Statistical Sciences 2244A/B Lecture 10: 2244- lec 10
15 Nov 2016
School
Department
Professor
Document Summary
Take s and use it as best estimate for sigma. use s/rt(n) to estimate sigma. Once you estimate sigma you cant use z scores, now we use t statistic which uses s and x-bar. Standard error (t=) is the estimate standard deviation of the sampling distribution of means. T comes from a t distribution (student"s t distribution, guy who made this his name is student) Need to know sample size to use t. The mean of every t distribution is 0 so there are positive t values and negative. On each graph the red is z distribution and the black is t distribution. As degree of freedom increases t and z become more similar. Z has standard deviation of 1 and t standard deviation depends on the sample size. In t distribution there are more values farther from the mean meaning the standard deviation is bigger in t. T approaches normality as sample size approaches in nity.
