Is there a way to compute the correlation between media coverage and....

[Lord Xar]

I hear alot of how people say "IF he was doing better in the polls, the media would cover him..."

YET, this then raises the question. What came first. The chicken or the egg.

I tend to believe that immediately the MSM had chosen WHO is wanted. Whatever reason that may be, ratings, or agenda.

SO, I know someone here has the media numbers from the candidates for the past month or months. I wonder if there is an algorithm that can determine the number of stories - positive and the rise in polling numbers and possibly the rise in negative ads, and a reduction in polling numbers.

Do you see what I am getting at? I am trying to figure out if such a correlation exists, and therefore a 'manipulation' of the presidential process would have taken place.

Also, there was a chart that in the 1980's over 50 different interests controlled the all media avenues. NOW, only 5. So, I am curious if in the 80's and previous, more liberty minded or "people" candidates had more influence rather than now, when a FULL AGENDA can be subtlly pushed via MSM thru these '5' owners/special interests.

[matratzac]

closest thing to it:
http://www.ronpaultimeline.com/

[hawkeyenick]

It's too hard to quantify, but much like carbon dating, you don't throw it out because of a few outliers.

There should be a way to correlate them, but only loosely.

[drfm]

In google trends, type ron paul and all the candidates separated by a comma like ron paul, mccain, huckabee, thompson, giuliani . Then select US region and select ur timeframe. U will get 2 graphs, the top one shows online search volume and the lower one quantifies media exposure.

U will find that thru out last year, Ron Paul was the most searched candidate but had the least media references (almost zero) of them all.

What even google trends does not show is that even those media references were on the most part negative trying to imapart a negative impression of Dr. Paul.

Also notice that the Huckabee surge was preceeded by a spike in media exposure. Ron got far less media exposure than Huckabee who was behind him both in fundraising and polls at that time. So to say that he was ignored because of polls is misleading to say the least.

[thoughtbombing]

What would the American Cancer Society do for work if, all of the sudden, there was no more cancer?

What would the media do for work if, all of the sudden, there were no more wars... no more scandals... no more police state... ?

Now you know.

[rodent]

I hear alot of how people say "IF he was doing better in the polls, the media would cover him..."

YET, this then raises the question. What came first. The chicken or the egg.

I tend to believe that immediately the MSM had chosen WHO is wanted. Whatever reason that may be, ratings, or agenda.

SO, I know someone here has the media numbers from the candidates for the past month or months. I wonder if there is an algorithm that can determine the number of stories - positive and the rise in polling numbers and possibly the rise in negative ads, and a reduction in polling numbers.

Do you see what I am getting at? I am trying to figure out if such a correlation exists, and therefore a 'manipulation' of the presidential process would have taken place.

Also, there was a chart that in the 1980's over 50 different interests controlled the all media avenues. NOW, only 5. So, I am curious if in the 80's and previous, more liberty minded or "people" candidates had more influence rather than now, when a FULL AGENDA can be subtlly pushed via MSM thru these '5' owners/special interests.

This is very doable and quantifiable. Suppose there are two sources of time-series data with n points:

pollingnumbers[1 through n], and msmarticlecount[ 1 through n].

For pollingnumbers[n], plot it against msmarticlecount[ n - timelag ]. For for example, say month twelve's polling numbers was pollingnumbers[12] -- plot that against msmarticlecount[ 12 - 1 ] (one month ago). Do that for each month on a graph, with msmarticlecount on the x axis and the polling numbers on the y axis. You can use Excel and try to see how well the model fits. Try a few different time lags and look at the correlation between the two data sets. You're going to want to try a few basic regression models.

So the best way to do it, I think, would be to take each of the Rasmussen poll numbers for each of the candidates on a given date range and plot it against the number of articles/stories on google for the previous month, or week. Plot the regression, and see how much variance in the polling numbers is affected by the volume of the press, without regard to the good/bad or quality. You can then further quantify the msmarticlecount series into good press, bad press, etc.

The hardest part is not the data analysis; rather, it's coming up with some software or program that goes through google news or some other newsfeed systematically and categorizes whether the article mentioned a particular candidate.

-Statistics student