Talk:Clinical study design

Untitled
I think the introductory sentence is too scientific. It will be better if layman terms can be used. — Preceding unsigned comment added by 147.8.33.225 (talk • contribs) 07:59, 29 April 2006 (UTC)

Please use the discussion page to comment on the articles.
User:65.96.168.157 added the following passage to the article: "The current author of this page states: "Some of the most popular designs are sorted below, with the ones at the top being the most powerful at reducing observer-expectancy effect (but also most expensive, and in some cases introducing ethical concerns) and the ones at the bottom being the most affordable." However, experimental designs are, in my mind, most well known for allowing the researcher to control the conditions in treatment/sample and control/comparison groups - not just the so-called observer-expectancy effect (which I do not beleive to be in the mainstream of social science research design literature, to my knowledge - could be wrong). I have a PhD in health services research and health policy and I teach research methods, so I encourage the Wiki editors to look into this. Thanks!" Since this does obviously not belong to the article I've moved it to this page. Please use the discussion page to comment on articles. Falk Lieder 16:54, 14 October 2006 (UTC)

Potential Merger
While I like the content on this page, it could be applied to a much wider scope of studies, as implied by the page name, which implies a general scope. I also note that Design of experiments exists, and contains a lot of non-overlapping material. Perhaps one or other of these pages (possibly even the currently redirected Research design) would be a good home to a combined page? --Limegreen 22:58, 16 October 2006 (UTC)

Inappropriate title - suggest "Epidemiological study designs"
The title is too vague - study design could refer to many different disciplines - all the designs mentioned are epidemiological study designs, and I think this would be a more sensible title.

Astaines 20:25, 9 May 2007 (UTC)

The term "research design" has a much broader meaning than discussed in this article. This article is about a specific type of study in epidemiology. The term "research design" can refer to research in all of the sciences, in the social sciences, and in other fields as well. I have designed research published in reputable scientific journals. I'm not a "layman".Richard Dates 18:36, 17 May 2007 (UTC)

I support this idea; this page is a useful one and it is about epidemiological study designs, nothing else. Any disagreement? Richard Keatinge (talk) 13:42, 12 September 2008 (UTC)
 * I disagree - while the current title may be too vague, this article covers more than traditional epidemiological designs. Perhaps "Clinical study designs" may be better, or maybe "Medical research study designs"&mdash; G716  &lt;T·C&gt; 14:08, 12 September 2008 (UTC)
 * BTW, I've tried to get this article more visibility by adding it to Medical research studies and Clinical_trial&mdash; G716  &lt;T·C&gt; 14:16, 12 September 2008 (UTC)
 * Support move to list of clinical study designs or similar. It doesn't read as an ordinary WP article, more as a glossary of miscellaneous terms.  The title as it is could be read to include studies in physics, literary criticism.... --94.194.57.116 (talk) 06:47, 3 May 2009 (UTC)

Experimental designs
A nonrandomized clinical trial cannot be "experimental". Randomization is part of the design of experiments. I am not good at editing images (although I can improve). I think someone should edit (or replace) the image, such that nonrandomized clinical trial is not considered "experimental."Iss246 (talk) 05:17, 14 October 2008 (UTC)


 * An "experimental" study is one in which the study organizers decide treatment allocation.  Nonrandomized clinical trials fit perfectly into that definition. Richard Keatinge (talk) 06:48, 14 October 2008 (UTC)

Randomized experimental designs are true experimental designs. Without random allocation to rival treatments, you don't have an experiment. Let's not get into an argument about this. Please consult http://www.socialresearchmethods.net/kb/desexper.php or a reputable text on research design.

We shouldn't be arguing about this. We should be improving the entry. Someone got the entry started, which is good. Now we have to refine it.Iss246 (talk) 00:34, 15 October 2008 (UTC)


 * Sorry to nitpick, but as your text says, while randomization is a highly desirable part of epidemiological experiments and is now usual, it is not required by the definition. There are many nonrandomized experimental studies; Google Scholar finds 24,500 hits for "non-randomized trial". I quite agree, rather than nitpicking, we should be improving this article. Richard Keatinge (talk) 07:56, 15 October 2008 (UTC)

Nitpicking is okay. The web site I cited indicated "The experiment relies on this idea of random assignment to groups as the basis for obtaining two groups that are similar. Then, we give one the program or treatment and we don't give it to the other. We observe the same outcomes in both groups." Random assignment is a critical ingredient in an experiment.

I think where we part company is here. The term quasi-experiment refers to a trial in which there are rival treatments, but the research team does not randomly assign subjects to groups. The team works with pre-existing groups. I think we will help our readers (who, I think, are primarily college students) if we clearly delineate an experiment from a quasi-experiment. I think the diagram will confuse them. Do you want to make the changes?Iss246 (talk) 00:57, 16 October 2008 (UTC)


 * I'd call a study of pre-existing groups "observational". A design in which the team assigns subjects to groups is "experimental", if that is done randomly it is a randomized controlled trial, if assignment is by the team but is not random, it is a nonrandomized experimental study. I think that's in accordance with normal usage. And "quasi-experiment" usually describes an observational study. Richard Keatinge (talk) 08:39, 16 October 2008 (UTC)

Not really. The reseach evaluating Head Start has been quasi-experimental. So has the research evaluating WIC. Research on programs used schools has been quasi-experimental (e.g., phonics v. whole language) because researchers cannot randomly assign children to schools or classes. Researchers must implement the rival treatments (phonics v. whole language; Head Start v. other preschools) in preexisting groups.Iss246 (talk) 14:01, 16 October 2008 (UTC)


 * You seem to be a social scientist, and I'm an epidemiologist. It occurs to me that "quasi-experiment" may be used differently by those two disciplines. I haven't actually seen it used much in recent epidemiology. But, in epidemiology, it describes observational studies, situations where groups are placed in very different situations, not by the research team but by some other circumstance. I've heard the term used to described the classic comparison between bus drivers and bus conductors, the latter having lower rates of heart disease and the difference being attributed to greater exercise. Going back to etymology, a quasi-anything is similar to, but not the same as, that anything. A quasi-experiment is similar to, but not, an experiment. The work you describe in the previous paragraph seems to be what I'd call nonrandomized cluster trials. But if you can provide usage to the contrary, I will have learned something. Richard Keatinge (talk) 11:09, 17 October 2008 (UTC)

I am a social scientist but I completed a post-doc in epi. There is some difference in nomenclature between epi and social science. I think it would helpful to users of the entry--they are most likely undergraduates--if the studies were divided between those that have treatments/interventions and those that do not. Perhaps you should augment the entry by adding a section devoted to epidemiologic research.Iss246 (talk) 00:24, 19 October 2008 (UTC)

Note of merge discussion Jan 2010
Duscussion of a merge proposal is at Talk:Design_of_experiments/Archive_1. The resuly was to keep separate, but there may be some points worth following up. Melcombe (talk) 10:21, 11 January 2010 (UTC)

Delete this page?
I suggest this page being deleted if not substantially improved. As it is it doesn't say much which isn't covered elsewhere (in particular under clinical trial and it is poorly written.Helenuh (talk) 14:38, 26 January 2017 (UTC)
 * I see you do not do a lot of Wiki editing. Anyone can start a deletion discussion by following the instructions at WP:AFD, but if it helps, I can start a deletion discussion for you if you can get it started with some more information. Can you say what, if anything, in this article could be moved to clinical trial? If there is something to salvage here then it should be moved, and the rest can be discarded. Alternatively, if you could provide citations to sources on "clinical study design", then those could be listed here to get a start on expanding this article. Would you like to contribute something in either of these directions?  Blue Rasberry   (talk)  15:50, 26 January 2017 (UTC)
 * This article could be significantly expanded so it would a step backwards to merge it into clinical trial. - Rod57 (talk) 12:26, 11 November 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 external links on Clinical study design. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20061217105946/http://www.epidemiologic.org/ to http://www.epidemiologic.org/
 * Added archive https://web.archive.org/web/20070228224146/http://www.epidemiologic.org/forum/ to http://www.epidemiologic.org/forum/

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 13:53, 9 August 2017 (UTC)
 * Not useful as external links, WP:ELNO. --Zefr (talk) 15:08, 9 August 2017 (UTC)

Could mention Basket studies
Article does not mention Basket studies. says "Basket studies use an innovative clinical trial design that helps collect data faster, and may accelerate the development of medicines for diseases with high unmet need. Instead of enrolling people based primarily on their disease or its location, basket studies match a disease’s underlying genetic profile to the mechanism of action of the medicine." Seems worth mentioning. - is more detailed. ~ RS ? ... Basket Trials in Oncology: A Trade-Off Between Complexity and Efficiency compares them to Umbrella trials. Not sure where they fit into the current article. - Rod57 (talk) 11:56, 11 November 2017 (UTC)
 * If the term is used outside medical/clinical studies perhaps it should go in Design of experiments. - Rod57 (talk) 12:02, 11 November 2017 (UTC)