tag:blogger.com,1999:blog-5841910768079015534.post9126776046224406842..comments2024-03-25T17:14:36.888+00:00Comments on BishopBlog: Sample selection in genetic studies: impact of restricted rangedeevybeehttp://www.blogger.com/profile/15118040887173718391noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-5841910768079015534.post-82924115390030376332017-04-27T07:30:54.644+01:002017-04-27T07:30:54.644+01:00I think if you are offering your services as someo...I think if you are offering your services as someone writing CVs, you should polish up your written English.deevybeehttps://www.blogger.com/profile/15118040887173718391noreply@blogger.comtag:blogger.com,1999:blog-5841910768079015534.post-5652992780583211602017-04-26T15:39:26.425+01:002017-04-26T15:39:26.425+01:00That’s a very good work that you and your colleagu...That’s a very good work that you and your colleagues are doing. Quality and methods of teaching and learning needs a lot of improvement. Glad someone is working on it. Lewis Clarkhttps://www.blogger.com/profile/05470410316942686555noreply@blogger.comtag:blogger.com,1999:blog-5841910768079015534.post-45301925955477485312017-04-25T22:44:41.139+01:002017-04-25T22:44:41.139+01:00Off topic: even though I don't really understa...Off topic: even though I don't really understand github I managed to grab the code and it runs in R.<br /><br />More or less on topic, I read the post, without understanding 50% and still said. Oh yes, obvious once someone points this out. <br /><br />Now I need to go back and read this carefully and see if I really agree. :) It is considerably outside my area but the implications are fascinating.jrkrideauhttps://www.blogger.com/profile/04869979887929067657noreply@blogger.comtag:blogger.com,1999:blog-5841910768079015534.post-54198637088881758652017-04-25T05:36:59.144+01:002017-04-25T05:36:59.144+01:00Thanks Franck. It is, of course, much harder to ge...Thanks Franck. It is, of course, much harder to get truly representative samples, and with cognitive phenotypes it is often the case that those who have major difficulties are less likely to volunteer.<br />Marcus Munafo also drew my attention to this paper on 'collider bias' which is relevant:<br />http://biorxiv.org/content/early/2016/10/07/079707deevybeehttps://www.blogger.com/profile/15118040887173718391noreply@blogger.comtag:blogger.com,1999:blog-5841910768079015534.post-55335233168801090042017-04-24T12:48:00.269+01:002017-04-24T12:48:00.269+01:00A familiar variant of this is, if you aim to compa...A familiar variant of this is, if you aim to compare two groups on something (gene frequencies, brain measures, cognitive scores), sampling from a population where the two groups differ less than in the general population is not a sensible strategy.<br />As I'm arguing in a forthcoming paper, looking for brain correlates of dyslexia in a university student population (or in children reading just 1SD below the norm) is likely to decrease the expected effect size, hence the likelihood of finding a reliable difference. A more sensible sampling strategy would be to recruit the most severe dyslexic individuals that can be found. Yet my feeling is that neuroimaging studies of dyslexia have particular lax inclusion criteria.Franck Ramushttps://www.blogger.com/profile/02656240693713885894noreply@blogger.com