I used to work on a team which would collect data for PR purposes for the company. The data is often very heavily skewed towards only what data is available to the company. This is far from a proper sample of the population. For example, instead of '24% of people do x' should be '24% of people that use our site do x'.
Zero statistical methodology is ever used to make sure the data is accurate for the broader population.
Also, just because some data is associated with someone with a PhD, doesn't mean it is any more accurate. Many PhDs and professors are put on part time pay rolls of these companies so as to give the bs data an air of accuracy and authority.
Zero statistical methodology is ever used to make sure the data is accurate for the broader population.
Also, just because some data is associated with someone with a PhD, doesn't mean it is any more accurate. Many PhDs and professors are put on part time pay rolls of these companies so as to give the bs data an air of accuracy and authority.