statistics

  • Critical Thinking with Statistics, #1

    Seen at Eduwonkette: "You know you’re in trouble when the number of authors on the paper exceeds the sample size." Tip from Joanne Jacobs. Read more


    by

    in

  • The straight scoop on antidepressants

    Mark Liberman has an outstanding article about the recent meta-analysis of anti-depressant effectiveness over at Language Log.  Not only does he use clear and succinct language to explain what’s going on, he has some simple and effective graphics to back it all up.  Check it out. Update (2 March).  Liberman is on a roll.  In… Read more


    by

    in

  • Learn all about polticial polling

    AAAPOR President Nancy Mathiowetz is a new blog contributor to the Huffington Post, and has a great debut article about an online course at NewsU: "Understanding and Interpreting Polls."  The course is informative and entertaining, registration is free, and what are you doing lollygagging around here for? Go  register and get smart! Tip from Mark… Read more


    by

    in

  • Ha Ha, RoundEye–You have squeaky voice!

    Here’s a great post at Language Log about gender pitch differences in Japanese speech…which gives way to a great discussion of why we should be wary of "scientific" findings based on small sample sizes. Update (12 November).  Language Log follows up with another great post, this one on the use and misuse of brain imaging. … Read more


    by

    in

  • One Big Fat Polling Project

    The latest update from the Disclosure Project describes the effort required to poll 6,357 Republican voters across five states in just two nights.  My ears hurt just from thinking about the phone calls. Read more


    by

    in

  • The Pollster.com Disclosure Project

    My favorite Democrat, pollster Mark Blumenthal, has been "nitpicking" opinion polls for some time now, and explaining the methods and shortcomings of polls with a great series of posts.  Now he’s formalizing his critiques: So I have come to this conclusion: Starting today we will begin to formally request answers to a limited but fundamental… Read more


    by

    in

  • Not so fast with that p-value, there, buddy…

    Don’t believe everything you read:  Why Most Published Research Findings are False. Tip from mobathome, commenting on the Volokh Conspiracy. Update (15 September).  FuturePundit weighs in. Read more


    by

    in

  • Mean vs. Median

    Tom Maguire has a great exposition on the difference between mean and median.  Cool. I’m using it in my new Applied Statistics class. Tip from the Instapundit. Update (14 August)  Looks like I had my 15 seconds of fame yesterday, and blew it.  Here’s a bar graph depicting the distribution of dance partners in Maguire’s… Read more


    by

    in

  • A Rose by any other name would be a mathematician

    Chew on this: He [David Figlio, professor of economics at the University of Florida] also showed how harmful giving your child a ‘chav’ or lower-status name can be. In a study of 55,000 children, the exam marks of those with ‘lower-status’ names – often spelled in an unusual way or including punctuation – were on… Read more


    by

    in

  • Do-it Yourself Tag Clouds

    Thanks to a savvy reader at Pollster.com, I’ve discovered TagCrowd, an online application for generating word-frequency tag clouds. Janet Harris generated some from last night’s Democrat debate. Cool. Update (7 May).  The folks at Many Eyes have their own version of a tag cloud, with lots of features. Read more


    by

    in