4 ms·

Significant enough for what?

• pbaka 5y ago
  "Statistical significance" is a qualification in statistical jargon regarding how certain you can be in the absolute of your result (how the experiment was designed, how precise the measurement inherently is, number of measures, presence of cofounding phenomena, how much the measurement inherently varies, etc). You may have a very clear positive trend, but, because of other factors, you can't say without a trace of doubt : "This is what happens, that's a bulletproof correlation". This is why when you say there's not a "Statistical Significance" it doesn't mean the result is wrong if the trend shown is strong, rather just that you need to improve your experiment and measurements, and check for cofounders to be certain, to validate it to a sufficient degree.