PowerPoint-Präsentation

PowerPoint-Präsentation

INVESTIGATING THE OTHER-RACE EFFECT USING MULTINOMIAL PROCESSING TREE MODELS Henrik Singmann David Kellen Christoph Klauer Johannes Falck Learning Phase: HOUSE BOAT FOOD SEAL RAIN Test phase: RAIN CAT KING FOOD old? SEALHEAVE new? old? new? old?

N new? old? new? old? new? old? new? Learning Phase: HOUSE BOAT Hit CAT SEAL RAIN Old Word RAIN FOOD New Word KING FOOD

old? SEALHEAVE new? old? new? old? N new? old? new? old? Miss new? False old? new? Alarm Correct Rejectio n Learning Phase: P(hits) + P(misses) = 1 P(false alarms)HOUSE + P(correct rejections) = 1 BOAT FOOD = two independent data points SEAL Old Word RAIN Hit CAT

RAIN New Word HEAVE N KING FOOD SEALold? old? new? old? new? old? new? new? old? new? old? Miss new? False Alarm Correct Rejectio n Two item classes: central-european faces turkish/arabic faces Learning phase:

Test phase: Data: old? new? hits false alarms 02/29/20 20 own-race faces old? new? old? new? other-race faces old? new? old? new? old? new? 5 Two item classes: central-european faces turkish/arabic faces

Learning phase: Test phase: Expected data: old? new? hits false alarms 02/29/20 20 own-race faces old? other-race faces > new? old? (1.4 ) < Other Race Effect new? old? new? (1.6 ) (ORE) (Meissner & Brigham,

2001) old? new? old? new? 6 RESEARCH QUESTIONS ORE in Germany with Turkish/Arabic faces? Turks/Arabs largest ethnic minority (Statistisches Bundesamt, 2011) most-researched other-race: Blacks (USA) Source for ORE: memory- or response processes? Memory processes: better memory = hits + false alarm worse memory = hits + false alarm Response processes: bias "old" = hits + false alarm bias "new" = hits + false alarm 02/29/2020 7 EXTENDED 2HTM Do Old Items "old" gunsure 1 - Do 1

gunsure Dn New Items "unsure" gold/new "old" 1 "new" gold/new gunsure 1 - Dn 1 gunsure "new" "unsure" gold/new "old" 1 "new" gold/new EXTENDED 2HTM Do

Old Items 1. detection "old" gunsure 1 - Do 1 gunsure Dn New Items "unsure" gold/new "old" 1 "new" gold/new gunsure 1 - Dn 1 gunsure "new" "unsure" gold/new

"old" 1 "new" gold/new EXTENDED 2HTM Do Old Items 1. detection 2. uncertaint y "old" gunsure 1 - Do 1 gunsure Dn New Items "unsure" gold/new "old" 1 "new"

gold/new gunsure 1 - Dn 1 gunsure "new" "unsure" gold/new "old" 1 "new" gold/new EXPERIMENTS 1 (N = 42) & 2 (N = 36) 100 pictures each (turkish/arabic and white): from websites of turkish, arabic, or central european football teams (no known leagues) Matched after pretest: ethnicity, valence, distinctivness Exp. 1: Singmann, Kellen, & Klauer (2013), CogSci Proceedings 11 RESPONSE PROPORTIONS Exp. 2:

Exp. 1: White 1 0.8 Arabic *** 0.8 0.6 0.4 0.4 * 0 t hi s e r s i u s m

n u *** FA re u s n u White 1 0.6 0.2 : p < .1 *: p < .05 **: p < .01 ***: p < .001 * C R *** **

0.2 0 Arabic t hi s e r s i u s m n u *** ** FA re u s n u C R error bars: 95%-Cosineau-Morey-Baguley within-subjects CIs Hierarchical Bayesian

(posterior grandmean ) Exp. 1: * Exp. 2: * EXPERIMENT 3 (N = 37) 100 pictures each (black and white): Color FERET (Facial Recognition Technology) database RESPONSE PROPORTIONS White 1 Black * *** 0.8 0.6 0.4 ** ** miss FA

* 0.2 0 hit unsure unsure CR error bars: 95%-Cosineau-Morey-Baguley within-subjects CIs Hierarchical Bayesian (posterior grandmean ) * RESEARCH QUESTIONS: TAKE HOME MESSAGE ORE in Germany with Turkish/Arabic faces? YES, but other than expected (replicated for Black faces) Source for ORE: memory- or response processes? Some evidence for influence memory processes. Stronger evidence for differences in response processes (stronger "old"-bias for other-race faces) Preliminary results point towards different ORE groups. 02/29/2020 17

Bild durch Klicken auf Symbol hinzufgen THANKS TO MY COLLABORATORS EXP 1 & 2 (N = 78) * *: p < .05 **: p < .01 ***: p < .001 EXP 2: RESPONSE PROPORTIONS Exp. 2 (binary): 1 White Arabic *** 0.8 1 0.8 0.6

0.6 0.4 0.4 *** 0.2 0 t hi s e r s i u s m n u FA re u s n u Exp. 2

(ternary): White * C R *** ** 0.2 0 Arabic t hi s e r s i u s m n u *** ** FA

re u s n u C R error bars: 95%-Cosineau-Morey-Baguley within-subjects CIs Exp. 2 White faces Do Dn gu Arabic f. go/n Do Dn gu p < .05 p < .1 Dn gu go/n

Do Dn gu go/n .7 0 .55 -.6 0 .9 0 White faces Do Arabic faces White faces Arabic f. Exp. 1 Do Dn gu go/n

Arabic faces gu go/n gu .9 1 -.3 1 go/n -.3 3 .6 0 Do CORRELATIONS White faces White faces .5 6 Exp.D3 n gDu o

.6 1 White faces Do Dn gu go/n .5 5 Dn gu go/n Dn .6 5 Dn Do Do .35 Black faces Do

Dn .7 5 .3 9 .5 2 .5 2 gu go/n -.3 4 .8 6 .6 0 Hierar. Bayesian (posterior ) MLE (individual fits) Exp. 1

** * * Hierar. Bayesian (posterior ) MLE (individual fits) Exp. 2 ** * * Hierar. Bayesian (posterior ) MLE (individual fits) Exp. 3 ** * * Hierarchical Bayesian (posterior grand-mean )

RESULTS EXPERIMENT 1 * Hierarchical Bayesian (posterior grand-mean ) RESULTS EXPERIMENT 2 *

Recently Viewed Presentations

  • Young Lee Amazon Case Study - California State University ...

    Young Lee Amazon Case Study - California State University ...

    History. Started by Jeff Bezos in 1994. Started his career as a product manager in banking . VP of D.E. Shaw & Co (hedge Funds) Lost 3 Billion in first 7 years. Many skeptical of ever making it as a...
  • HR PAYROLL SYSTEM Originally known as BEACON HR

    HR PAYROLL SYSTEM Originally known as BEACON HR

    Under "My Working Time," you can enter working time and leave, request leave from your supervisor, and view and print your leave quota balances. Remember, ONLY employees at locations that are authorized to enter time/leave shall enter work time and...
  • Diapositiva 1 - Le Baobab Bleu

    Diapositiva 1 - Le Baobab Bleu

    À désigner des objets ou des personnes présents dans la situation: C'est qui l'administrateur? Celui-là À éviter des répétitions: Quel homme politique je choisis pour l'Europe? Celui qui a des projets d'avenir. Les pronoms démonstratifs sont toujours accompagnés: d'un adverbe:...
  • Romanticism - tamdistrict.org

    Romanticism - tamdistrict.org

    Return to Christian Mysteries Elohim Creating Abraham William Blake, 1805 Body of Abel Found by Adam and Eve William Blake, 1825 The Cathedral Gaspar David Friedrich, 1818 The Paradox of Romanticism: Romanticism could reinforce the greatest themes of political liberalism...
  • Formal Software Engineering Processes

    Formal Software Engineering Processes

    Disparate and Heterogeneous Systems Used During The Design, Construction, Commissioning, Operation, and Maintenance of an Experimental Physics Facility (EPF) ... Process Requests. Naming Convention - Reports . Screenshots - Unit Conversion ... Control System Studio, EPICS V4, ChannelFinder ...
  • DIMENSIONS - unifr.ch

    DIMENSIONS - unifr.ch

    1972-75 Primary school teacher (Locarno) 1975-79 Graduation in Biology UNI ZH 1979-82 PhD curriculum UNI ZH, Mol. Bio. 1982-84 Research assistant UNI ZH 1984-86 Postdoc UCSF, K Yamamoto, (S. Francisco) 1987-91 Principal Investigator, Privaztdozent, UNI ZH 1994-today Professor Biochemistry UNI...
  • Vocabulary Development

    Vocabulary Development

    What is a synonym for the word . retrieving? A. getting. B. turning. C. locking. D. polishing. CONNOTATION AND DENOTATION. The canal was so parched that the fish lined up. with their towels to take turns swimming in the few....
  • Charter Critics Revisited

    Charter Critics Revisited

    On the right: F.L. Morton, Rainer Knopff Charter is used for "social engineering" by left-wing groups Too much emphasis on social equality, not enough on individual liberty "court party" activists capture Charter litigation Richard Sigurdson: Since 1982, most criticism of...