Introduction to Statistics inPsychologyPSY 201Professor Greg FrancisLecture 01How to avoid being a schmuck.MAKING JUDGMENTSwe have to make judgments all thetime• Do nicotine patches help people stopsmoking?• Is Pepsi better than Coke?• How is alcohol consumption relatedto depression?• Is this a good buy for a stereo?• Is diabetes related to arsenic?2A PROBLEMpeople are not very good at answeringthese kinds of questions. we makesystematic errors (take PSY 200, PSY285 or PSY 318)people in the “know” can takeadvantage of these tendencies• politicians• retailers• drug companies• “activists”3SIGNIFICANCEpeople are easily influenced by thecontext in which they make decisionsthis is a problem, because context iseasily (and subtly) manipulatedto avoid being a schmuck you need tolearn how to make decisions properlySTATISTICSit is not always easy...but it is worth it4AVOIDING COLDSMany p eople recommend the herbechinacea to reduce the severity ofcolds and/or to increase yourimmunity to getting a coldWhat should happen if echinacea doeswork?What should happen if echinacea doesnot work?Why is it popular?5A GOOD STUDYthere appears to be a link between low-levelsare arsenic and type 2 diabetes788 adult medical tests found a fourfoldincrease in the risk of diabetes for thoseexposed to low-levels of arsenic in their urine(compared to people with even lower levels ofarsenic)previous studies have found high levels ofarsenic in some drinking watera follow up study with 4000 people is planned6A GOOD STUDYBut, even a good study has limitsarsenic comes from lots of different sources: drinkingwater, coal burning, seafoodit’s not known how arsenic could contribute to diabetesit could be that people with diabetes tend to retainarsenic in their urinethe press (CNN) tries to scare people aboutthe issueno indication of the actual risk numbers7A BAD STUDYThe Daily Express (subheading “TheWorld’s greatest newspaper”) ran astory claiming statistical evidence thatsuicides were higher around mobilephone masts8A BAD STUDYAccording to the quoted researcher: masts areplaced an average of 800 meters from eachhome across the country, while for 22 (highlypublicized) suicides the masts were an averageof only 356 meters awayBut what does this actually mean?How was “average” distance computed?Mobile phone masts are not distributed evenly acrossthe countryside, so is an average even a reasonablemeasure of anything?Maybe more suicides are in urban areas, which alsohappen to have more mobile phone masts.If you look long enough you can find something incommon among any number of people.9COURSE GOALSWhat are the differences between goodand bad studies?What are the limitations of goodstudies?What is needed to get goodexperimental data?How is statistics used in psychologicalsciences?10COURSE GOALS1. Descriptive statistics• How to describe data.• Using graphs.• How to summarize da ta.2. Inferential statistics• Hypothesis testing.• Comparing descriptive statistics11WHY IS IT HARD?several reasons1. Little differences in presentation canmake a big difference in understand-ing.2. It is hard to get good measurements.3. It involves mathematics.4. It goes against our intuitions (anec-dotal evidence).5. If you don’t ask the right type ofquestion it is worthless.6. Sometimes the answer is “I don’t know.”12COURSE OUTLINEstatistical termsdescribing dataidentifying relationshipsEXAM I (10%)probabilityhypothesis testingestimationEXAM II (10%)hypothesis testingstatistical powerEXAM III (10%)ANOVAFINAL (15%)(cumulative)Beware scheduling of the final exam!13TEXTOnline Statistics: An InteractiveMultimedia Course of Studydesigned for the behavioral sciencesfree!You will earn course credit by finishingreading assignments by the due dategiven in the sy llabus (10% of yourclass grade)You must login to earn the coursecredit and stay logged in for thespecified duration required to read thematerial (rarely more than 10minutes).14TEXTprovides a pretty good introduction tomany topics we want to discusssometimes weak in presenting exactlyhow to carry out the computationsour discussion follows a different order,so we jump around the chapters a bitIf you log off, or point the web browserto a different page, you have to startover15LECTURE NOTESreduced format of 6 slides to a pageavailable before lecture!available on the world wide web inPortable Do cument Format (readableby Adob e Acrobat)http://www.psych.purdue.edu/∼gfrancis/Classes/PSY201/index.htmlnecessary for some topics we willdiscuss16HOMEWORKhomework problems are available fordownload from the class web sitedue on the day specified in the syllabusif turned in by the next class, it isworth 1/2 credit.any other time is worth no credit.17HOMEWORKgrading will focus on whether you aredoing the problems correctly, not juston whether you get the right answer.make sure that you show that youknow what you are doingtotal homework grade is worth 20% ofyour class grade18STATLABa new part of the class is a set ofon-line experiments that measure someaspect of your behaviorat the end of each experiment you areguided through a statistical analysis ofthe resultsYou need to know what you are doingto complete the statistical analysis.When you finish the lab (answer allthe statistical question correctly), youget credit for that labfree!19STATLABYou will earn course credit by finishinga lab assignment by the due date givenin the syllabus (10% of your classgrade)You must login to do the labs.20COMPUTERSsome software packages handle lots ofstatisticsI don’t recommend you use them tocomplete the homework problemsbecause you need to show intermediatestepsa spreadsheet program like MS Excelis pretty useful for computing termsquickly, but be sure you know how tocompute things by hand. Excel isespecially useful for the StatLabassignments.you will not be able to use thepackages on the exams!21CLASS PARTICIPATIONevery now and then I requireassistance during classI will go through the class list and picknames at random. I will note if you arepresent or not and that will counttoward your class participation grade.Your total class participation grade isworth 10% of your final class grade.I view the class participation grade as“gimme” points!22GRADINGstraight scale, no rounding• 98–100% A+• 93–97% A• 90–92% A-• 88–89% B+• 83–87% B• 80–82% B-• 78–79% C+• 73–77% C• 70–72% C-• 68–69% D+• 63–67% D• 60–62% D-• < 60%