Curiosity maps of MSNBC viewers
Google has a little used feature called “Insights for Search” which will give you an idea which areas of the country have the highest per capita curiosity about particular subjects. Similarity in regional patterns can be amusing.
I am calling these curiosity maps because these are not really the same as viewership statistics that ratings agencies produce. These maps are measuring the incidence of people curious about the search term. After they have satisfied their interest, they do not necessarily keep on searching for the same term. So the patterns can be significantly different than actual viewership.
For fun, I was looking for curiosity patterns that were similar to MSNBC hosts. The following map suggests a correlation between states with populations sophisticated enough to know what a Higgs field, and those intensely interested in what Lawrence O’Donnell has to say. This is just for fun- the similarity in patterns could be for complete different reasons.
Lawrence O’Donnell vs. “Higgs Field”
What should be emphasized here is that this doesn’t indicate popularity. Last Word has good ratings. If there is very intense interest in some areas compared to others, you get patterns like this. It indicates heavily polarized curiosity.
Chris Hayes vs. Santayana
Hayes has less polarization, but departs from typical liberal patterns. He has not just the liberal coasts, but the south and midwest are interested. What sort of non political curiosity patterns are like that? Well- those curious about not-so-common not-currently-trendy thinkers. Like Santayana. Again, maybe this is entirely an accidental co-occurrence.
Melissa Harris-Perry vs. “Magic School Bus” (a children’s PBS show)
MHP. Why no interest in Oregon? Why strong in GA and NC, but zip curiosity from neighbors SC and AL? Why no Louisiana? Why does this pattern resemble that of a children’s educational show? Odd.
Naomi Klein vs. Hercule Poirot
Diving a little deeper into these maps, you notice some common structure. That is, excluding hosts like Matthews and Maddow (those two pretty much touch every state). Rachel does have some regional strengths but is tough. Mary Poppins comes close, but RM is much stronger in the Pacific Northwest, and weaker in the south and Utah than Poppins is.
For the rest of the hosts, there is a base of states that is always included: California, Texas, Florida, New York- maybe this is because they have sufficiently large populations to support subculture diversity. So which states are consistently incurios about what MSNBC has to say? Almost always the mountain and plains states (except CO), Maine Vermont, and Hew Hampshire, Mississippi, Arkansas has as much interest in MSNBC as a Vegan does in BBQ ribs recipes. Scarborough can get Maine, NH and even Miss. and Arkansas. But West Virginia and the northern plains/mountain states can’t relate even to him.
This is not due to some other factor like access to the internet. It is interest. Consider the pattern for George Strait (a country singer). It is almost the complete Yang to the MSNBC Yin.
The Rachel Maddow Anomaly
Ok, I am an engineering kind of guy and so I get obsessively interested in any event that “shouldn’t happen”. A saying attributed to Thomas Kuhn is that “Discovery commences with the awareness of anomaly.” So this is a well known character flaw among scientists and technology folks- we secretly love it when our understanding breaks in spectacular ways. It’s not pleasure of destruction or nihilism or anything like that- it’s because such failures indicate your model for reality is broken. The more broken it is, the greater the possibility you are near something ground breaking. So oftentimes when you fix its broken-ness you open whole new worlds you were blinded to due to flaws in the prism through which you were examining the world. Anyway, the aggregate map of regional curiosity in Maddow is pretty anomalous. One might think at first glance that she is an unlikely host to quickly overwhelm MSNBC’s regional barriers nation wide.
Given the context of pretty consistent regional penetration barriers faced by the other hosts at MSNBC, this is a weird weird map. The second anomaly is the tremendous underlying volatility over time that this aggregate map hides. The map above is the aggregate statistics since 2004 but looking at these “curiosity maps” as they progress in time presents a different, more peculiar story. The google query for maddow has many more maps that can be played as an animation (clickon “View change over time “), but here I select just a few for illustration. Between campaigns, the “curiosity maps” for Maddow looks more like the following periods: Ambivalent in CA, TX, FL and even NY, but a consistently base in WA, OR, WI and MA.
So for some periods she will entirely lose large swaths of states- including even Texas. Then the next period she will surge back like the tide. For example, during the GOP primaries, she had remarkable curiosity from Idaho and Montana, Kansas and even Nebraska, but virtually nothing from Texas.
It would be interesting to know if there is any correlation between her stories and these patterns of curiousity.
Finally, it is important to take note of the phenomenon of Blue city islands in Red States. Take a look at Maddow California versus George Strait California. It’s almost oil and water.
It is tempting to feel warm and fluffy when we see Maddow’s strength in places like Kansas, but which Kansas is open to thinking about information she presents? Is George Strait’s California the same as Maddow World California? There isn’t much intersection except for mixed cities like LA, Sacramento and San Diego. Maddow has Berkeley, Oakland, SF, Santa Barbara and Pleasanton. Strait has Bakersfield, Patterson, Fresno, Stockton and Anaheim.
———– Technical Notes ————–
- Be careful you don’t misinterpret the charts: The shorthand warning for math geeks is: “each state has a normalized scale.” The meaning for civilians is that if two states are the same darkness, it does not mean they had the same number of searches. It is a per capita measurement of the concentration of searches. For example in DC and NY the same number of people out of 10 find Hayes interesting, so they are the same darkness- even though NY had many more searches. Also, the same color in separate maps does not indicate that they had the identical number of searches, so dark blue NY in the Santayana map is not anywhere near the number of searches for Chris Hayes in NY even though they are the same darkness.
- There is something funky with Insight’s numbers. For example, the aggregate chart for Maddow 2004-2012 shows she has moderately strong queries from states MSNBC never gets: West Virginia, Mississippi, Arkansas and the Dakotas. For this to be true, at least one of the shorter periods within this timeframe would have had to have some non zero number of queries, but there are none for any of these states. I can imagine two reasons for this statistical anomaly, so I am not saying something is necessarily bogus. I’m just raising a red flag on the data if anyone wants to use Google insight for analysis like this. I am optimistic Google would allow access to the datasets if a person needed to do a rigorous study.