by Ji Liu
From the dataset listed in the article, it can be seen that after restrictions in different states, there was sharp increase of gun sales. The author of the article tried to accuse the increase to strict restrictions. However, personally, this accusation is lack of proof in statistics.
Two evidences even listed by the author could repute him. Firstly, there are other obvious spikes on the chart, which did not appear after the restrictions policy days. This means, if restriction is one of the reason of the gun sales, it is not the only reason. Secondly, there was another sharp increase of gun sales after hurricane Katrina, which could hardly be seen any association with gun sales. This reveals that the chronological sequence does not have to be causation links.
The data the author had at hand is census, which makes the correlation test easier. Knowing parameter mean and standard deviation, the author could check the correlation of gun sales and restrictions via statistics Z-test, with the hypothesis that these two variables are correlated. In this case, the null hypothesis will be there is no correlation between them. Confidence interval, p-value and extreme values could all tell if there is a significant correlation between national restrictions of gun usages and gun sales. At that time, the author’s claim will be more scientific and convincing.