1462: Blind Trials

Explain xkcd: It's 'cause you're dumb.
Jump to: navigation, search
Blind Trials
Plus, you have to control for the fact that some people are into being blindfolded.
Title text: Plus, you have to control for the fact that some people are into being blindfolded.

Explanation[edit]

In research, a blind trial is an experiment where certain information about the test is concealed from the subjects and/or the testers, in order to reduce sources of bias in the results. A double-blind trial is one where neither the subject nor the testers know who has or has not received treatment (or for multiple treatments, which treatment).

A scientific approach also requires the use of control groups to determine the significance of observations in (clinical) trials. The members of the control group receive either no treatment or the "standard" treatment. However, to ensure "blindness" in the study, even if a control group is to receive no treatment, they must be given a placebo: an ineffective treatment given to ensure the doctors and/or patients are unaware whether they are being given the treatment.

For example, in clinical drug trials, when a treatment being tested is administered in the form of a pill, a visually-identical inert pill is given to the control group so no one will know if a subject has been given the treatment or a placebo. In pop culture, placebos in pill-form are often made of sugar, which has negligible medical effects.

Controls and blinding are crucial to distinguish the actual effects of the treatment from the placebo effect, or the psychologically-induced effects of a subject's belief that a treatment will or will not help them, which may have real physiologic effects or influence the reporting of subjective measures such as pain level or the presence of side effects. It is vital that there are no clues available to distinguish between the different groups. Even subtle cues from the body language of the testers are sufficient to trigger placebo effect, making double-blind trials necessary.

Challenges exist in designing placebo alternatives to certain physical treatments that might be tested, such as acupuncture; in this case the best quality trials have typically used either special 'joke' retractable needles that only give the illusion of proper penetration or the practitioner/researcher deliberately and safely avoids the traditional meridians on the body for the treatment concerned so that the patient remains 'blind' to their role in the trial. The practitioner must otherwise be consistent in treatment between groups and not be involved in the medical assessment phase for properly double-blinded conditions, where the most reliable results still seem to only show a significant placebo effect at work.

There are, however, certain cases where it is almost impossible to make the experience of the control group identical to that of the test group. Making a real and fake pill appear the same is a relatively trivial task, and the ignorance of participants to the details of a given established practice or procedure can allow for a certain level of blinding. However, it would be challenging (to say the least) to make the control group in the described experiment think that they are having lots of sex, when in fact they are not. The description of the control group as taking sugar pills is a laughably poor placebo substitute, as the pills themselves may affect cardiovascular health, therefore ruining the concept of a controlled experiment, especially if it is a double-blind trial, and nobody knows who got sugar and who got sex.

Scientific research involving humans is extremely challenging to conduct because of the difficulty in finding appropriate control groups. This is one of the reasons animal experiments (for instance involving inbred strains of mice) are so common.

The title text adds another twist by taking “blind” literally, and noting that for some people, being blindfolded increases their enjoyment of sexual activity, thereby acting as a confounding variable.

Despite this, it should be noted that Cardiovascular health is typically measured in terms of objective data such as cholesterol levels, ejection fraction, and morbidity/mortality data like the frequency of myocardial infarctions, strokes, or sudden cardiac death. Even sighted, it would be difficult for either subjects or researchers to manipulate this kind of data.

Transcript[edit]

[Megan is pointing at charts hanging on the wall.]
Megan: We've designed a double-blind trial to test the effect of sexual activity on cardiovascular health.
Both groups will think they're having lots of sex, but one group will actually be getting sugar pills.
The limitations of blind trials


comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!

Discussion

I think I finished the transcript 108.162.250.219 09:21, 19 December 2014 (UTC)

I think the item Megan is pointing at should not be referred to as "chart", but as "charts", since a chart is "a sheet of information in the form of a table, graph, or diagram", while there is clearly both a table and a diagram, therefore two "charts".17jiangz1 (talk) 12:45, 19 December 2014 (UTC)

I wonder though, if it makes sense to account for Placebo Effect when you're mesuring something that does not involve subjective perception. In the example case, one can measure, say, heart rate for the group that practices sex and the monastic control group and have valid results. Of course, you lose the option of passing a survey with questions like "Have you lately felt tired?" or the like. 173.245.49.106 09:33, 19 December 2014 (UTC)

Of course it does make sense, placebo effect can be observed on test for disease such as cancer. The idea is that thinking you are taking a cure might, will trigger physiological change. It is quitte likely, that the testing of pills for the heart would induce some placebo effect. What you might be refering too is some kind of perseptive bias, which could be observed in psychological test. 108.162.229.100 11:51, 19 December 2014 (UTC)

Hmmmm ... maybe you could make the control group think they have sex with hallucinogens ... but somehow I don't think it would make useful control group, not speaking about most hallucinogens being illegal. Alternatively, you can get both groups drunk and - ... ok, this doesn't sound like good idea either. -- Hkmaly (talk) 12:44, 19 December 2014 (UTC)

I was thinking maybe hypnotherapy... --Pudder (talk) 12:48, 19 December 2014 (UTC)
the only legitimate thing I could think of is to get the control group to use various forms of cardiovascular exercise to get a few measurable statistics (like heart rate, perspiration, respiration, blood pressure, etc) to be roughly equivalent to those experienced by the experimental group, then the experiment would be to show MOSTLY whatever else sex does (hormones, psychological stimulation, etc), if cancer susceptibility was shown to be lower in the experimental group, it MIGHT have some validity... it would be a fairly complex set of requirements to design and especially hard to control for things like sexual preference(s), but I could imagine a well-designed study being able to show a few things -- Brettpeirce (talk) 14:42, 19 December 2014 (UTC)

When I was diagnosed with accute hypochondria and then suffered a number of psychosomatic injuries, those sugar pills actually saved my life! 141.101.98.245 12:55, 19 December 2014 (UTC)

So, does a triple blind trial mean a threesome? Nialpxe (talk) 16:27, 19 December 2014 (UTC)

Just my luck - I'm in the second group! --RenniePet (talk) 03:18, 20 December 2014 (UTC)