In case anybody is wondering what we really spend our time talking about . . .

## High-level intellectual discussions in the Columbia statistics department

Posted by Andrew on 4 June 2011, 9:56 am

In case anybody is wondering what we really spend our time talking about . . .

## Recent Comments

- Rahul on “The Statistical Crisis in Science”: My talk this Thurs at the Harvard psychology department
- Shravan Vasishth on “What then should we teach about hypothesis testing?”
- Mayo on “What then should we teach about hypothesis testing?”
- Mayo on “What then should we teach about hypothesis testing?”
- Mayo on “What then should we teach about hypothesis testing?”
- Martha on The (hypothetical) phase diagram of a statistical or computational method
- Martha on The (hypothetical) phase diagram of a statistical or computational method
- Martha on The (hypothetical) phase diagram of a statistical or computational method
- Martha on The (hypothetical) phase diagram of a statistical or computational method
- Martha on The (hypothetical) phase diagram of a statistical or computational method
- Rahul on The (hypothetical) phase diagram of a statistical or computational method
- Lee Sechrest on Tell me what you don’t know
- Leon on Tell me what you don’t know
- Jeremy Fox on The (hypothetical) phase diagram of a statistical or computational method
- John on The (hypothetical) phase diagram of a statistical or computational method
- zbicyclist on The (hypothetical) phase diagram of a statistical or computational method
- Anonymous on The (hypothetical) phase diagram of a statistical or computational method
- Jeff Walker on The (hypothetical) phase diagram of a statistical or computational method
- Jeff Walker on The (hypothetical) phase diagram of a statistical or computational method
- Martha on “What then should we teach about hypothesis testing?”

## Categories

I'm sure you'll be devastated, but the first person that comes up when I search for "Gelman" is the producer of "Live with Regis and Kelly" (after two links for the Gelman library system at George Washington University). You are right after that, though!

Sorry, but when I search for "Gelman", the first 2 sites I get are related to the "George Washington University's Gelman Library", your web page is 3rd. Personalized search strikes again!

Weirdly enough I routinely navigate to your blog by search "gelman blog", so you'd think my personalized search would be trained by now, but who knows, maybe they're trying not to overfit my search history. I wish there was a way of doing searches with these personalization options off.

Michael Gelman comes up before you do in a google search of 'gelman'. Makes sense that you would show up first when you search it though. There was an interesting TEDTalk about internet filters, which I believe is at play here: http://www.ted.com/talks/eli_pariser_beware_onlin…

Yeah, you're fourth for me (in the Mid-Atlantic), after GW's library, some weird website on the Study Room Reservation System, and Michael Gelman.

I like personalized search though–when I type in the first two letters of my school's library (two very common letters), it spits out my library's name–saves me a lot of typing every day. Of course, I could just put it in my favorites tab up top or whatever, but I've never managed to do it.

This is a great example of what Eli Pariser calls the "filter bubble." Many search engines and websites (such as Amazon) try hard to tailor their search results to what they think you, specifically, are interested in. To do this, they use information including your location, the time of day, other websites you have visited (if their cookies are available to you), your previous search history (if known), and so on. Sounds good, right? Why not use all of the available information to try to deliver the search that is best for you? But Pariser suggests there is a problem: this can create a "search bubble" that makes it hard for you to see opposing points of view. Search for "Pawlenty" and you will likely be pointed to sites that support your existing point of view (whatever it is). And you won't even know it.

Pariser did a TED talk about this that might be worth a look.

Anddrew Gellman or something related to him is in the first 4 hits for me. Michael comes next.

I'm in Canada