Predicting Outcomes From Conversational Dynamics Within the First 5 minutes

Thin Slices

Curhan & Pentland

Power of First ImpressionsThin Slices

Evidence that conversational dynamics might play a critical role in negotiation

Formal Microanalyses of highly specific speech features

Predictive validity of computers

✓

7

6

5

4

3

2

1 First impression based on 30sec and 6sec of silent video

Interviewer’s impressions are formed in the early stages of the interview and tend to persist throughout the interaction

Degree to which 4 conventional dynamics occurring within the 1st 5min of a two-party, simulated employment negotiation predict the outcomes of that negotiation

Behavioral outcome of a Transactional Negotiation

Have been shown to predict a broad range of consequences

Key to success lies in understanding social signaling, often nonverbal in nature

Thin Slices of behavioral data

Negotiation context, Pentland hopes to provide a useful diagnostic instrument that might facilitate future research on negotiation processes as well as applications for training and evaluating negotiators.

Computer Algorithms- Thin Slice

1 2

Nonlinguistic social signals, body language, facial expressions, & tone of voice are as important as linguistic content in predicting behavioral outcomes

Which social signals participants might use to judge outcomes?

That have already been suggested in the literature with behavioral outcomes, to determine which signals (if any) have predictive power similar to that of human judges.

Compare candidate signal features

3How these predictive social signals relate to existing theories of mental function and social interaction.

Examine

Social Signaling and Conversational

Dynamics2

1

Curhan & Pentland Research

PENTLAND’ S FOUR MEASURES OF VOCAL QUALITY AND CONVERSATIONAL

INTERACTION THAT COULD SERVE AS PREDICTIVE SOCIAL SIGNALS

An individual’s activity level during the first 5 min of the negotiation will be positively correlated with his or her own individual outcome.

Hypothesis

Time a person speak. Studies’ involving competitive settings, speaking time was positively correlated with dominance over the outcome

Activity

1 2

PENTLAND’ S FOUR MEASURES OF VOCAL QUALITY AND CONVERSATIONAL

INTERACTION THAT COULD SERVE AS PREDICTIVE SOCIAL SIGNALS

An individual’s level of engagement during the first 5 min of the negotiation will be positively correlated with his or her own individual outcome.

Hypothesis

Measured by the influence that one person has on the other’s conversational turn taking. (Markov process).

One sided engagement occurs when one person is energetically questioning another and the other begins speaking only after the questioner ceases speaking.

Engagement

1 2

PENTLAND’ S FOUR MEASURES OF VOCAL QUALITY AND CONVERSATIONAL

INTERACTION THAT COULD SERVE AS PREDICTIVE SOCIAL SIGNALS

An individual’s level of emphasis during the first 5 min of the negotiation will be positively correlated with his or her own individual outcome but positively correlated with the counterpart’s individual outcome.

Hypothesis

Measured by variation in speech prosody, specifically, variations in pitch and volume.

Pentland & Curhan emphasis measure may be an indication of the importance a speaker attaches to the interaction.

Emphasis

1 2

PENTLAND’ S FOUR MEASURES OF VOCAL QUALITY AND CONVERSATIONAL

INTERACTION THAT COULD SERVE AS PREDICTIVE SOCIAL SIGNALS

Hypothesis: an individual’s freq. of mirroring during the first 5 min of the negotiation will be positively correlated with his or her own individual outcome.

Hypothesis

Behavior of one individual is mimicked or mirrored by another, this could signal empathy, which has shown to positively influence the smoothness of an interaction as well as mutual liking.

Mirroring

1 2

Past studies relied on human

2

Method

Participants engaged in a second, multi-issue employment negotiation task, which were digitally recorded. Conversational speech features of the first 5 min of dialogue were taken by a computer using algorithm.

Research findings

Past studies relied on human intuition

Conversational dynamics associated with individual success among high status parties tended to be different from those associated with individual success among low status parties.

Technology can be used to give real time feedback to diagnose and improve individual negotiation skills

Microcoding using a computer it may ensure high test-retest reliability. Because computer measures objective, physical property of the audio signal, not subjective or psychological properties of the measures are 100% consistent.

Present study uses computer algorithms to identified specific features of thin slices that correlated with subsequent behavioral outcomes

Pros

Conclusion

1

2

3

Cons

No real evidence to prove the success of technology based on the algorithms used in this research is a better predictor of thin slice outcome then human judge predictors

Research increase the original concept of Thin Slice 30sec and 6sec to 5min first impression to predict outcome

For more complicated speech features, the link between specific measures and their respective conversational dynamics is more tenuous

1

2

3

Notwithstanding this potential shortcoming, one advantage of microcoding using a computer is that it ensures high test-retest reliability.

4

Going beyond Mining Social Networks

Wearable Networks

The belief that by adding mathematical understanding of behavior based on the feature analysis, will enhance distance-separated team work behavior, collaboration, decision-making and the sharing of social experiences.

GRS

Group Network

Social Intelligence

THANK YOU!

Your Logo