3 ms·
Sample size is immaterial unless you clearly can see and evaluate the criteria under which they were chosen. 8000 of 9000 dentists choose Brand X toothpaste mea
by coldcode 13y ago
Sample size is immaterial unless you clearly can see and evaluate the criteria under which they were chosen. 8000 of 9000 dentists choose Brand X toothpaste means nothing if they asked 100,000 people and found 8,000 willing to say they liked Brand X. By itself the sample size that is indicated may have little to do with the actual test.
- VLM 13y agoI was going to suggest the same thing, that given about 1/5 of people use Bing (why?) then you simply sample 5000 people, select the 1000 or so Bing users... However, it turns out there is an easier way. Simply select your sample from existing heavy Bing users. Oddly enough, they'll probably prefer Bing.
- mattwallaert 13y agoAnd this is precisely why a 3rd party does our testing: so that these sort of manipulations don't creep in. (Note: I work for Bing.)
- VLM 13y agoWell fair enough then, I'll take your word for it. As some marketing advice, beyond a certain (very low) level of savvy, people know the only reason statistics are used is because they're manipulated. So, don't use statistics in your ad campaign unless you want everyone above a certain (very low) level of savvy to reflexively believe they're faked, which isn't going to help your product. Hey did you know that 9 out of 10 marketing execs endorse everything VLM posts on HN? Yeah thats exactly what I'm talking about. Doesn't matter if it true or even if it matters, its just not a good selling strategy.