MSM Citations in Republican, Democrat Blogs

March 19th, 2007

A number of qualitative and quantitative analysis of Main Stream Media (MSM) sources have caused heated debates about bias and trustworthiness (or in Stephen Colbert’s lingo shall we say “truthiness“? ;-) ) in MSM. Commentary on news and current affairs once used to be the exclusive prerogative of a handful of political analysts on new channels and sites. Today, blogs and citizen journalism are the new form of punditry. It’s importance is also being recognized by some of the 2008 presidential aspirants.

So, the question is — which MSM sources are going to play an important role on the Blogosphere during the election year? To analyze this we first look at the most cited MSM sources from the ICWSM dataset shown on the right (the complete list is here). Next, we use a list of 113 Republican and 144 Democrat blogs. This list was compiled using a data set provided by Dr. Lada Adamic and by querying Technorati. We count the number of citations for MSM in each of the sets. The MSM sources most frequently cited by democrats and Republican blogs is as follows:

These counts also include multiple citations from each blog. We would like to rank the list in a more meaningful way that would indicate how “influential” an MSM is for a particular group. To do this, we first use KL Divergence based scoring to find the difference in the distribution of citations of MSM in the two groups. For example, a MSM would have a high score in the democrat MSM listing (see below) if it has a high probability of being linked by each of the democratic blogs in the set while having a low chance of being linked to by republicans (and vice versa for republican set). We also modify the scoring function to give importance to citations from multiple distinct blogs (vs. many links from a single blog). The final scoring function produces a ranked list of MSM based on preference of being linked to by either Republican or Democrat blogs. This shows some interesting results (complete list here and here):

Of course, this does not explain any bias of the MSM source itself, but provides a good indication of sources that might influence Republicans and Democrats. Here is a questions I ask our readers: “Bias seems to be quite subjective, according to the side of the political spectrum one may associate with. Do you think it would even be possible to agree on the neutrality of a MSM source?”


  1. The popularity of the blog that links to the MSM is not considered here, and it would be useful to incorporate it.
  2. The results are limited to a small sample of Republican/Democrat blogs.
  3. No content analysis was performed and results are solely based on citations.
  4. The presence of a link does not always indicate influence and we need to use Link Polarity to improve the scoring function.
  5. There is scope for improvement in the ranking function itself. But I think its a first order approximation (to rank distinctively democrat vs. republican MSM preferences).


MSM is influential and there are selective preferences of each community towards different sources. Some of the sources that are categorized under MSM in the dataset almost have a blog like quality. As people rely on blogs for information and opinions, the indirect influence that MSM sources (and perhaps, its biases) can not be ignored. While blogs and MSM seem to almost have a symbiotic relation, (IMHO) this election season might see a fierce competition between the two.

[Acknowledgment: Buzzmetrics for the dataset, Dr. Lada Adamic for the Republican/Democrat labels]