You can either pre-determine the desired statistical **power** and **significance** of a test to [[AB Test Sample Sizes]], or you can define a significance to calculate the critical value at which to reject $H_0$ when doing Null Hypothesis Significance Testing (NHST).  The above image is a visualization that explains how power ($1-\beta$) and significance ($\alpha$) influence the false negative and false positive rate of your test outcome, possibly driven by a critical value above which you consider that you will reject $H_0$. The larger the sample size, the larger the power of a test, and the less likely you are to not reject the $H_0$ when it is false. Similarly, with larger sample sizes, the likelihood of getting a significant result increases, and the less likely you are to reject the $H_0$ when it is true. Because in both cases, the bell-shaped curves above get narrower ("thinner", or more "squashed") the more samples you have.