posted by Anonymous on .
A recent study of 100 employees from six departments of a major corporation found 65% to be sleep deprived. The researchers concluded that the majority of corporate employees are sleep deprived.
The researcher’s conclusion is suspect because it
For the conclusion to be valid, the results of the study need to be representative of those that would have been obtained if the entire company had been surveyed. Are they? A major corporation could have thousands of enployees in any number of departments, whereas we're only looking at 100 employees from six departments. Were the employees appropriately sampled, perhaps using stratified random sampling to reflect the differences in the sizes of the departments? Were the employees required to respond to the survey? If not, wouldn't the ones who DID feel sleep-deprived be more likely to respond than those who didn't? And finally, is the result actually statistically significant? The researcher would need to demonstrate that an appropriate sampling technique was used to obtain the data in the first place, and also to calculate within what limits the estimate of 65% could realistically lie, given the size of the sample, before the conclusion could be considered credible.