specific people and that's why it's more reliable. Its results are based on finding the labels and seeing how they are applied when they are used regardless of who or what org they are being appended to. That's the only way to measure what is suppoed to be at issue here: how often the media in general feel the need to label conservative people/groups as opposed to liberal ones. The only variable I can see affecting these results is a coverage difference, but even that can't explain a 4:1 ratio of how the labels are applied.
I did. As was said, looking for candidates first as opposed to labels can lead to a lot of other factors affecting the results. What should be done, and what MRC did, was look for the labels and then see how they were applied. The only thing missing is a control for coverage differences, but that was actually mildly accounted for by eliminating repeated references. Bottom line is the methods you are pushing are actually studying something completely different than what's at issue. Looking for names first is only useful if you want information concerning that group of people and how the media treat them specifically, not how the media behave in general on this issue.
Well, I've already stated my problem with MRC study, it doesn't account for how many times liberals and conservatives were actually in the news overall and it doesn't factor in that there are more conservatives than liberals in America and specifically more Republicans who want to be called conservatives than Dems who want to be called liberals. In fact, every election I've seen the Republican has attacked the Democrat as a liberal or too leftwing, etc. If the MRC had just restricted it's search to the labeling of think tanks and their representatives, and normalized for actual appearances, then I'd have no problem with that. At least the MRC should have restricted it's study actual conservative and liberals by matching them with ADA or other ratings.
But there's a larger point here; the motive. As I understand it the idea is that reporters are more liberal and hence the bias. However, reporters only tend to be more liberal than average Americans on social issues, but they tend to be more conservative on economic issues. And the media is owned by corporations anyway which are economically even more conservative and their interests are more aligned with the Republicans than Democrats.
