So a publicly traded company cannot validate one of its key metrics? And the CEO, who has a fiduciary duty to shareholders is telling the world that essentially, he doesnβt know how many real, living human beings actually use Twitter. Wow!
Lol the people reeeing over statistics. "100 people isn't representative of 260 million users" to validate their existence meanwhile "100 people is representative of 260 million voters political beliefs, leanings, opinions, etc."
The sample needs to be a certain size AND must also be representative of the group being sampled to yield accurate projections. Larger sample sizes generally produce better results, but the only accurate result is counting every user. Think how politcal polls are manipulated (oversample democrats then declare majority of likely voters support their democrat cadidate). They carefully contrive the sample to ensure the desired result to use as a tool to manipulate public opinion. For sure twitter carefully selected the 100 accounts to yield a number they felt the public would believe, in this case 5% bots.
I completely agree. I was just offering the observation that, from a strictly numerical standpoint, their "sample" size was far too small---aside from the issue of cooking the sample.
So a publicly traded company cannot validate one of its key metrics? And the CEO, who has a fiduciary duty to shareholders is telling the world that essentially, he doesnβt know how many real, living human beings actually use Twitter. Wow!
Lol the people reeeing over statistics. "100 people isn't representative of 260 million users" to validate their existence meanwhile "100 people is representative of 260 million voters political beliefs, leanings, opinions, etc."
Using a rough rule of thumb that the square root of the population would be the minimum sample size, it works out to 5,100.
The sample needs to be a certain size AND must also be representative of the group being sampled to yield accurate projections. Larger sample sizes generally produce better results, but the only accurate result is counting every user. Think how politcal polls are manipulated (oversample democrats then declare majority of likely voters support their democrat cadidate). They carefully contrive the sample to ensure the desired result to use as a tool to manipulate public opinion. For sure twitter carefully selected the 100 accounts to yield a number they felt the public would believe, in this case 5% bots.
I completely agree. I was just offering the observation that, from a strictly numerical standpoint, their "sample" size was far too small---aside from the issue of cooking the sample.