In our daily work, as a data department, we often say: data is leadership. Many times, if there is a conflict at work, from the perspective of A, there is a reason for A, and from the perspective of B, there is a reason for B. It is difficult Whatsapp Database to reach an agreement in terms of reason. However, it is possible to resolve differences by talking about data, because talking about data is talking about facts, and there can Whatsapp Database only be one fact. However, in work and life, we often find that even if the data is completely real, we may still be deceived. Today, we will share some basic methods to avoid being deceived by real data. 1.
Sample Lies The sample lie refers to the fact that the objective objects we face are huge and complex. Under the constraints Whatsapp Database of time, cost, ability and other factors, there is no way to measure the full amount, but only the samples drawn. The problem with sampling is how to ensure that the sample is representative of the Whatsapp Database population. There are examples of extremely small samples or no samples: Small sample: The Chinese men's football team has only lost three games in the long history of the World Cup. No sample:
The Chinese men's football team has never lost a World Cup game since the 2002 World Cup. Of course, this is a joke. Whatsapp Database Some people say that I can avoid this problem by increasing the sample size as much as possible. In fact, it is not. It is familiar to the 1936 US election. Literary Digest magazine surveyed 2.4 million voters, while Gallup surveyed only 5,000 people. The result was that Gallup's prediction was successful. The biggest Whatsapp Database difference is that Gallup is sampling in the same proportion according to the demographic structure of all voters. Although the sample is small, it is typical enough. In real life, there is also a scenario in which the sample is selectively enlarged or selected artificially, resulting in misleading information receivers.