There seems to be an interesting relationship here…
This is trend chart from our product InfoNgen plotting significant media mentions of both John McCain and Barack Obama for the period July 3rd to November 3rd.
Click Chart To Enlarge
You can see from the chart that Obama tended to have more media coverage than McCain for most periods outside of a brief window following the announcement of Sarah Palin as the Republican Vice Presidential nominee. You can also see a significant Obama spike that happened around the Democratic national convention.
After looking at that chart, I wanted to see how this media centric view to election coverage compared with contemporaneous public sentiment. To do that, I grabbed this chart from Pollster.com showing polling data trends for the two candidates over that same period.
While not identical, there are some interesting similarities between the two charts…
Both charts follow roughly the same shape, though with less overall volatility in the polling data. You can see the same brief crossover for McCain at the same point in both charts, as well as the broadening gap in favor of Obama as the election drew closer.
There isn’t enough evidence here to draw any hard correlations between the two data sets, but the similarities do suggest that the two are not completely unrelated.
There are three possible factors at work here:
- Advertising spend by the two campaigns influenced both public sentiment and media coverage. That would certainly be an interesting additional pair of trend lines in the chart.
- Media coverage directly influenced public sentiment. Media bias has often been cited as a factor in various election cycles.
- The media is simply responding to swings in public sentiment. This could raise some interesting questions about the balance in media between integrity and profits.
No doubt, the truth here lies in some combination of each of these.
But this does point to something broader and very significant…
As we start to collect more and more data points around any complex subject we follow, we’ll start to see trends, patterns, and relationships emerge that were never apparent to us in the past. Some will help to illuminate the causes of past behaviors, and some will even be useful in predicting future behaviors. Our continually improving ability to connect the dots in these data sets is probably one of the most exciting area in information discovery.
Finding the data that is in no one place, but is hidden everywhere…
*Many thanks to Karen Smith, one of my colleagues at InfoNgen, for originally sending me the trend chart of media coverage of the candidates.