Study Guides
(238,122)

Canada
(114,941)

University of Toronto St. George
(7,977)

Psychology
(682)

PSY201H1
(26)

Alison Luby
(7)

Quiz

# Lecture 8: Ch. 8 - Hypothesis Testing, Effect Size, Power

by
OneClass3700

Unlock Document

University of Toronto St. George

Psychology

PSY201H1

Alison Luby

Fall

Description

PSY201H1F L8Ex Previous exampleH800nd2 scientific alternative hypothesis H states 1Ch 8 Hypothesis Testingthere is effect of treatment on dep variableie there is change dif or relationship for I The Logic of Hypothesis Testing general popGeneral goal of a hypothesis test is to rule out Ex Previous exampleH 80 chance sampling error as plausible explanation 1meaningfullt dif from 80for results from a research studyFrame significance testing in terms of null Hypothesis testing allows us to use sample data to hypothesis because easier to prove something is draw inferences about pop of interestfalse than to prove it is trueHelps determine whether a specific treatment Step 2has an effect on the indivs in a popUse sample data to determine likelihood of H0 Can never know 100 if we are right but can being correctassign a statistical probability to the likelihood that Do by determining what sample means would be our conclusion is correctconsistent w Hwhat sample means would be 0inconsistent w itie look at distr of sample means to determine which means are near enough to popwhich are too far fromfor Hto hold0 Find where sample score is relative to cutoff on normal curveTrying to determine whether treatment has effect on the indivs in the popHowever usually only can give treatment to sample of indivs so use sample to make inferences about unknown treated popIf indivs in sample we administer treatment to are noticeably dif from indivs in original pop have evidence treatment has an effectTypically expect effect of treatment to be addition of constant to each indivs scoreMean changes but not shapeSDAlso possible that dif btwn samplepop means is simply sampling errorTo determine what is high vs lowprobabilityPurpose of hypothesis test is to decide btwn 2 sample select specific probability valueexplanationslevel of significance or alpha level level for 1Any dif btwn the samplepop can be hypothesis testexplained by sampling errorestablishes criterion or cutoff for making ie does not appear to be treatment effectdecision about H01The dif btwn samplepop is too large to be Usually set 05 5 01 1 or 001 explained by sampling error01ie does appear to be treatment effectEx 05 separates most unlikely 5 of sample means from most likely 95 of sample Steps involved in hypothesis testingmeansState hypothesis about pop05 most often used standardUse hypothesis to predict characteristics that az 196 unit table for p025 in the tailsample should haveWhen 2tailed In unit normal table look for Obtain random sample from poptail containing 2nd Compare obtained sample data w prediction in 2step to see if have support for hypothesis or whether hypothesis is wrongStep 1Step 1State 2 competing hypotheses always 2 about unknown popst1null hypothesis H states treatment 0independent variable has no effect on dep variable for popie pop mean after treatment is same as before treatmentHhypothesis 0zeroeffect

More
Less
Related notes for PSY201H1