@zeynep @seanjtaylor incidentally, your example of election polls isn't one of random sampling--N=people home & willing to answer phone, etc
-
-
Replying to @MichelleNMeyer
@zeynep@seanjtaylor truly random sampling hard. Best recent example that comes to mind ironically mood contagion experiment.1 reply 0 retweets 0 likes -
Replying to @MichelleNMeyer
@MichelleNMeyer@seanjtaylor Contagion study random in terms of experimental designation, not random representation of Facebook. Different.1 reply 0 retweets 0 likes -
Replying to @zeynep
@zeynep@seanjtaylor 700K were randomly selected by FB ID from among all who view FB in English--a rep sample of Englsh-viewing FB users no?1 reply 0 retweets 0 likes -
Replying to @MichelleNMeyer
@MichelleNMeyer@seanjtaylor "who posted one status update"2 replies 0 retweets 0 likes -
Replying to @zeynep
@zeynep@seanjtaylor course study analyzing text couldn't have been done w/users who produced no text. So results perhaps not gen to lurkers3 replies 0 retweets 0 likes -
Replying to @MichelleNMeyer
@MichelleNMeyer@seanjtaylor But big diff between class-based non-random (political-id + frequent use) and imperfect random while attempting1 reply 0 retweets 0 likes -
Replying to @zeynep
@zeynep@seanjtaylor true, but what were alternatives? Could use what self-ID pol users post to create predictive algorithm of ideology?6 replies 0 retweets 0 likes -
Replying to @MichelleNMeyer
@MichelleNMeyer@seanjtaylor 4-Inferential/imputational methods also confound, but everything is imperfect, triangulation is best.1 reply 0 retweets 0 likes -
Replying to @zeynep
@zeynep@seanjtaylor Suspect imputation-based study wld hv led 2criticisms of FB trying 2reduce complex beh 2algo; voting recs "creepy" etc2 replies 0 retweets 0 likes
@MichelleNMeyer @seanjtaylor I agree, they'll come under criticism. They can handle it, as a $150 billion company. :-)
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.