The magic numbers 4 and 7. Modeling chunking in immediate memory


Lörch, Lucas


[img]
Vorschau
PDF
TeaP_2019_talk.pdf - Veröffentlichte Version

Download (408kB)

URL: https://ub-madoc.bib.uni-mannheim.de/49653
URN: urn:nbn:de:bsz:180-madoc-496530
Dokumenttyp: Präsentation auf Konferenz
Erscheinungsjahr: 2019
Veranstaltungstitel: 61. Tagung experimentell arbeitender Psychologen (TeaP)
Veranstaltungsort: London, UK
Veranstaltungsdatum: 15.-17.4.2019
Sprache der Veröffentlichung: Englisch
Einrichtung: Fakultät für Sozialwissenschaften > Bildungspsychologie (Münzer 2012-)
Lizenz: CC BY 4.0 Creative Commons Namensnennung 4.0 International (CC BY 4.0)
Fachgebiet: 150 Psychologie
Freie Schlagwörter (Englisch): working memory , expertise , chunking
Abstract: The main assumption of chunking theory is that knowledge about semantic units in a certain task domain can be used to compress incoming information. Although the maximal capacity of immediate memory is 4±1 information units, this information compression allows to enlarge the memory capacity by increasing the size of information units. As even random information can be compressed to a certain extent, a capacity limit of 7 single bits of information is typically found with random material. These assumptions were tested with a complex span task. Expert musicians (n=75) memorized the pitch of a single note and then played a simple melody on a piano at first sight. This procedure was repeated twelve times and then a memory test followed in which all memorized notes had to be recalled in correct order. The presence of semantic units was varied within-participants: in experimental trials, consecutive notes formed major chords, while in control trials they did not. As expected, mixed model analyses revealed that memory capacity was larger when the material contained semantic units. In experimental trials, the capacity limit was 3.18 chords while in control trials it was 7.32 single notes. The TBRS*C computational model integrates time-based decay, refreshing mechanisms and chunking. It is explicitly suited to model recall performance in complex span tasks. This model will be used to analyze the data in order to confirm its validity.




Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt.




Metadaten-Export


Zitation


+ Suche Autoren in

+ Download-Statistik

Downloads im letzten Jahr

Detaillierte Angaben



Sie haben einen Fehler gefunden? Teilen Sie uns Ihren Korrekturwunsch bitte hier mit: E-Mail


Actions (login required)

Eintrag anzeigen Eintrag anzeigen