44 WORKING WITH SAMPLE DATA more than enough evidence to conclude the alternative hypothesis is true to warrant changing the status quo. A final but important note about forming hypotheses: while the contexts given in this book are predefined as appropriate for a one- or two-tailed test, the researcher should be careful about structuring a one-tailed hypothesis test. Two-tailed tests are generally most appropri- ate except where reason, historic record, or known effects persuade the researcher to structure and perform a one-tailed test. The Decision Rule: Setting Error Tolerance With hypotheses in place, the researcher next must play out the null hypothesis. The null hypothesis is assumed to be true. The researcher must ask: if the null hypothesis were true, what should we expect to see? In the formal hypothesis test, this step creates a decision rule. For exam- ple, if, in Scenario 1, the mean is equal to 6 ounces, then sample means will likely fall no further than some specified number of units of standard