Sequential Analysis of Nominal Data in Mother-Infant Communication: Quantifying Dominance and Bidirectionality
| Autor: | Jan B. Hoeksma, Y. van Beek, Brian Hopkins, B. De Roos |
|---|---|
| Rok vydání: | 1992 |
| Předmět: |
Contingency table
business.industry Transition (fiction) Information theory Machine learning computer.software_genre Behavioral Neuroscience Dominance (ethology) Animal Science and Zoology Artificial intelligence Psychology Contingency business Social psychology computer Preference (economics) Reliability (statistics) Event (probability theory) |
| Zdroj: | Behaviour. 122:306-328 |
| ISSN: | 1568-539X 0005-7959 |
| DOI: | 10.1163/156853992x00561 |
| Popis: | In recent studies on early communication a preference seems to exist for applying statistical methods to behaviours that have in some way been scaled. As an alternative, two methods for analyzing the interaction between two sequences of nominal data are discussed: an information-statistical analysis as developed by VAN DEN BERCKEN & COOLS (1980a) and log-linear modelling. Both methods make use of contingency (transition) tables, which should be constructed according to common requirements that are often ignored in practice. Firstly, the relationships between the number of behavioural categories, the length of the recording, the complexity of the model and the reliability of the statistics need to be considered. Secondly, choices have to be made regarding the use of either time or event sequences. Although these requirements are explained and illustrated using theoretical considerations and data on early mother-infant communication in humans, the application of these models may be much broader. Since informational and log-linear statistics are mathematically related, both can be used to describe the same measures of directionality in the interaction of individual pairs. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|