Alpha Level

You can also check this article out on Medium

Concept of the Alpha Level

First, we know that the null hypothesis and the alternative hypothesis are the main questions we aim to test. Its easy to think about the two-sided hypothesis test between two means:

Null hypothesis - H0: Mean(1) equals Mean(2)

Alternative hypothesis - Ha: Mean(1) does-not-equal Mean(2)

We can think about the alpha level quite simply as the probability of rejecting the null hypothesis when, in actuality, the null hypothesis is true. Basically, it represents the probability of us making the mistake of saying that H0 is not true when it is true. When this mistake is made, we say that we have made a Type I error.

Math notation-wise we just write that as

Another name commonly used for the alpha level is “significance level”. The probability, α is also referred to as the “Type I error rate” (the chance of making a type I error)

Application of the Alpha Level ()

On how the alpha level is used in applied statistics we as the analyst, researcher, data scientist or student are allowed to specify this probability, α. So, we get to decide what the chance of making a type I error is (or in other words what the probability of incorrectly rejecting the null hypothesis when the null hypothesis is in fact true).

Typical values that are chosen for alpha are

· 0.05 (Most commonly used)

· 0.01 (Used if you really want to see if the results are meaningful or in some precise applications)

· 0.10 (Used when you want to loosen the restriction of finding a significant difference)

After you conduct a hypothesis test, like the two means T-test specified above, and get a p-value, you basically compare the obtained p-value and your specified alpha:

· If (p-value) < , then we can reject the null hypothesis that the two means are the same.

· If (p-value) > , then we failed to reject the null hypothesis and that is the one we keep.

So that’s the alpha level in a nutshell. Specifying the alpha level is actually a major consideration depending on the data you have or the questions your trying to answer. So, before you start crunching and coding give it some thought. Thanks for reading. Until next time, happy mining.

For more articles and content check out our YouTube Channel and our website, if you like what we have to offer then please show your support and like us on Facebook and Follow us on Twitter