This morning, there was this exchange in the comments of his post 'Ears, camera angles, and the "null hypothesis":'
Brad posted 'Ears, camera angles, and the "null hypothesis"' at his own blog, after posting at Political Gates.
From Ghostbuster's comment, the hypothesis is "the images we have access to can prove this point one way or another," and its null hypothesis is "the images we have access to cannot prove this point one way or another."
An alternative hypothesis is "the images we have access to cannot prove this point one way or another," and its null hypothesis is "the images we have access to can prove this point one way or another."
Sharlott wrote of the first hypothesis, "Exactly. That being true, you would be ill-advised to push the argument that there are different children involved. That is precisely the point I want everyone to take from this series I have done."
It appears as though Sharlott concurs with Ghostbuster's assertion that he (Sharlott) presented a "false dilemna."
Would Sharlott say of the alternate hypothesis, "you would be ill-advised to push the argument that there are not different children involved?"
Couldn't Sharlott conclude that it would be ill-advised to decide anything about the baby count with just one of those hypotheses? Either of them? Both of them?
The hypotheses are easier to understand written this way:
1) The images we have access to can prove this point, and its null hypothesis is the images we have access to cannot prove this point.Here is an even easier way to grasp the concept of hypothesis (and null hypothesis). Hypothesis: Otomorphenanan is an effective ear analgesic; Null hypothesis: it is not an effective ear analgesic. It is a concept used in drug trials, but the first time I've seen it used to identify a person was at Brad Sharlott's post.
2) The images we have access to cannot prove this point, and its null hypothesis is the images we have access to can prove this point.
Whatever your opinion of the foregoing may be, there is one very important hypothesis that should have been tested first, before the post at Political Gates appeared: Does ear shape individually identify a person? Like fingerprints? DNA? Even facial features? Isn't the post at Political Gates dependent -- assuming for the sake of argument that the pictures appear to be pictures of the same baby -- dependent on an assumption that ear shape individually identifies a person? Even if that hypothesis could be established with 95% certainty, would the pictures presented at Political Gates be adequate for an analysis under that hypothesis? Are they sufficiently sharp? Are they controlled for camera, lens, sensor, camera and lens settings, lighting, camera distance from the subject, camera shift and camera rotation relative to the subject to draw any conclusions using a hypothesis (if it could be established) that ears individually identify a person?
With these hypotheses, it can be said that we cannot be absolutely sure, using the published pictures, whether there was one baby or more than one baby. And, of course, before using any pictures, we would have to be certain that ears individually identify a person. And then, of course, we would need to be sure that our pictures were adequate for analysis based on that hypothesis.
The cat that should be out of the bag is that Brad Sharlott doesn't know how many babies there were, even though he appeared to be arguing at Political Gates that there was only one baby.
Note: In his post at Political Gates, Professor Sharlott attacked the blogger whose post began the more-than-one-baby discussion. This morning, that blogger expressed his ideas about Sharlott's post, here.
That's my $0.02. I suppose we're going to see more slow Palin-news days.