Dan Chambliss Defines Sampling on the Dependent Variable

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:10

      DANIEL CHAMBLISS: A major problem that a lot of students,and even a fair number of professionals, run intois what we call sampling on the dependent variable.Sampling on the dependent variablemeans you're picking your cases to study on the outcome,on the result of what it is you're thinking about.

    • 00:31

      DANIEL CHAMBLISS [continued]: For instance, you might say you want to study murders, and findout what causes murder.Well, so you go out and through some miracle of personality,you get access to a whole bunch of murdererswho are in prison or something.You're going to talk to them all and try to figure outwhat caused their problems.

    • 00:52

      DANIEL CHAMBLISS [continued]: Why they were bad to people, and you start working backwards.You discover maybe all of these murderers were drug abusers.You go, aha, it's drug abuse.You keep working backwards and you discover, well,they all started out drinking a bit too much whenthey were young and that led to their drug abuse.

    • 01:12

      DANIEL CHAMBLISS [continued]: You think, Ah, that's what it's really about-- drinkingtoo much when you're young.You keep working back and you discoverthat every single one of those murderers-- every single oneof them-- started out when they were young drinking milk.And you think, aha, it's milk.

    • 01:34

      DANIEL CHAMBLISS [continued]: You go, wait a minute.Everybody drinks milk.That's the problem with sampling on the dependent variable.If you start out with the people at the end of the story,you will see a lot of things that they all do,and you might think that's what'sreally causing the problem.What you have to do instead is try

    • 01:55

      DANIEL CHAMBLISS [continued]: to start out by looking at, so to speak, everybody,and then figure out what is it that led to the specific peoplewho wound up say as murderers.That's an example of sampling on the dependent variable.I'll give you another example.Some years ago, I studied athleteswho were training to compete in the Olympics in swimming.

    • 02:20

      DANIEL CHAMBLISS [continued]: I noticed that a lot of these athletesworked incredibly hard at swimming.I mean really, really hard.It was quite noticeable.I thought, maybe that's what makes them great swimmers.And then I went back and I looked at pretty good swimmers,at different levels of the sport.And even people who never really went very far at all.

    • 02:43

      DANIEL CHAMBLISS [continued]: It turns out, a lot of them work hard too.It wasn't the hard work that made the difference.It was other things in the techniques,and in the coaching, and things like that I discovered later.But the point is I couldn't just lookat the successful Olympic athletes,because I would miss all the stuff that theyshared with everybody else.

    • 03:03

      DANIEL CHAMBLISS [continued]: You have to look at both.You have to look at the successfuland the unsuccessful.You have to go back and look from the whole population, notjust the result people.Sampling on the dependent variablecan be incredibly misleading.

Dan Chambliss Defines Sampling on the Dependent Variable

View Segments Segment :

Abstract

Professor Dan Chambliss explains sampling on the dependent variable as trying to determine causation based on a limited sample with similar traits. Without comparing a sample against the wider population, it's easy to draw incorrect conclusions.

SAGE Video Shorts
Dan Chambliss Defines Sampling on the Dependent Variable

Professor Dan Chambliss explains sampling on the dependent variable as trying to determine causation based on a limited sample with similar traits. Without comparing a sample against the wider population, it's easy to draw incorrect conclusions.

Copy and paste the following HTML into your website

Back to Top