Correlations in Social Neuroscience Aren't Voodoo. Lieberman, M. D, Berkman, E. T, & Wager, T. D Perspectives on Psychological Science, 4(3):299–307, 2009. abstract bibtex Vul, Harris, Winkielman, and Pashler (2009, this issue) claim that many brain–personality correlations in fMRI studies are ‘‘likely . . . spurious’’ (p. 274), and ‘‘should not be believed’’ (p. 285). Several of their con- clusions are incorrect. First, they incorrectly claim that whole-brain regressions use an invalid and ‘‘nonindepen- dent’’ two-step inferential procedure, a determination based on a survey sent to researchers that only included nondiagnostic questions about the descriptive process of plotting one’s data. We explain how whole-brain regres- sions are a valid single-step method of identifying brain regions that have reliable correlations with individual difference measures. Second, they claim that large corre- lations from whole-brain regression analyses may be the result of noise alone. We provide a simulation to demon- strate that typical fMRI sample sizes will only rarely produce large correlations in the absence of any true ef- fect. Third, they claim that the reported correlations are inflated to the point of being ‘‘implausibly high.’’ Though biased post hoc correlation estimates are a well-known consequence of conducting multiple tests, Vul et al. make inaccurate assumptions when estimating the theoretical ceiling of such correlations. Moreover, their own ‘‘meta- analysis’’ suggests that the magnitude of the bias is ap- proximately .12—a rather modest bias.
@article{lieberman_correlations_2009,
title = {Correlations in {Social} {Neuroscience} {Aren}'t {Voodoo}},
volume = {4},
abstract = {Vul, Harris, Winkielman, and Pashler (2009, this issue) claim that many brain–personality correlations in fMRI studies are ‘‘likely . . . spurious’’ (p. 274), and ‘‘should not be believed’’ (p. 285). Several of their con- clusions are incorrect. First, they incorrectly claim that whole-brain regressions use an invalid and ‘‘nonindepen- dent’’ two-step inferential procedure, a determination based on a survey sent to researchers that only included nondiagnostic questions about the descriptive process of plotting one’s data. We explain how whole-brain regres- sions are a valid single-step method of identifying brain regions that have reliable correlations with individual difference measures. Second, they claim that large corre- lations from whole-brain regression analyses may be the result of noise alone. We provide a simulation to demon- strate that typical fMRI sample sizes will only rarely produce large correlations in the absence of any true ef- fect. Third, they claim that the reported correlations are inflated to the point of being ‘‘implausibly high.’’ Though biased post hoc correlation estimates are a well-known consequence of conducting multiple tests, Vul et al. make inaccurate assumptions when estimating the theoretical ceiling of such correlations. Moreover, their own ‘‘meta- analysis’’ suggests that the magnitude of the bias is ap- proximately .12—a rather modest bias.},
number = {3},
journal = {Perspectives on Psychological Science},
author = {Lieberman, Matthew D and Berkman, Elliot T and Wager, Tor D},
year = {2009},
pages = {299--307},
}
Downloads: 0
{"_id":"YAszXortJ36Jc7RqS","bibbaseid":"lieberman-berkman-wager-correlationsinsocialneurosciencearentvoodoo-2009","author_short":["Lieberman, M. D","Berkman, E. T","Wager, T. D"],"bibdata":{"bibtype":"article","type":"article","title":"Correlations in Social Neuroscience Aren't Voodoo","volume":"4","abstract":"Vul, Harris, Winkielman, and Pashler (2009, this issue) claim that many brain–personality correlations in fMRI studies are ‘‘likely . . . spurious’’ (p. 274), and ‘‘should not be believed’’ (p. 285). Several of their con- clusions are incorrect. First, they incorrectly claim that whole-brain regressions use an invalid and ‘‘nonindepen- dent’’ two-step inferential procedure, a determination based on a survey sent to researchers that only included nondiagnostic questions about the descriptive process of plotting one’s data. We explain how whole-brain regres- sions are a valid single-step method of identifying brain regions that have reliable correlations with individual difference measures. Second, they claim that large corre- lations from whole-brain regression analyses may be the result of noise alone. We provide a simulation to demon- strate that typical fMRI sample sizes will only rarely produce large correlations in the absence of any true ef- fect. Third, they claim that the reported correlations are inflated to the point of being ‘‘implausibly high.’’ Though biased post hoc correlation estimates are a well-known consequence of conducting multiple tests, Vul et al. make inaccurate assumptions when estimating the theoretical ceiling of such correlations. Moreover, their own ‘‘meta- analysis’’ suggests that the magnitude of the bias is ap- proximately .12—a rather modest bias.","number":"3","journal":"Perspectives on Psychological Science","author":[{"propositions":[],"lastnames":["Lieberman"],"firstnames":["Matthew","D"],"suffixes":[]},{"propositions":[],"lastnames":["Berkman"],"firstnames":["Elliot","T"],"suffixes":[]},{"propositions":[],"lastnames":["Wager"],"firstnames":["Tor","D"],"suffixes":[]}],"year":"2009","pages":"299–307","bibtex":"@article{lieberman_correlations_2009,\n\ttitle = {Correlations in {Social} {Neuroscience} {Aren}'t {Voodoo}},\n\tvolume = {4},\n\tabstract = {Vul, Harris, Winkielman, and Pashler (2009, this issue) claim that many brain–personality correlations in fMRI studies are ‘‘likely . . . spurious’’ (p. 274), and ‘‘should not be believed’’ (p. 285). Several of their con- clusions are incorrect. First, they incorrectly claim that whole-brain regressions use an invalid and ‘‘nonindepen- dent’’ two-step inferential procedure, a determination based on a survey sent to researchers that only included nondiagnostic questions about the descriptive process of plotting one’s data. We explain how whole-brain regres- sions are a valid single-step method of identifying brain regions that have reliable correlations with individual difference measures. Second, they claim that large corre- lations from whole-brain regression analyses may be the result of noise alone. We provide a simulation to demon- strate that typical fMRI sample sizes will only rarely produce large correlations in the absence of any true ef- fect. Third, they claim that the reported correlations are inflated to the point of being ‘‘implausibly high.’’ Though biased post hoc correlation estimates are a well-known consequence of conducting multiple tests, Vul et al. make inaccurate assumptions when estimating the theoretical ceiling of such correlations. Moreover, their own ‘‘meta- analysis’’ suggests that the magnitude of the bias is ap- proximately .12—a rather modest bias.},\n\tnumber = {3},\n\tjournal = {Perspectives on Psychological Science},\n\tauthor = {Lieberman, Matthew D and Berkman, Elliot T and Wager, Tor D},\n\tyear = {2009},\n\tpages = {299--307},\n}\n\n","author_short":["Lieberman, M. D","Berkman, E. T","Wager, T. D"],"key":"lieberman_correlations_2009","id":"lieberman_correlations_2009","bibbaseid":"lieberman-berkman-wager-correlationsinsocialneurosciencearentvoodoo-2009","role":"author","urls":{},"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/michaeltarr","dataSources":["5qXSH7BrePnXHtcrf"],"keywords":[],"search_terms":["correlations","social","neuroscience","aren","voodoo","lieberman","berkman","wager"],"title":"Correlations in Social Neuroscience Aren't Voodoo","year":2009}