Does Twitter language reliably predict heart disease? A commentary on Eichstaedt et al. (2015a)

View article
Language modeling has been thoroughly debunked by the one and only Nick Brown @sTeamTraen with a minor assist from me. Bottom line: real people have to peek at and fix arbitrary machine coding to avoid ridiculous findings. https://t.co/xVtUkh92Fz https://t.co/k7FlOu5SGn
reply repost on X like
1581 days ago
See https://t.co/K0OZCkvgeZ for examples of how Twitter censors datasets of tweets, including removing those containing the words "Jew" and "Muslim". https://t.co/vCfTp0ZMyP
reply repost on X like
1856 days ago
@OpinionCovid Here is the PDF of our paper: https://t.co/cKEWPWZPKH The original article is in theory behind a paywall, but I think there are copies available with a bit of googling.
reply repost on X like
2070 days ago
RT @TjeerdWBoonstra: @Neuro_Skeptic I handled a commentary by @sTeamTraen on a related paper for @thePeerJ https://t.co/yUcaYbnvsa Reviews…
reply repost on X like
2072 days ago
RT @TjeerdWBoonstra: @Neuro_Skeptic I handled a commentary by @sTeamTraen on a related paper for @thePeerJ https://t.co/yUcaYbnvsa Reviews…
reply repost on X like
RT @TjeerdWBoonstra: @Neuro_Skeptic I handled a commentary by @sTeamTraen on a related paper for @thePeerJ https://t.co/yUcaYbnvsa Reviews…
reply repost on X like
@Neuro_Skeptic I handled a commentary by @sTeamTraen on a related paper for @thePeerJ https://t.co/yUcaYbnvsa Reviews and author replies are publicly available for those who are interested
reply repost on X like
2304 days ago
@jessesingal @caladrylclear7 Data all public, so they just had to reassemble. And the error (failing to control for population) is flagrant. Aggregating stuff (here, into counties) is full of pitfalls. Clinton "causing" an even worse effect is a cool idea. Compare with my paper https://t.co/K0OZCkvgeZ
reply repost on X like
RT @sTeamTraen: "Tweeted Anger Predicts County-Level Results of 2016 US Presidential Election" (https://t.co/8E21ufxc8E) cites and indeed r…
reply repost on X like
2423 days ago
RT @sTeamTraen: "Tweeted Anger Predicts County-Level Results of 2016 US Presidential Election" (https://t.co/8E21ufxc8E) cites and indeed r…
reply repost on X like
2423 days ago
"Tweeted Anger Predicts County-Level Results of 2016 US Presidential Election" (https://t.co/8E21ufxc8E) cites and indeed re-uses dataset from Eichstaedt et al. (2015); I'm just wondering if the authors read Brown & Coyne (2018) https://t.co/K0OZCkvgeZ, which they don't cite?
reply repost on X like
RT @toddkashdan: A plea for corrections controversies on data quality to be published in original study journal. @sTeamTraen comment is not…
reply repost on X like
2539 days ago
RT @toddkashdan: A plea for corrections controversies on data quality to be published in original study journal. @sTeamTraen comment is not…
reply repost on X like
A plea for corrections controversies on data quality to be published in original study journal. @sTeamTraen comment is not published in the highly cited @aps psych science article arguing positive Twitter language predicts major health outcome. Read both! https://t.co/4EkTVYzezc
reply repost on X like
My article with @sTeamTraen 'Does Twitter language reliably predict heart disease? A commentary on Eichstaedt et al. (2015a)' was one of the top 5 most viewed #Epidemiology and #HealthPolicy articles published in @thePeerJ journal in 2018!
reply repost on X like
We had a similar issue with this commentary submitted to PeerJ: https://t.co/yUcaYbnvsa, which was posted as preprint here: https://t.co/6z0cJ15A0Q. The authors of the original paper then responded here https://t.co/DCcrEPU07e before the commentary was submitted for peer review.
reply repost on X like
2584 days ago
@jimcoan @jayvanbavel @LindaSkitka I think these may be from Twitter, in which case see https://t.co/K0OZCkvgeZ and especially the section entitled "Apparent censorship of the Twitter data".
reply repost on X like
2629 days ago
The answer is this commentary is ‘no’. #ICRE2018 via ⁦@pash22⁩ https://t.co/8dCobjvjaX
reply repost on X like
2629 days ago
RT @usama_bilal: @VPplenarysesh "We conclude that there is no good evidence that analyzing Twitter data in bulk in this way can add anythin…
reply repost on X like
RT @usama_bilal: @VPplenarysesh "We conclude that there is no good evidence that analyzing Twitter data in bulk in this way can add anythin…
reply repost on X like
2630 days ago
@VPplenarysesh "We conclude that there is no good evidence that analyzing Twitter data in bulk in this way can add anything useful to our ability to understand geographical variation in AHD mortality rates." https://t.co/hKc5S0Ma39
reply repost on X like
RT @k_kasiao: Does #Twitter language reliably predict #heartdisease? https://t.co/C1ZdalNQXx #socialmedia #hcsm #SoMe #hcmktg #healthcare #…
reply repost on X like
2631 days ago
RT @k_kasiao: Does #Twitter language reliably predict #heartdisease? https://t.co/C1ZdalNQXx #socialmedia #hcsm #SoMe #hcmktg #healthcare #…
reply repost on X like
2631 days ago
Does #Twitter language reliably predict #heartdisease? https://t.co/C1ZdalNQXx #socialmedia #hcsm #SoMe #hcmktg #healthcare #mhealth
reply repost on X like
2634 days ago
"Yet the rejection of a statistical null hypothesis cannot in itself justify the acceptance of any particular alternative hypothesis in the absence of any coherent theoretical explanation." On the importance of #theory for #computation: https://t.co/jMbh7n20B7
reply repost on X like