←back to thread

1293 points rmason | 1 comments | | HN request time: 0s | source
Show context
b_tterc_p ◴[] No.19323150[source]
> How the study was conducted: A total of 1,500 persons were interviewed to explore Americans’ use of digital platforms and new media. From January 3rd through February 4th, 2019, telephone interviews were conducted with respondents age 12 and older who were selected via Random Digit Dial (RDD) sampling through both landline phones and mobile phones. The survey was offered in both Spanish and English. Data was weighted to national 12+ U.S. population estimates.

I wouldn’t put much faith in this estimate. While facebook’s is probably an overestimate of people actually engaged in their platform, this survey doesn’t seem very useful to me.

replies(2): >>19323453 #>>19323615 #
1. throwawaymath ◴[] No.19323615[source]
Is there a specific reason you doubt the estimate, or is your problem that the sample size is small? Small sample sizes don’t imply incorrect conclusions.

EDIT: To whoever has downvoted this, I politely (but urgently) recommend you read up on statistical significance. The idea that a small sample size implies a study’s findings are unreliable is one of the most widely held misconceptions in modern statistics.