Transcript Investigating the Effects of Electronic Self

Impacts of Electronic SelfPortrayal on Newly Formed
Virtual Teams
Shoshana Altschuller, Iona College
Raquel Benbunan-Fich, Baruch College
October 12, 2007
Motivation for Study

Dispersed team members interacting remotely via
electronic means are common.

Trust has a positive effect on group performance.

The challenge of attaining trust is increased using
computer-mediated communication systems (CMCS).

“True identity” exposure should have profound effects
on the way team members interact.

The Electronic self-portrayal continuum
Research Question

To what extent do changes in the selfportrayal characteristics of a synchronous
CMCS affect the outcomes of ad-hoc virtual
group meetings using the system?

Does electronic self-portrayal impact


trust ?
performance?
Electronic Self-Portrayal

Electronic self-portrayal (ESP)




The extent to which a CMCS portrays true identity (“true-to-life”)
Expression of “true self” on the Internet leads to close relationships
(Bargh et al., 2002; McKenna et al., 2002).
No research studies a CMCS in terms of its ability to reveal who
really is on the other end.
The way that a person is portrayed online is based on two aspects
of their online participation:

Personal Representation -- “who they are”



Photos
Avatars
Messaging-based Representation -- “how they formulate messages”

Rehearsability (Dennis & Valacich, 1999):

The extent to which a system allows users to reread and edit their own
messages before submitting them to be processed by the system.
Research Model
Mechanisms
Independent Variable:
Electronic Self-Portrayal
VERBAL
COMMUNICATION
PERSONAL
REPRESENTATION
MESSAGING-BASED
REPRESENTATION
Dependent
Variables
PERFORMANCE
NON-VERBAL
COMMUNICATION
TRUST
Hypotheses: Dependent Variable
Electronic self-portrayal impacts team trust based on Walther’s
Hyperpersonal Model (1997):

Ha. Communication using a more true-to-life mode of personal
representation will result in less trust than will communication using
a less true-to-life mode of personal representation.

Hb. Communication using a more true-to-life mode of messagingbased representation will result in less trust than will communication
using a less true-to-life mode of messaging-based representation.

Hc. Communication using a more true-to-life mode of only one form
of representation (“Partial ESP”) will result in more trust than will
communication using any other combination of representation
modes.
Methodology: Experimental Design
• 2x2 Factorial Design
Personal
Representation
Messaging-Based Representation
Avataridentified
Photoidentified
High
Rehearsability
Low
Rehearsability
(no keystrokes)
(keystrokes)
Methodology: Procedure

Ad-hoc teams of three Baruch students

Perform a given task together using the system




Each individual filled out pre- and post-task questionnaires


Decision-making task
Case scenario
Current, relevant to the students, ethically charged
All items based on previously validated and used instruments
Each group randomly assigned to a condition


Identification manipulations: avatar, photo
Rehearsability manipulations:


High – no visible keystrokes: users type one entire message at a time
before submitting them.
Low – visible keystrokes: users could see the letters appearing on the
screen as their partners type them.
Methodology: Manipulations

Avataridentified

Photoidentified
BACK
Methodology: Manipulations

Chat software: ICQ 2003b
Low Rehearsability
Split Layout
(visible keystrokes)
High Rehearsability
IRC Style Layout
(no visible keystrokes)
BACK
Data Analysis

Sample
 251 usable observations
 60-66 individual observations in each condition

Trust survey instrument
 Adapted from (Jarvenpaa et al., 1998)
 Confirmatory Factor Analysis
 Cronbach’s alpha = .90

HANOVA
 To test for differences in means
 Hierarchical - adjusts for group effects
High
Reh.
Low
Reh.
Avatar
60
65
Photo
60
66
Results Highlights
• HANOVA describes significant differences in the trust means among
treatments (p<.05).
•Ha and Hb are not supported.
•Trust is increased in the “Partial ESP” condition – Hc supported.
Crossover Interaction Effects
Photo
Low
Rehears.
(no keystrokes)
(keystrokes)
4.39
4.82
4.86
4.64
4.9
4.8
Trust Means
Avatar
High
Rehars.
4.7
Avatar
4.6
Photo
4.5
4.4
4.3
No Keystrokes
Keystrokes
Results Highlights

Partial electronic self-portrayal




at least one form of true representation and at
least one of non-disclosure
associated with higher levels of trust
ideal balance between idealization and realism
Low rehearsability (can see the keystrokes)


conveys the “person behind the avatar”
exposes the “fool behind the photo”
Results Highlights
Electronic self-portrayal continuum confirmed
Electronic Self-Portrayal Continuum
4.9
4.8
Trust

4.7
Representation
4.6
Rehearsability
4.5
4.4
4.3
no ESP
1
1 dimension
2
"true-to-life"ness
2 dimensions
3
Next Steps

Partial least squares analysis to test the
structural properties of the mechanisms in the
research model

Continued analysis in conjunction with
HANOVA on group performance
Contribution

Examines two dimensions of electronic selfportrayal and their interaction

First to empirically test effects of rehearsability

Practical implications for the use of
synchronous communication systems

The element of “unrealism” that exists in CMC
could be one of its most useful characteristics.