I repeat this procedure 1000 times and save the p-value of the β coefficient each time. The intuition underlying this test is that if spillover effects were present, I would disproportionately reject the (true) null hypothesis of β=0. If spillover effects are not present the null hypothesis should only be rejected by chance (type-1 errors). The rejection rates at the 10%, 5%, and 1% significance levels are 4%, 1.1% and 0% respectively indicating the control group was unaffected by the treatment