Chapter 9: Introduction to the t Statistic (One Sample)

fo-r

nł+/ c.-L -

2,/.. o Co.TG cr·1t/ co-/ .Rel1on u.s in.. '1

r'f', Ł-t o./ .st r/ bL/ -ion -/ O-b/ c.. l-d-.t/ ·LC, S IRL 0,./

'C / '

- novj c

M

.)

-.tr.. F-e x.a.. m_p/ Ł

, o t co nL, o

-t_l/ J/ $" _

b I/ .N /

and Ł __ tJ

I/

problems with z-scores

require more info than is usually available; requires knowledge of the population when usually you only have sample data

t-statistic

alternative to z-score - considered "approximate" z

t-distribution

family of distributions, one for each value of df; approximates the shape of the normal distribution - flatter and more spread out

hypothesis tests with a t-statistic

1. state the null and alt hypotheses; 2. locate critical region using the t distribution table, df, and desired alpha level; 3. calculate t-statistic; 4. make a decision regarding the null

estimated Cohen's d

uses sample SD instead of population parameters

confidence intervals

treated sample mean is an estimate of treated population mean

CI steps

1. pick a degree of confidence; 2. use t-distribution table to find the value of t; 3. set up and solve; 4. you are now x% confident that the mean fall within this interval

factors affecting CI

more confidence = larger interval; less confidence = smaller interval