Regression assumptions in clinical psychology research practice—a systematic review of common misconceptions

View article
@byrd_nick @rstatstweet @lakens @learnfromerror @CaAl has a nice paper with Ernst: https://t.co/nbE9d7JQtB
reply repost on X like
1891 days ago
@unnombrealazar https://t.co/Ov4w5GknBA
reply repost on X like
@HenrikSingmann @CaAl (oops, should link to the 2017 version: https://t.co/nbE9d7JQtB)
reply repost on X like
RT @CaAl: @LisaDeBruine You can even make it much 'more' bimodal and still satisfy the normality assumption. "DV ~ normal" probably is the…
reply repost on X like
https://t.co/i1VDQBs9Vl
reply repost on X like
2002 days ago
@LisaDeBruine You can even make it much 'more' bimodal and still satisfy the normality assumption. "DV ~ normal" probably is the most common mistake w.r.t. statistical assumptions. https://t.co/dZimZjX0DZ https://t.co/OW7xqehpQP
reply repost on X like
RT @SarahSDermody: @KMKing_Psych Yes! I assign this review of common misconceptions to my grad quant students in clinical psychology resear…
reply repost on X like
RT @SarahSDermody: @KMKing_Psych Yes! I assign this review of common misconceptions to my grad quant students in clinical psychology resear…
reply repost on X like
@KMKing_Psych Yes! I assign this review of common misconceptions to my grad quant students in clinical psychology research. I think Figure 1 is a nice simple example of non-normally distributed outcome that could have normal residuals following analysis. https://t.co/p1ERbTMwGt via @thePeerJ
reply repost on X like
https://t.co/auvhmtFd8K
reply repost on X like
RT @richarddmorey: @Research_Tim Simulate some data; show them a scatterplot of with a two-level x variable with a large slope, then show t…
reply repost on X like
@Research_Tim Simulate some data; show them a scatterplot of with a two-level x variable with a large slope, then show them the (obviously bimodal) y variable, then show the (Gaussian) residuals. (Fig 1 from @CaAl's https://t.co/nbE9d7JQtB) https://t.co/vdu49ukTDN
reply repost on X like
RT @CaAl: @Research_Tim Thinking that X or Y themselves have to be normally distributed rather than the residuals, is a common misconceptio…
reply repost on X like
RT @CaAl: @Research_Tim Thinking that X or Y themselves have to be normally distributed rather than the residuals, is a common misconceptio…
reply repost on X like
2457 days ago
RT @CaAl: Our paper in @thePeerJ on misconceptions surrounding regression assumptions is now fully published (OA) https://t.co/dZimZjX0DZ h…
reply repost on X like
2457 days ago
RT @CaAl: Our paper in @thePeerJ on misconceptions surrounding regression assumptions is now fully published (OA) https://t.co/dZimZjX0DZ h…
reply repost on X like
2457 days ago
@Research_Tim Thinking that X or Y themselves have to be normally distributed rather than the residuals, is a common misconception https://t.co/dZimZjX0DZ Here: Depending on the values of the covariate, the distribution of the residuals can take almost any form
reply repost on X like
2627 days ago
@MaartenvSmeden @DanielOberski @statsepi @robertstats Indeed, it's common. What's even more common, is that people said "We checked for normality" without explicitly stating *what* they checked and *how* they checked it. See this meta-analysis by @anja_franziska and me: https://t.co/dZimZjX0DZ
reply repost on X like
Eek! More than 90% of pubs in clinical psyc journals using regression don't report tests of regression assumptions...those that do report them, often report incorrect assumptions (normality of variable instead of residual). https://t.co/p1ERbTuVhT via @thePeerJ
reply repost on X like
2995 days ago
@seriousstats @deevybee To biased esitmators, just sub-optimal ones (cf https://t.co/dZimZjX0DZ ) 3/2 (apologies for my inability to count tweets properly)
reply repost on X like
3027 days ago
@richarddmorey @EikoFried @VandekerckhoveJ I know (but that didn't fit in 140 chars). https://t.co/dZimZjX0DZ
reply repost on X like
RT @CaAl: Our paper in @thePeerJ on misconceptions surrounding regression assumptions is now fully published (OA) https://t.co/dZimZjX0DZ h…
reply repost on X like
I was a painful Reviewer 2 on this great article by Anja Ernst & @CaAl : https://t.co/gCgu4LVC04
reply repost on X like
RT @CaAl: Our paper in @thePeerJ on misconceptions surrounding regression assumptions is now fully published (OA) https://t.co/dZimZjX0DZ h…
reply repost on X like
Regression assumptions in clinical psychology research practice—a systematic review of common misconceptions https://t.co/ys9YHKR9p8
reply repost on X like