Posted by shar on January 22, 2003, at 0:30:37
In reply to Re: Research--Mitchell » shar, posted by Mitchell on January 20, 2003, at 21:13:22
>
>In the design of an experiment, it is acceptable to examine one.........do you mean "one" (the numerical value, like one and only one) or "a" hypothesis?
>hypothesis that is not supported by prior research, unless human subjects are involved. When human subjects are involved, it is still okay, if subjects are not harmed and the research is overseen by an IRB.
.........we agree!
>
> In published findings, the evidence should support the conclusions. A title should accurately summarize content. "Best" in the title of any scientific publication smacks of promotionalism rather than professionalism.........Except, I would guess, if the title had something to do with "best practices" (which we know are not best among *all*, but best among those that the author decided to look at).
>A research paper that concludes something is the best can be expected to qualitatively examine the other side of the comparison. Hsuing explained in a research paper his opinion and his reasoning why banishment is part of the best world, but did not offer evidence that alternative methods had proven unworkable in other settings, nor that anything other than his opinion informed his conclusion.
.......ok, yes, I get what you're saying. Previously I believed you were saying he shouldn't be using "untested" methods in research, period; but you're saying he shouldn't say his methods (tested or otherwise, I suppose) are "best" if he isn't comparing them to something else?
>would probably be more effective if techniques identified in meta-analysis as best practices were not mixed in with techniques for which is available less evidence of efficacy.
>
.........I agree with this statement, but believe that it is nearly impossible to accomplish if research is taking place 'in the field.' It is rare that in a basically uncontrolled setting, a researcher will be able to adhere only to an existing set of techniques that may well not cover all the situations that the researcher must handle.........That's why I believe research publications (journal articles, scholarly reviews, even conference documents and white papers) should be excrutiatingly well-documented. It is possible to be absolutely misled about an experiment or research study, and the outcomes, without such detail. Even with strict adherence to an external set of standards, it's possible to leave out extremely important information about the study that could compromise the reliability and validity of any results. Or, as some might say, the devil is in the details. But, there seems to be a belief in the social science community (starting about 15 years ago) that details are passe'.
........As an aside, dr. bob has control over the techniques he uses and chooses, and he is also dealing with the chaos that makes up real life, and makes decisions about the expression of ideas by humans from all walks of life, and all the complexity of humans and expression and communication--and, there is not likely to be a set of techniques that will satisfactorily address how to deal with all possible events, occurrences, and outcomes he encounters. It is reasonable to consider the setting of any study when analyzing research models and designs, and determining what can effectively be implemented.
> BTW, my preference would be that bob use a machine to delete words like "shit" from his board. My experience is that introduction of vulgar metaphor in open conversation can degrade the quality of interaction. I think use of the word "shit" in published dialogue requires one to reject the preferences of many adults who would rather words like that not be used in conversations children might join.
...........I responded to the "shit" section in a separate post.
Shar
poster:shar
thread:8860
URL: http://www.dr-bob.org/babble/admin/20021128/msgs/8946.html