Experiential product placements in Virtual Environments:

1
Experiential product placements in Virtual
Environments
Presence, information processing and advertising
effectiveness

• Dan Grigorovici
• Pennsylvania State University
• ICA 2003 Presence Panel

2
Roadmap

• IVE as invisible persuasive medium
• Persuasive effects of presence via physiological
  (arousal) cognitive moderators (depth level
  of processing)
• Application to persuasion advertising
• Commercial messages
• Product placements

3
Background

• Two-step theoretical model of persuasion-related
  effects is proposed1
• More presence ? more arousal and affect ? impact
  on depth of processing (user more likely
  to process information affectively, more
  implicitly and heuristically)
• Effect of making the user less aware of an
  embedded persuasive message ? moderating effect
  on various advertising-related outcomes less ad
  recall, but more positive brand attitudes and
  favorable purchase intentions.
• Product placements more effective than blatant
  advertisements
• 1 Grigorovici, D. (2003), Persuasive effects of
  presence in Immersive Virtual Environments, in G.
  Riva, F. Davide W. Ijjsselsteijn (Eds.), Being
  There concepts, effects and measurement of user
  presence in synthetic environments (Emerging
  Communication Book Series, Vol. 5) (pp. 191-208).
  Amsterdam, the Netherlands IOS Press.

4
Research objectives

• 1. empirical testing of an original model of
  cognitive effects of presence on persuasion
  (applications in advertising effectiveness,
  persuasive arguments, etc.)
• 2. establishing a link between form
  content-based presence and persuasion models
  (ELM, HSM)
• 3. empirical testing validating standardized
  presence scales using SEM (ITC-SOPI IPQ Witmer
  Singer)

5
Methods Study 1-pretest

• N of observations 248
• Sample size 31 undergraduate students
• 24 worlds varying in
• VE content arousal (high, medium, low)
• scene type (outdoor, indoor)
• sound (yes, no)
• interactivity (yes, no)
• random order and world viewing sequencing was
  performed

6
Methods Study 2

• 2 (task type) x 2 (VE content type) x 2 (3D ad
  type) x 4 (3D Ad product type) x 2 (3D Ad brand
  name) x 2 (product type) x 4 (3D Site ID) mixed
  design
• task type and VE content type
  between-participants variables
• 3D ad type, ad product type, ad brand name,
  product type and 3D site ID within-participants
  variables
• Eight 3D-VRML immersive Web pages 4 high
  arousal, 4 low arousal (pretested in Study 1)
• N 240 undergraduate students
• Stimuli

7
Methods Study 2

• Independent variables
• task type categorical, 2 levels
  experiential-navigation, search-selection
• VE content arousal ordinal, 2 levels high and
  low
• 3D ad type categorical, 2 levels a 3D product
  placement, vs. same 3D on a billboard
• 3D ad product type categorical, 4 levels soda
  can, car, cell phone, solar panel
• Dependent variables
• Presence
• ITC-SOPI
• IPQ Sense of Presence Questionnaire
• Cognitive capacity
• skin conductance response (SCR), heart-rate (EEG)
  and secondary task reaction time (RT) measures
  throughout the test period
• Attention
• via Skin Conductance Response (SCR) measurement,
  as well as through Electrocardiogram (ECG)

8
Methods Study 2

• Dependent variables
• recall and recognition memory
• memory for the 3D content was also assessed
  post-test with free and cued recall as well as
  recognition memory items included in the
  post-test questionnaire.
• attitude towards the site
• ad awareness
• eight recognition items, each with four
  multiple-choice, closed-ended response
• free recall
• attitude towards the ads
• ten 7-point Likert scale items derived from
  Machleit and Wilson (1988)
• attitude towards the brands (advertised in the 3D
  ads)
• three 7-point Likert scale items used by Muehling
  and Laczniak (1988)
• purchase intention three-item, 7-point Likert
  scale
• attitudes towards environment scale
• navigation patterns (movement speed and routes
  followed, stop timestamps, etc.) tracked in
  real time

9
Preliminary Results

• Spatial presence factor of ITC-SOPI positively
  predicts arousal level F (1,230)142.34, plt.0001
• When controlling for engagement, main effect of
  spatial presence on arousal is not significant
• Main effect of previous VR experience on spatial
  presence not significant
• Main effect of gender on spatial presence,
  engagement, naturalness non significant
• Main effect of gender on negative FX
  F(1,230)78.60, plt.0001

10
Preliminary Results

• Validation of ITC-SOPI, using first order CFA
  (Lisrel)
• Original model fits our data poorly, since Chi
  square (df1184)4930.14, p0.0 Chi square/df
  4.16 (gt3) RMSEA0.14 (90 CI 0.14, 0.14) gt
  recommended 0.05 AGFI0.41 CFI0.89 SRMR0.29
  ECVI30.53, with 90 CI 29.4131.47 (while ECVI
  for saturated model11.04)
• Suggested model modification
• Correlating SpPres-Eng SpPres-Natural
  Eng-Natural
• Chi square (df1181)5767.66, p0.00 Chi
  square/df4.88 RMSEA0.13
• Correlating error covariances between several
  observed variables from the initial ITC-SOPI
• The factors of ITC-SOPI are not
  independent/mutually exclusive
• Analysis of the physiological data (SCR, ECG)
  secondary reaction time measures
• currently on going

11
Preliminary Results

• Validation of ITC-SOPI, using first order CFA
  (Lisrel)

12
Discussion pattern of findings

• 1. Results from ITC-SOPI CFA suggest the need for
  further refinement of the scale using SEM
• 2. The engagement factor from ITC-SOPI presence
  scale seem to make an important contribution
  above and beyond the remaining 3 factors in the
  overall sense of presence ratings
• When controlling for engagement level, main
  effect of arousal on presence is not significant
• 2. When highly arousing, AD-related variables
  (Aad, Ad recall, etc.) do not show significant
  effects for product placements but they do for
  billboards (due to the nature of the
  ad-presentation) however, BRAND-related
  variables (brand perceptions, purchase intention,
  etc.), show significant positive effects for
  product placements, but not for billboards
• 3. The reverse is true for the low arousing VE
  content
• 4. Data analysis for the ELM-type persuasive
  argument manipulation is on going (expected
  findings similar to those at 3 4 above),
  which would confirm our model of cognitive
  effects of presence

13
Applications

• Is there a distinction between product placements
  and classical advertisements in IVE?
• NO, same 3D object can be both
• Task type differentiates between them

14
Further research (on going)

• Study 3
• 3 x 2 x 2 between participants design
• Task type 3 levels (navigation, search,
  cognitive)
• Immersion level high, low (fishtank VE,
  ImmersaDesk, CAVE)
• Stereoscopy (presence, absence)
• Expected findings
• Study 4
• Classic ELM persuasive argument manipulation in
  an environmental context 2 (presence) x 2
  (argument strength) mixed design
• Expected findings
• Significant interaction
• Study 3 4 currently in the design stage