CrimsonEclipse
Elio Addict
- Joined
- May 24, 2014
- Messages
- 959
- Reaction score
- 2,182
I am always willing to listen to a well reasoned argument.
Please explain how 9% of a group represents a reasonable sample.
Always remembering that famous old statement popularized by Mark Twain .......
"There are three kinds of lies: lies, damned lies, and statistics"
Incidentally, I plan to install an OBD II sourced HUD in my Elio.
A sample size of 9% is an unusually large size. A sample size of less than 1% is absolutely acceptable assuming it's a well distributed sample.
A questionnaire can be used as a "lie" and its results skewed to make Mr. Clemens correct, but that usually includes loaded questions.
Is the Elgin Dash:
1. bad
2. really bad
3. gave me cancer
4. don't mess with Texas
This poll is REALLY simple, yes or no. No double negatives, no diversions, yes or no.
This simplicity actually makes the survey more accurate.
This poll does not, in any way, provide any meaningful sampling. It's a self-selected subset of the most rabid fans. Most of us have a penchant for customizing our vehicles, and have strong opinions for what we do and do not like. We already know that most of us are older men eligible for AARP membership. We can, by no stretch of the imagination, be considered "typical" customers. This is so far from a statically significant, representative sampling that you'd be laughed out the door if you tried to present it as such. In fact, you could almost represent this poll as a deliberate attempt to sample only those already known to be biased *against* it.
In order to claim statistical significance, sample selection is critical. Random sampling of a significant portion of the population is the only reliable method. A poll on this website can't claim that.
That two pounds can be a valid sample, so long as it is a randomly selected sample, that could reasonably assumed to be representative of the load as a whole. If, on the other hand, you grabbed a bunch that were next to an access hatch with a leaky seal, that's a bad sample.
You can achieve accurate results with smaller sample sizes, assuming your sampling methodology is sound. You just have to be aware that the smaller your sample size the greater the level of uncertainty, and the lower the confidence level.
You assume that the users on this forum are ALL rabid fans.
This forum actually represents an accurate cross section of the Elio population.
Being an Elio forum, a sampling of the population here is more accurate that a sampling of the general population since the mind set of the population on Elio Owners is what matters vs the mind set of the general population that may or may not desire the car.