MultiView: Improving Trust in Group Video Conferencing through Spatial Faithfulness

BERKELEY INSTITUTE OF DESIGNMultiView: Improving

Trust in Group Video

Conferencing through Spatial

Faithfulness

David NguyenJohn CannyUC Berkeley

ACM SIGCHI 2007San Jose, CAApril 28 – May 3

Perspective Invariance

Perspective Invariance

Perspective InvarianceMona Lisa

(the painting)

Your viewingposition

YourPerspective/

Leonardo DaVinci’sPerspective

21 3

CL R

21 3

Perspective Invariance andGroup Video Conferencing

21 3

21 3

21 3

CL R

21 3

Perspective Invariance andGroup Video Conferencing

21 3

21 3

The Apprentice…

Donald Accidentally Fires Entire Staff

Video

Conference

System to

BlameYou’re

Fired!

Effects of Video Conferencing

• Turn Taking(Vertegaal et al., 2000)

• Cooperation(Bradner and Mark, 2002)

• Persuasion (Bradner and Mark, 2002)

• Deception (Bradner and Mark, 2002)

• Trust(Bos et al., 2002)

Research Question

How do spatial distortions affect trust formation between two meeting groups?

Before that, we need a spatially faithful video conferencing system.

Our Approach

Our Approach

Gen 1

Gen 2

Gen 3

Gen 4

Our Approach

MultiView: Spatially Faithful Group Video Conferencing

Cameras

Projectors

MultiViewDisplay

1 32

L RC

1 32

1 32

1 32

1 32

L RC

1 32

1 32

1 32

Research Question

How do spatial distortions affect trust formation between

two meeting groups?

Trust Measure: DayTrader (Bos et al, 2002)

• Daytrader is a measure of trust.

• Daytrader is an Iterated Prisoner’s Dilemma (PD) Game with Noise

• There were >30 rounds. In each round, groups were given 60 credits

• They chose how much to invest cooperatively and how much to keep individually

• Cooperative investments had a fluctuating market (average 50%) and split evenly, regardless of initial cooperative investment by each team (teams only knew their earnings)

• Every 5 rounds, a bonus is split between the two teams. The higher a team’s earning, the bigger their share of the bonus

0 60

0

60

A

B

60

10560

909045

45

105

Team A Investment

Tea

m B

In

vest

men

t

Does meeting through standard video conferencing affect trust formation when compared to face-to-face?

Can we improve trust formation patterns by using a spatially faithful video conferencing system such as MultiView?

Experimental Conditions

Face to

Face

MultiView

Standard Video Conferencing

Experimental Design

•N = 169 participantso 110 females, and 59 maleso 156 students (20), 13 staff members (39)o Formed 29 groups of 2 and 37 groups of 3o Groups were randomly formed

Results: Overview

Results: Overall Trust

2600.09 2627.641928.28

Face-to-Face vs. non-Directional Video Conferencing, p<0.05Directional Video Conferencing vs. Non-Directional Video Conferencing, p<0.05

Face-to-Face vs. Directional Video Conferencing, p>0.05

Results: Overall Trust

4.424 4.3883.562

Face-to-Face vs. non-Directional Video Conferencing, p<0.05Directional Video Conferencing vs. Non-Directional Video Conferencing, p<0.05

Face-to-Face vs. Directional Video Conferencing, p>0.05

Results: Overview

Results: Fragile Trust

-1.968 -2.348-4.520

Face-to-Face vs. non-Directional Video Conferencing, p<0.05Directional Video Conferencing vs. Non-Directional Video Conferencing, p<0.05

Face-to-Face vs. Directional Video Conferencing, p>0.05

Ability to Build Trust…… not always good!

Conclusions

Does meeting through standard video conferencing affect trust formation when compared to face-to-face? YES!

Can we improve trust formation patterns by using a spatially faithful video conferencing system such as MultiView? YES!

Acknowledgements

Nathan Bos

Questions?• Developed a spatially faithful video

conferencing system• Extended existing trust measure to support

group-to-group experimentation• Experimentally compared trust formation

patterns between groups meeting face-to-face, through standard video conferencing, and through spatially faithful video conferencing.

• Shown that spatial fidelity plays key role in trust formation between two groups meeting over video conferencing.

Prior Work: (Bos et al, 2002)

Trust Formation and CMC

Face-to-Face

Video

Audio Text

Results: Overall Trust

Results: Delayed Trust

Results: Fragile Trust

Face-to-Face vs. non-Directional Video Conferencing, p<0.05Directional Video Conferencing vs. Non-Directional Video Conferencing, p<0.05

Face-to-Face vs. Directional Video Conferencing, p>0.05

Construction

• Retroreflective LayerReflects image back in direction of source

• Vertical DiffuserDiffuses image vertically to accommodate varying viewing heights

• Antireflective/Antiglare Reduces distracting glares due to glossy surface and front projection setup

MultiView Directional Display

• Big, Bright, High Resolution Display

• Each view is provided by a projector

• The projected image is reflected directly back in the direction of the projector

• The image can be seen at varying heights only behind the projector

0

50

100

150

200

250

-30 -10 10 30

viewing angle (degrees)

illu

min

an

ce

(lu

x)

Illuminance vs. Viewing Angle

0O

α-20O

• JND Power Half Width = 7.5o (15”)

Planned Comparisons

•When doing pair wise comparison you can…o Perform and omnibus ANOVA followed by

pair-wise comparisons technique with proper adjustments

o Or, if you have specific comparisons driven by theory or prior data, you can use Planned Comparisons without any adjustment