Study Guides (248,019)
Canada (121,233)
Psychology (775)
PSYC 2002 (22)
all (5)

PSYC2002 Full Notes for the Entire Year

44 Pages
Unlock Document

PSYC 2002
All Professors

PSYC2002STATS MIDTERM NOTESLECTURE 1Statistics DefinedMathematical body of scienceCollection analysis interpretation explanations and presentations of observations and measurementsData symbols that represent some measureable aspectoData is plural data were obtainedoDatum is singular a datum was obtained SamplesStatistics describe samplesWords qualitativeNumbers quantitative PopulationsParameters describe populationsUse measurements from samples to infer onto populations Role of statisticsSummarize dataCreative quantitative relationshipsMake predictionsOrganize and communicate facts and findingsTest hypotheses Hypothesis TestingProbability of a descriptive statistic being due to chanceNull hypothesisoDifference is due to chance and not to a real differenceoRetain it when p05oCoined by Fishero1st experimentlady tasting teabelieved she could tell if milk or tea was poured in firstoIf p must be5 to reject null hypothesis and there are 70 permutations theres a 14 chance of correctly guessing which 4 cups had tea in it firstshe guessed the 4 correctly and theres a 14 chance of her doing so thus she must have a skillTruth tableoType 1 error when null hypothesis is rejected but it shouldnt have been probability of making type 1 error is known as alphaoType 2 error when null hypothesis isnt rejected but it should have been probability of making a type 2 error is known as beta Type 1 and 2 errors can be known as mistakes but a decision error rejectingnot rejecting cannot be known as a mistake Power of a Statistical TestAbility of stats to detect an effect when the effect is actually presentProbability that test will result in rejection of null hypothesis when null hypothesis is falsePower1betaTo get necessary statistical power need a lot of participantsAs power increases chances of getting a type 2 error decreases Goldilocks Principle for Statistical PowerUnderpowered studiesoEffects of importance may not be detectedoEspecially in a type 1 error conclude theres an effect when there actually isnt an effectoWhen type 1 error is present the effect size will appear larger in a small n design than in a large n designoLeads to inconsistent resultsOverpowered studiesoWaste resources time and money may also be unethicaloThe closest to 1 you are the stronger the powerJust right studies ie appropriately poweredoBalance the risks of type 1 and type 2 errorsoAim for a power of about 08 4 Ways to Increase Test PowerIncrease probability of making type 1 errorChange 5 to greater numberoCan strengthen the difference making it easier to detectDecrease variability of dataoNarrow parametersoCareful selectionoCareful trainingoPretesting of specific criterionoPerform experiment extraneous variables are controlledIncrease sample sizeoMore data gives more info Parameter descriptive term for something that cant be measured directlySample used in descriptive statsSampling frame who what where when and howRandom purposeful stratified convenience representative etc Variables 6Measurable quantity that variesDependent variableoValue depends upon experimental hypothesisoReflects any effect caused by the manipulation of independent variableoExperiment with only one dependent variable univariate experimentIndependent variableoValue controlled within experimentoMust be at least two levels in order to compareoAlso known as a factorIntervening variableoAbstract process linking independent and dependent variablesoNot necessarily observable or measuredoeg if teaching technique is IV and mastery of teaching is the DV then the learning processes is the intervening variableModerator variableoModerates the intervening variable therefore affecting the relationship between the independent and dependent variablesoCan be controlled to certain extentoeg gender age cultureControl variableoMust be held constant balanced or eliminated to test impact of independent variableExtraneous variablesoAlternative explanation for dependent variableoIf cannot be controlled must at least identify them and take them into consideration when interpreting resultsoeg confounds ConstantsValues that dont changeOften used for scaling purposesMeasured variable can become a constant eg sample mean Measurement Scales 4Nominal scaleoAssign a unique symbol to an event as the measurementoBut there is no quantity associated to the symboloCan only count the symbols and sort them into frequency distributionsoeg license plate number phone number social insurance numberCannot find the mean or standard deviation of these symbols
More Less

Related notes for PSYC 2002

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.