13 Comments

1. 

   acilius

    /  January 21, 2010

   The Residents video reminded me of similar productions I saw in the 80s. So I got nostalgic for that time. Then again, those videos seemed to appeal to nostalgia for 50s sci-fi movies, especially for watching 50s sci-fi movies on TV in the 70s. It’s funny how video turns time in on itself that way.

2. 

   cymast

    /  January 21, 2010

   I much prefer the campy violence of the 50s sci-fi movies to the realistic violence of modern movies.

   Vinnie-P’s convulsing Christian is notable. When I first saw him in the video, I thought he was naked except for his belt and shoes. That, coupled with the onlooker casually observing him, seemed like a torture scene. I thought maybe he was stripped and drugged, and now he’s seizuring. I wonder if there have been any studies done in which the brain activity of these people- while convulsing- is monitored, then compared to the brain activity of people convulsing for non-religious reasons. Perhaps the wires and machinery would be inhibitive. But maybe there’s a way to do it wirelessly.

3. 

   acilius

    /  January 21, 2010

   I also prefer silly pretend violence to realistic pretend violence.

   There have been a bunch of studies like the one you suggest. For example, this one and the one described here. I get the impression that the topic is in a very early stage of development.

4. 

   cymast

    /  January 21, 2010

   Wow, so a lot of seizure-prone Christians actually agreed to be fitted with electrodes for the purpose of analyzing their religious seizures? Christians never cease to amaze me.

5. 

   acilius

    /  January 21, 2010

   Well, there are billions of Christians on earth, so it shouldn’t be too surprising that scientists could talk a couple of dozen into going along with their study.

6. 

   cymast

    /  January 21, 2010

   Only a couple dozen? That doesn’t seem like a large enough sample at all.

7. 

   cymast

    /  January 21, 2010

   Besides, we’re talking about the Christians who actually supposedly have faith-induced seizures. I somehow don’t think those specific Christians would be too keen on being “scientifically tested.”

8. 

   acilius

    /  January 21, 2010

   Depends on what you’re doing. For a telephone survey, you might need hundreds of respondents; for an experiment like that, 10 or 15 might be enough. It depends on how much control the researcher has over the situation.

9. 

   cymast

    /  January 22, 2010

   For example, in a sample size that small, depending on the test subject selection process, some of the people may be related. That would possibly skew the variables and result in flawed data. In science, the bigger the sample size, the more reliable the results.

10. 

    acilius

     /  January 22, 2010

    “In science, the bigger the sample size, the more reliable the results.”
    That’s what I used to think. I said something like that to Mrs Acilius when she was taking a graduate level class on sample size in sociology, and she explained why that isn’t necessarily the case. A few weeks later, I was at a meeting with faculty members from various departments. The topic of sample size came up; if Mrs A hadn’t given me that explanation, I’m sure I would have said the same thing there. As it was, I knew enough to know I didn’t grasp the issues. But someone else did make a sweeping generalization about bigger samples being better, and a psychology prof who’d written a couple of books about the subject explained how under certain conditions a group of 10 subjects can produce more reliable results than a group of thousands.

11. 

    cymast

     /  January 22, 2010

    “under certain conditions a group of 10 subjects can produce more reliable results than a group of thousands.”

    *under certain conditions*

    Back to the convulsing Christians experiments, and MOST other science experiment: “In science, the bigger the sample size, the more reliable the results.”

12. 

    acilius

     /  January 22, 2010

    As I recall, the key point about sample size is that the best sample size for a study is determined by the number of statistical variables you need to control for, and the amount of variation in the subject population on those variables. So there really is no number that applies to “most science experiments.” One field tends to differ from another in the sample size of the average study, and within fields sample size can vary dramatically from one study to another. According to Mrs Acilius when she was taking that class, sociologists tend to look for sample sizes in the low hundreds when they are doing telephone surveys and in the mid-dozens when they are doing interviews; smaller numbers make it hard to separate one variable from another, but bigger sizes just add paperwork, not validity. According to the psychologist I heard from shortly after, psychologists rarely find any difference in results when the same experiment is conducted with a sample size of 20 and repeated with a sample larger than that. Brain anatomists, she said, find that even smaller sample sizes are usually sufficient to produce results that can be replicated consistently.

13. 

    cymast

     /  January 22, 2010

    “So there really is no number that applies to ‘most science experiments.'”

    Well I agree with you there.

    “sociologists tend to look for sample sizes in the low hundreds when they are doing telephone surveys and in the mid-dozens when they are doing interviews; smaller numbers make it hard to separate one variable from another, but bigger sizes just add paperwork, not validity”

    So you’re saying sample size should be based on whether the (in this case) sociologist is doing telephone surveys or interviews? That doesn’t make sense. If that was the case, why not just make it easy and just do interviews? Telephone surveys are very easily translated into interviews.

    “psychologists rarely find any difference in results when the same experiment is conducted with a sample size of 20 and repeated with a sample larger than that”

    So I guess if one psychologist presents results from an experiment in which 10 of the 20 test subjects are related, and another psychologist presents differing results from a duplicate test in which no test subjects are related, we get to choose which results are valid?

    “Brain anatomists . . find that even smaller sample sizes are usually sufficient to produce results that can be replicated consistently”

    For brain anatomists, that sounds reasonable.