Gonzales, A.L., Hancock, J.T., and Pennebaker, J.L.(2009).
Language style matching as a predictor of social dynamics in small groups
Communication Research, 37(1), 3-19
Main aim to validate an algorithmic method for measuring the amount of linguistic style matching as an p13’ efficient means of assessing mimicry from language and to determine whether such a metric could predict two social dynamics, group cohesiveness & task performance
Literature review on social interaction/affinity and mimicry and also synchrony.
Bernieri & Rosenthal, 1991 – p4 ‘people imitate each other’s nonverbal movements and gestures’
Association with positive relationships Chartagh & Burgh (1999)
More refs in the last paragraph of p5.
Pickering& Garrod (2004)
Verbal mimicry – ‘at the level of syntatctic structure or word-by-word matching’
Linked to team performance ( Koylowski & Ilgen, 2006)
Measuring verbal mimicry
requires a micro approach therefore very time consuming and difficult to measure reliably
This article proposes an automated alternative the linguistic style (LSM) metric based on two characteristics of language
- ‘Entire conversations can be parsed into psychologically relevant dimensions quickly and with a high degree of accuracy’
- Function words – ‘These do not contain semantic information rather they are the backbone of language and are characterised by’
- High frequency
- Context independent
- Non consciously ( KRO – how known?- ?Pennebaker & King, 1999, Chung & Pennebeker (2007 ) produced, therefore difficult for an individual to manipulate
The LSM measures the degree to which two or more participants are producing similar rates of function words – 9 categories
| Category | Examples |
| 1. Adverb | Completly, often |
| 2. Article | A,an,the |
| 3. Auxillary verb | Am,have |
| 4. conjunction | And, but, or |
| 5. Indefinite pronoun | It, those |
| 6. negation | No, not, never |
| 7. Personal pronoun | I, you, we |
| 8. preposition | At, for, into |
| 9. quantifier | All, few, some |
RQs
- Does verbal mimicry occur at equal rates in FTF groups and online groups?
- Does LTM predict cohesiveness and task performance
Method
Same sex groups (4-6) people, 34 CMC, 41 FTF. Strangers. Material reward.
10 minutes socialising
20 minutes complex information searching task require some co-operation but not collaboration as understood for the purposes of my thesis.
5 minutes post task discussion
at the end each participant was given an Interaction Rating questionnaire.
Results & Discussion
Note these are based on correlations
- LSM metric predictive of group cohesiveness irrespective of medium (CMC and FTF) , gender and group size. The more individuals liked a group, the more members of the group used the same function words.
- There was a positive relationship between LSM and task performance but only for the FTF groups
Other relationships
1. Language and cohesiveness
- Word count +
- First person plural (we) - authors explained this unexpected result as follows . Perhaps it is used deliberately in an otherwise non cohesive group.
2. Language and task performance
- Future orientated ( could, would, should, must, ought, and will) +
- Use of achievement orientated words ( success) - authors explained this unexpected result as follows . Perhaps it is used deliberately in an otherwise non cohesive group.
