2/10/15 What if you wanted to test… Factorial designs •  Whether FOXP2 knockout mice vs. regular mice show differenCal __________ in motor learning tasks vs. __________ learning tasks •  MulCple __________ –  KO and regular: motor task –  KO and regular: ________ task •  Or put it all into a __________ What if you wanted to test… •  Whether driving in ______________ condiCons is more impaired by visual or auditory __________ ______ •  MulCple experiments –  Driving in __________ : Vis and Aud –  Driving in __________ : Vis and Aud –  __________ : easy driving, hard driving –  __________ : easy driving, hard driving •  Or put it all into a __________ What if you wanted to test… •  Do French speakers and English speakers show different ______________in speech? –  Speech discriminaCon: French ________; Eng. ________ •  MulCple __________ –  French on French and English sounds –  English on French and English sounds •  Or put it all into a __________ 1 2/10/15 Simple factorial design: 2 __________ •  Two independent variables or __________ •  TesCng mulCple hypotheses –  Main effects: –  IV 1 effect (H0: __________ ) –  IV 2 effect (H0: __________ ) –  Interac3on: –  IV 1 x IV 2: effects of IV1 are __________ depending on ______________ (or vice versa) –  H0: __________ FOXP2 knockout and learning Simple factorial design: 2 factors •  Two independent variables or __________ •  TesCng mulCple hypotheses –  Main effects: –  Task difficulty: __________ –  Regular vs. KO mice: probably ________ worse –  Interac3on: –  ___________: motor learning shows bigger impairment than _________ for KO mice Simple factorial design: 2 factors Motor learning Regular mice FOXP2 KO mice 100 150 200 150 350 300 250 200 150 100 50 0 Regular or KO 200 200 Motor Olfactory 300 •  Two independent variables or __________ •  TesCng mulCple hypotheses 250 200 150 100 50 0 Regular KO mice Marginal means show you the main effects Trials to reach criterion performance Trials to reach criterion performance Task type Olfactory learning Trials to reach criterion performance 350 300 250 200 150 –  Main effects: –  Driving condiCons: _________ increase errors –  Task type: _________ increases errors (vs. aud) –  Interac3on: –  _________________: visual task increases errors _________ when driving condiCons are difficult 100 50 0 Regular mice FOXP2 KO mice Motor learning Olfactory learning 2 2/10/15 Driving and visual/auditory distracCon Easy driving Hard driving Driving condiCons 560 570 790 DistracCon type 705 685 900 800 700 600 500 400 300 200 100 0 900 800 700 600 500 400 300 200 100 0 900 800 700 600 500 400 300 200 100 0 Easy driving Hard driving French vs. English sounds French 50 94 ppts. English 95 ppts. Sound type: 72.5 71.5 Listeners’ language 72 72 Main effects and interacCons 100 Percent correct English sounds –  Main effects: –  Language background: _________ –  Speech sound language: _________ –  Interac3on: –  LB x SSL: French speakers beber at _________, English beber at _________ Auditory Visual Distrac3on is Auditory Visual French sounds Simple factorial design: 2 factors •  Two independent variables or __________ •  TesCng mulCple hypotheses Easy Hard driving driving Braking reac3on 3me (ms) 590 Braking reaciton 3me (ms) Braking reac3on 3me (ms) Auditory Visual 80 60 40 20 0 French English ppts. ppts. •  SomeCmes you’re looking for _________ . •  SomeCmes you’re looking for _________ . •  SomeCmes you’re looking for _________ . 100 Percent correct Percent correct 100 80 60 40 20 0 French English sounds sounds 80 60 40 20 0 ParCcipant speaks French ParCcipant speaks English French sounds English sounds 3 2/10/15 Examples of interacCons Examples of interacCons Parallel or not parallel? 0.2 Older adults 0 Exercise Young adults 0.2 Older adults 0 Lumosity Exercise 0.4 Young adults 0.2 Older adults 0 Exercise Lumosity 0.4 Young adults 0.2 Older adults 0 Exercise 0.6 0.4 Young adults 0.2 Older adults 0 Exercise 0.6 0.4 Young adults 0.2 Older adults 0 Exercise Lumosity What if you have both main effect(s) and interacCons? •  SomeCmes you want an _________ ! •  Or not. 0.4 Young adults 0.2 Older adults 0 Exercise Lumosity 0.8 0.6 0.4 Young adults 0.2 Older adults 0 Exercise 1 0.8 Lumosity 0.6 Lumosity 1 0.8 Flanker task performance 0.6 0.6 Lumosity 1 0.8 Flanker task performance 1 Flanker task performance 0.4 0.8 0.8 Lumosity TV 1 0.8 0.6 0.4 Young adults 0.2 Older adults 0 Exercise Lumosity TV Flanker task performance Young adults 0.6 1 Flanker task performance 0.4 0.8 Flanker task performance 0.6 Flanker task performance 0.8 1 Flanker task performance 1 Flanker task performance Flanker task performance 1 Parallel or not parallel? 1 0.8 0.6 0.4 Young adults 0.2 Older adults 0 Exercise Lumosity What kinds of factors? •  Totally between-­‐groups •  Totally _____________ –  (Distracted driving example) •  _______ design (AKA split-­‐plot design) –  _________ is within-­‐subjects –  _________ is between-­‐groups –  (Both the FOXP2 mouse example and the English/
French speech percepCon example are ___________) 4 2/10/15 Higher-­‐order designs •  All of the examples we have considered are “_________” or “2x2” designs. –  Analyze with a 2-­‐way Analysis of Variance (_______) •  If one factor has _________ , it’s a 2x3. –  Analyze with a 2-­‐way Analysis of Variance (_______) •  If there are _________ , could be a 2x2x2. –  Analyze with a 3-­‐way Analysis of Variance (_______) How many factors? •  You can use lots and lots, but gets harder and harder to ___________ •  As few factors as you can use and sCll _____ ___________ •  Why? Levels of ___________: –  The boy moved to Germany –  The boy __________ moved to Germany –  The boy __________________ moved to Germany How many factors? •  As few as you can use and sCll draw ___________ •  SomeCmes, some of your IVs are just controls or ___________ (you don’t want to see effect) •  SomeCmes, you really do want to see a 3-­‐way interacCon! –  *(Also depends on how you code your __________________) •  Typical n00b error is to _______________________________________ 5