I am going to make a couple statements… please let me know if they are incorrect. *The standard error is always computed: std. dev / n^.5 *The test statistic is always computed (X - hypothesized value) / (std. dev / n^.5) *Constructing confidence intervals: X + and - {critical value * (std. dev/n^.5)} I am finally starting to feel okay with this jargon… please tell me my sprouting confidence has a reasonable foundation! But seriously, can someone please let me know if this is incorrect? Thanks
*The test statistic is always computed (X - hypothesized value) / (std. dev / n^.5) Thats wrong. There exist hundreds of test statistics. *Constructing confidence intervals: X + and - {critical value * (std. dev/n^.5)} X is the sample mean and what you call the critical value is just a so called fractile/quantile. Normally we don’t speak about critical values in this context (It’s a term from test theory.) The formula is only valid for µ. Other parameters would need in many cases others formulas.