Most research at university tends to be searching for a robust methodology. All I'm saying is that this one is hit and miss. I've read the article reasonably well now:
http://www.polisci.ucla.edu/faculty/groseclose/Media.Bias.8.htm
Essentially the methodology appears to be that politicans are the link from media to think tank. I believe it is telling me that think tanks are allocated an ADA-equivalent score (which is a score which ordinarily measures the degree to which politicians voted left vs right) based on the average score of the politicians which cited those think tanks.
Despite the authors protestations, I don't think this is a good measure of left vs right leaning. I've already noted that it supposed that US = centre (but do note they aren't basing it on the current congress, which is good), when I would say the world is centre.
There are many problems with the causality implied by "quoting" of think tanks. The implication is that a politician or media outlet quotes a think tank due to their own liberal/conservative bias. Other reasons they might quote the think tank might include the effectiveness of the think tank in promulgating its ideas, or the amount of info produced by the think tank.
Its not - based on my quick read - a terrible study, but it isn't beyond question and the results are not particularly strong statistically.