Hi Mike- In your last post you said, "Maybe we should only analyze the data of subjects whose beliefs were consistent with the treatment condition they were actually assigned to. Someone should at least analyze the relevant groups to see if there is a difference." I think that you may be overstating the case a bit when you suggest that no researchers look at whether subjects think that they got the real drug or the placebo. This is actually quite a common procedure in many experiments. For example:
Insomnia and the attribution process. Storms, Michael D.; Nisbett, Richard E. Journal of Personality and Social Psychology, Vol 16(2), Oct, 1970. pp. 319-328. In their paper the authors state, " In addition, subjects were asked how much they suffered from insomnia on each night and what drug they thought the pills actually contained" (my emphasis) They then later report that, " It is noteworthy that the experimental effects occurred only for subjects who believed the pill descriptions. So, yes subjects do develop hypotheses about what is going on in an experiment and these hypotheses can affect the outcome but to suggest that no one takes this into consideration when designing an experiment is a bit of an overstatement. -Don. ----- Original Message ----- From: Mike Wiliams <jmicha5...@aol.com> Date: Thursday, September 15, 2011 11:22 pm Subject: Re:[tips] Blinded Studies To: "Teaching in the Psychological Sciences (TIPS)" <tips@fsulist.frostburg.edu> > Hello All. > > This is starting to look like a response to journal > reviewers. Rather than make a long list of point-by-point > responses, I would like to just state some points that > might generate a paradigm shift in how we think of outcome > research. > The major point I want to make is one that everyone will agree > with but no one has thought of the consequences. > > This point is that humans are sentient beings who attempt to > understand and figure out which treatment condition they are in. > After they make a judgement about this, they behave consistent > with the social demands of the > research setting and other expectation biases that are > consistent with their judgments. All the research design > guides, such as Campbell and Stanley, assume that the subject is > a passive agent in the process and does > not interact with the research design. Some of the threats > to internal validity suggests something like this but they > really need a chapter called, "How Human Cognition Screws up > Research". > Imagine you could create a placebo that had the same side > effects as the medication but did not have an active > agent. Both groups would have the same side effects. > You then randomly assign the meds and > placebo and the subjects and investigators are not told who got > which one. > > Now, I'm a subject in the placebo condition and I start > experiencing dry mouth and constipation. I infer from this that > I am getting medication. I then endorse positive change on > the Beck Depression Scale because I want to help out > the researchers or otherwise demonstrate an expectation > bias. Some proportion of the subjects in the placebo > condition behave like they were treated on the self-report > measures. This is a scenario in which the side effects are > controlled and the result is still noise. I understand > that some of you will counter that you could inform all the subjects > that they might experience side effects. This is always > done on the consent form. However, it is commonly understood > that placebos do not cause effects and subjects will reason > consistent with this. There is also a limit to which any side > effects can be emphasized because you run the risk of actually > suggesting that subjects should experience a side effect. > > It is understood that all psychotherapy outcome studies cannot > be blinded. It is so widely understood that people > don't think about its consequences any longer. > > If the standard for empirical validation is one blinded study > showing a treatment effect then no study of meds or > psychotherapy meets the standard. > > I think this area would be fascinating to study. Survey > the subjects and ask them which condition they thought they > were in and why. > > Maybe we should only analyze the data of subjects whose beliefs > were consistent with the treatment condition they > were actually assigned to. Someone should at least analyze > the relevant groups to see if there is a difference. > > If the FDA mandated that the drug companies verify the blinding, > the research reps would piss their pants. They > know they are passing off a flimflam to make millions on > psychotropic meds. > > Mike Williams > > --- > You are currently subscribed to tips as: dap...@shaw.ca. > To unsubscribe click here: > http://fsulist.frostburg.edu/u?id=13157.966b795bc7f3ccb35e3da08aebe98f18&n=T&l=tips&o=12706or > send a blank email to > leave-12706-13157.966b795bc7f3ccb35e3da08aebe98...@fsulist.frostburg.edu > Don Allen Retired professor Langara College --- You are currently subscribed to tips as: arch...@jab.org. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=12723 or send a blank email to leave-12723-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
