There are 3 kinds of lies: lies, big lies and … statistics.
Many news topics are based on so called statistical studies.
In average … the conclusions are based on the average.
In average … there is no mention of the data range or other variation metrics, like standard deviation.
In average … there is no mention of sample size.
The other day I heard a great example of the above on the radio regarding the correlation of traffic jams and people working from their home office. The hypothesis attached to the story is : home working does not reduce traffic jams.
Due to the difficult traffic situation, like in any other densely populated area in the world, and the opportunities offered by modern technology withvirtual private networks set upover broadband internet connections, it becomes more and more popular that people work from their home office a couple of days per week.
According to a study, in average ;-), people working from home use their car more often then if they would work in an office. Compared to the latter, people that work from home use the car to go shopping and take their children to and from school, which coincides with traffic rush hours. As a result and conclusion the hypothesis is not rejected.
I don’t know what bothers me most. The fact that these people need to go shopping and take their children to school, regardless the location from which they work or the undoubtedly incomplete and simplistic reflection of the study details. As usual, there was no mention of sampling strategy and size, geographical area where the sample was taken, if there was a significant difference in mileage or time people are driving (which is a different metric with higher environmental and society impact), etc …
For me, incomplete statistics really are a 3rd kind of lie, whether they are used in the news or to solve business issues.