Eye Tracking, Journal of Marketing Research

Post on 21-Jan-2017

1.903 views 0 download

Transcript of Eye Tracking, Journal of Marketing Research

Meißner, Musalem, Huber1 Eye Tracking and Conjoint Choices

Eye Tracking Reveals Processes That Enable Conjoint Choices to Become Increasingly Efficient with Practice

Martin Meißner, University of Southern Denmark, meissner@sam.sdu.dk

Andrés Musalem, Universidad de Chile,amulsalem@dii.uchile.cl

Joel Huber, Duke University,joel.huber@duke.edu

Meißner, Musalem, Huber2 Eye Tracking and Conjoint Choices

Choice task from our coffee maker study

Meißner, Musalem, Huber3 Eye Tracking and Conjoint Choices

To what extent does the focus on each alternative reflect the respondent’s utility for the

option?

Meißner, Musalem, Huber4 Eye Tracking and Conjoint Choices

• Mere exposure effect: Manipulated attention to objects results in greater liking. (Armel, Beaumel and Rangel 2008)

• Mere selection effect: Allocating attention to an object increases probability of choice. (Janiszewski, Kuo and Tavassoli 2013)

• Visually salient alternatives generate greater attention and are more likely to be chosen. (Milosavljevic, Navalpakkam, Koch and Rangel 2012)

To what extent does the focus on each alternative reflect the respondent’s utility for the option?

Previous research on alternative focus

Meißner, Musalem, Huber5 Eye Tracking and Conjoint Choices

• The central position of a new product on a shelf leads to a greater choice probability. (Atalay, Bodur and Rasolofoarison 2012)

Is there evidence that incidental attention distorts choice?

Centrality bias as incidental alternative focus distortion?

Meißner, Musalem, Huber6 Eye Tracking and Conjoint Choices

• Biased search: Information search and comparison processes are more favorably disposed towards the features of the leading object. (Willemsen, Böckenholt and Johnson 2011)

• Biased encoding: Once a favorite is identified then evidence supporting that choice is bolstered. Biases are most likely to occur when the decision is important, difficult or emotional. (Russo, Meloy and Medvec 1998)

Biased search as motivated feature focus distortion

Meißner, Musalem, Huber7 Eye Tracking and Conjoint Choices

• Unbiased search: In a binary choice task information supporting the leader was not different from a 50%-50% chance. No particular bias towards supportive features. (Carlson and Guha 2011)

• Absence of a pre-choice bias toward searching for preference-supporting information. Tendency towards balancing positive and negative information in self reported search strategy. (Chaxel, Russo and Kerimi 2013)

Is there evidence of greater attention to positive information for chosen alternatives?

Evidence against feature focus distortion

Meißner, Musalem, Huber8 Eye Tracking and Conjoint Choices

Study design and sample:• Questionnaire design: Explanation of attributes and

levels, 4 warm-up choice tasks (CT), 12 CT being used for utility estimation, 2 additional holdout CT

• Random task generation method: Sawtooth’s complete enumeration algorithm

• Sample: n= 60 (of 110) respondents

A conjoint study of coffee makers

Product attributes:1. Brand 4. Design2. Material 5. Price of a

cup3. System 6. Price

Meißner, Musalem, Huber9 Eye Tracking and Conjoint Choices

Fixations and saccades in a choice task:

Measuring attention

FixationsSaccades

Meißner, Musalem, Huber10 Eye Tracking and Conjoint Choices

Counting feature fixation frequencies

Meißner, Musalem, Huber11 Eye Tracking and Conjoint Choices

Technical data of our eye tracking equipment: • Head-mounted, video-based SMI Eye-Link II eye

tracking system • Head movement compensation• Temporal resolution: 250 Hz (every 4ms)• 9-step calibration, screen resolution: 1280 by1024

pixels

Eye tracking technology used in our first study

Meißner, Musalem, Huber12 Eye Tracking and Conjoint Choices

• Fixation frequency: number of times an element is fixated

• Fixation duration: increases closer to choice (e.g., Velichkovsky, Dornhoefer, Pannasch and Unema 2002)

• Pupil dilation: increases with greater cognitive load arousal (Beatty 1982; Just and Carpenter 1993)

• Saccade distance: may reflect perceptual shifts• Blinking: indicates processing or cognitive shifts

(Stern, Walrath and Goldstein 1984)

We focus on fixation frequencies to generate measures of attention to particular objects or features.

Eye tracking measures

Meißner, Musalem, Huber13 Eye Tracking and Conjoint Choices

Results

Meißner, Musalem, Huber14 Eye Tracking and Conjoint Choices

Feature utilities with standard deviations across respondents for the coffee maker study

Meißner, Musalem, Huber15 Eye Tracking and Conjoint Choices

Hit rate increases and decision time decreases with experience for the coffee maker study

Meißner, Musalem, Huber16 Eye Tracking and Conjoint Choices

Attention to an alternative increases with the attractiveness of the alternative, and this

relationship increases with practice

Meißner, Musalem, Huber17 Eye Tracking and Conjoint Choices

Attention to attributes increases for important attributes but is not significantly higher with

practice

Meißner, Musalem, Huber18 Eye Tracking and Conjoint Choices

Attention to features increases with greater utility for chosen alternatives but decreases for rejected

alternatives

Meißner, Musalem, Huber19 Eye Tracking and Conjoint Choices

Shift in fixations on left, right, and central alternatives as decision approaches

Meißner, Musalem, Huber20 Eye Tracking and Conjoint Choices

Differences between the three eye tracking conjoint studies

Meißner, Musalem, Huber21 Eye Tracking and Conjoint Choices

Multilevel analysis of factors influencing fixations in three studies

Meißner, Musalem, Huber22 Eye Tracking and Conjoint Choices

Incidental fixations have no/little effect on choice

Meißner, Musalem, Huber23 Eye Tracking and Conjoint Choices

1. Alternative focus directs attention to options with high utility

2. Attribute focus directs attention to important attributes

3. After we account for alternative attractiveness and attribute importance, feature utility has a minor impact on attention

Summary of findings

Meißner, Musalem, Huber24 Eye Tracking and Conjoint Choices

4. Incidental Fixations Have Little Effect on Choice: • The middle alternative gets more attention, but this

attention does not alter choice probabilities• The first fixated alternative gets more attention, but

this attention does not alter choice probabilities

Take away: “The attentional processes are consistent with a goal-driven causality mechanism whereby attention depends on stable values that each respondent brings to the task.”

Summary of findings

Meißner, Musalem, Huber25 Eye Tracking and Conjoint Choices

Appendix

Meißner, Musalem, Huber26 Eye Tracking and Conjoint Choices

Feature utilities with standard deviations across respondents for the beach vacation study

Meißner, Musalem, Huber27 Eye Tracking and Conjoint Choices

Feature utilities with standard deviations across respondents for the laptop study

Meißner, Musalem, Huber28 Eye Tracking and Conjoint Choices

Are short fixations the reason for the increased attention to the central object?

Meißner, Musalem, Huber29 Eye Tracking and Conjoint Choices

Are short fixations the reason for the increased attention to the central object? - Simulation