Announcement

Collapse
No announcement yet.

Error Some Place

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Error Some Place

    A fairly recurring criticism I hear of parapsychology is the lack of a control condition in an experiment. For scientists used to having to sets of data to compare, it seems to worry them that parapsychology often compares itself to the null hypothesis as expected by chance.

    This raises concerns that errors in the method could sway the null hypothesis away from what is expected. This is why Blackmore’s account of her visit to the Cambridge lab is important. The idea that fraud was going on is almost secondary to her description of a convoluted method that wasn’t always followed as laid down in the paper. This is a worry, and it is not overly cautious to consider these as damaging to the paper’s standing.

    The problem with this is that experiments that get good results tend to be the ones that get most attention from sceptics. Can I be sure that the issues raised are genuine, and not just an attempt by sceptics to lessen the importance of a set of results? In the case of Sargent, I do think that the criticisms of his work come from a varied enough set of commentators that they can be taken seriously and that his early work shouldn’t be in the database. Although Sargent’s later work which removed the labelling issue that meant fraud wasn’t possible still stands, I think.

  • #2
    One possible problem is that the statistical model that represents the null hypothesis may be incorrect. For example, in presentiment experiments, the null hypothesis that the subject will guess calm/disturbing at random can't be right, since there must be some sort of anticipation going on. This is so regardless of whether you think that anticipation can explain all the results.

    The results can be skewed both by mistakes in the methodology and by errors in the statistical analysis. When the entire story is a comparison of subjects' behavior to a statistical model of behavior, the results are highly dependent on the accuracy of that model.

    ~~ Paul

    Comment


    • #3
      Ersby,

      Thanks for making the Ganzfeld wikipedia page fairer .

      I might suggest later a minor addition to the wording relating to '3' - as this is more subjective viewpoint than the others in list .... but I'm a busy at the moment.
      Last edited by Open Mind; October 18th, 2009, 10:04 AM.

      Comment

      Working...
      X