There is an open poll for everyone to express their opinion on the direction in which SmartThings is headed. Currently, there are only 25 participants. I know we can do better than that!
Basically nearly everyone in the last 5 hours voted “yes”. Having the poll buried in a thread with people complaining may have skewed the early results.
My calcs show that around 25 people voted no, wondering what it was like when only 25 votes were in.
If the results were skewed due to a confirmation bias at the outset due to crowd in the thread, how did that change so suddenly in the end? The thread didn’t change in the last 5 hours.
This thread was new thread with no biased attached.
The other thread was people talking about “mass exodus” so when a poll resulted I would guess at least 10-12 people voted no immediately. That is why I am curious about early results when only 25 votes were locked in.
Congrats! You have reached statistical significance with margin of error of 10% and 95% certainty (which is around 95 respondents)
1 Like
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
11
I forget my basic Stats class (and, advanced stats, actually…);
But how do you rate significance without:
Knowing the population size (i.e., the number of SmartThings’s customers or potential customers).
The quality of the sample (i.e., randomness of selection from the population – Forum users do not represent a random cross-section of all SmartThings Customers).
Of course, if you define the population differently (e.g., all Forum users in the past 30 days?), then …?
A population of 10,000 needs 95, anything over 100,000 needs 96. 95% accuracy with 10% margin.
Now the sample is another issue, since the forum people are a special bunch.
Sill its safe to say that that the majority of the forum members are happy with the direction smartthings is going. 65% with a 10% margin of error still puts worse case at 55%.