# When is a Ljung-Box test significant?

I have trouble understanding the output of the Ljung-Box test due to conflicting information:

The R documentation doesn't actually say how to interpret the output.

This site states that small p-values means that the data is likely to be stationary.

This otexts textbook states that large p-values means that the data is likely to be like white noise.

Clearly, one of them must be wrong. Which one is it?

Thanks!

The first one is simply wrong. If you just check the wiki on L-B test, you will see that the null hypothesis is that the data are White Noise, so small $p$-values (say $<0.05$) mean that you have evidence to reject the null hypothesis. Here is a good online reference. Moreover, if you would like to test your data for white noise, in addition to this, I could suggest you try our hwwntest package which contains several types of white noise tests with examples.