It could be said that confidence intervals, hypothesis testing, and p-values are three sides to the same coin.The p-valuetells you the probability that your results could have occurred due to chance alone. This is the number you want to focus on.

If you are in a discussion about a study involving a sample, and someone says they came up with a great new way to slice bread, your first question to them about the findings should be "what was the p-value?" P-values of 5% (i.e., 0.05) or less are generally considered good, i.e., 2%, 1%, less than 1% (i.e., 0.02, or 0.01, 0.005). Higher values are questionable, i.e., 10%, 15% (0.10, 0.15) or more. Report a p-value from your work or the news. Even if a p-value of less than 5% is commonly used in social science research, when would you want to use a more stringent significance level cutoff, such as alpha=1%, or even .1%? When might you be willing to relax the standard to 10%? Explain with an example.