TT97
Active Member
The CI is only accurate assuming this was a random sample. We know that this is not. This is a type 1 error. Ergo, can’t make inferences to population.
You can never have a 100% random survey when polling people as people always have the option to opt-in or opt-out for any survey. My guess is the two largest group of people that sign up for TeslaFi are the data geeks that love the info it provides and the business traveler that appreciate the reporting for tracking expenses. [Well, the largest group that signs up for TeslaFi are Tesla owners].
While not a 100% random sampling, there is no indication that the sampling frame would be different than the population.
If the updates were being push out 60% or even 70% to D models, I can chalk it up to an anomaly, however, 59 of the 67 Model 3s that upgraded from 36.2 to 39.7 have been Ds.