Sampling is like shining a light into a darkened room to see what is there. A small data sample gives you a view equivalent to a penlight. The larger the sample size, the brighter and wider the light. Sampling your entire population of customers, patients, or employees is like switching on the overhead lights, fully exposing everything in the room.
Rarely, though do we have the opportunity to ask questions of an entire population. And, even when we do have that ability (as we might with an employee survey), rarely do we get a 100% response. So, if we cannot ask everyone, how can we maximize the utility of the sample we do get? After all, if we do not shine the light in the correct part of the room, we might not find the answer to the question we are asking.
The solution is to increase the chances that the sample we get shines light in the correct part of the room. We can do this a couple of different ways. One way is to use a random sample. By randomly selecting who will participate, we are more likely to get a sample that represents everyone in a population. If we sample an intact group, such as just accounting, then we are less likely to represent those in other groups, such as those who work on the production line. Any answer we get using a non-representative sample will likely produce a biased (inaccurate) reflection of the population.
We can also more narrowly define the question we are asking. This effectively decreases the size of the room that we are trying to illuminate. In this case, if the question is focused on the accounting department, then surveying just that department makes sense and provides a more accurate glimpse into the room.
So, when you decide to ask a question, first define your population and then randomly sample that population. That way you will get the best possible illumination into the darkened room.
If you need more guidance or assistance in making the most out of your people data, we are glad to help. Contact us using our Contact Form.