Eye Tracking and Decision Making Flashcards
Vision
Visual system is… centrally implicated in learning, higher-order, cognitive affective processes, decision making and its behavioural implementation and co-ordination
Building on First Impressions - Does the halo effect always apply?
Fang, Singh & Ahuluwalia (2007)
232 participants from University of Kanvas
Procedure: Banner ads on the web page being read changed regularly
Participants saw target ads 0, 5, or 10 times
Participants rated ads after reading the web page
The more you like something initially, the more halo effect builds.
The mere exposure effect
Zajonc (1968) – the automatic increase of appeal through exposure to stimuli
Mere exposure effects in advertising
Lindgaard’s (2006) experiments – appeal decisions made reliably in 50 ms
Supporting evidence from research on perception – Oliva’s work
Limitations
Can create halo effects
To build on halo effects you need to like a website/advertisement first.
Just & Carpenter (1980)
Proposed the eye-mind hypothesis.
No appreciable lag between what is fixated and what is processed.
Pieters, Wedel & Rosbergen (1997, 2004)
- 73 seconds – the average time that consumers examine adverts [range of
- 37-5.30 seconds]
0-0.5 seconds - First fixations are usually on the pictorials, on the region that is more informative
0.5-1 second – large cluster of fixations on key text/editorial or subsidiary picture (all participants still looking at this stage)
1-4 seconds – exploring the editorial or pictures in more detail (50% of participants still looking)
5-20 seconds – sequences of fixations indicate reading.
Cowen, Ball, & Delin (2002)
Looked at measures used for eye movement measurement.
Number of fixations
Total fixations
Average fixation duration
Fixations spatial density - Global measure of total amount of processing performed on each page.
Scan paths
These are looked at most ^
Different measures online
Lindgaard et al (2006)
Rated liking for 50 web pages seen for 50 millisec
Rated again a 2nd time to check reliability
Rating were reliably the same on 2nd occasion
Mcdougall, Goodliffe, Ollis & Taylor
Do user evaluations of informativeness take longer?
Avg no. of fixations 500ms = 1.15
Reasons given for evaluations are the same even when only 1-2 fixations are possible
This may be the result of retro-fitting ‘reasons for decisions’ to task demands
BUT they are not based on changing patterns of eye fixations
Mere exposure effects - scene gist capture
Olivia et al (2004,2005,2006)
Found that the gist of a scene can captured in a single glance/fixation
This includes both ‘bottom-up’ information – colour, surface, volume – and ‘high-level’ information – objects, semantic knowledge activation
Thus, gist is captured ‘at a glance’ at both perceptual & conceptual levels
Decision making as a problem-solving process
Problem Recognition
Information Search
Evaluation of Alternatives
Pros and Cons
Heuristics
Product Choice
Outcomes
Multi Attribute models
They are extremely popular with market researchers.
Attributes – those which consumers usually take into consideration when evaluating a particular product
Beliefs – the extent to which a consumer thinks a brand has a particular attribute
Importance weights – the importance of each attribute for a consumer (this may vary considerably between individuals)
Multi Attribute models
Information makes it possible to
This information makes it possible to:-
Spot weaknesses in the brand profile
Capitalise on advantages which the product already has
Strengthen key product-attribute linkages
Create new unique selling points if required
Multi-attribute models – mismatches between predictions and behaviour
Fishbein (1983) – created the first, and most influential, attribute model
Other models have been created to improve the predictability of the model
These models include:-
The theory of reasoned action (Ajzen & Fishbein, 1977)
The theory of planned behaviour (Ajzen, 1991)
Multi-attribute models – mismatches between predictions and behaviour
Mismatches between predictions based on multi-attribute models and questionnaires arise because:-
There can be differences between intentions and actual behaviour
Social pressure may change our intentions or behaviour
Models concentrate on the evaluation of the product rather than evaluating what consumers feel about buying it (e.g. if the shop is far away or unpleasant)
Models also need to include what consumers feel about websites when shopping online
Attitudes can change for a variety of reasons (e.g. culture, shopping on-line, age)
Evaluation of Alternatives
The product alternatives a person considers compromise their evoked set.
Members of the evoked set usually share some characteristics; we categorize them similarly.
The way a person mentally groups or categorises products influences which alternatives he or she will consider.
Usually we associate some brands more strongly with these categories.
Evaluation of Alternatives
Evoked set
Consideration set
Inept set
Inert set
Evoked set: the alternatives a consumer knows about
Consideration set: the ones actually considered
Inept set: ones a consumer knows about but would not consider buying
Inert set: those not under consideration at all
Evaluation of Alternatives
Strategic implications of product grouping
Product positioning
This hinges on the marketer’s ability to convince consumers that a product should be considered within a given category.
Identifying competitors
Many different product forms compete for membership of a category.
Exemplar producers
Where a product is a really good example of a category, it is more familiar to consumers and is more easily recognized and recalled.
Locating products
Product categorization can also affect consumers’ expectations regarding the places where they can locate products.
Perspectives on decision making
Rational Perspective
Behavioural Influence Perspective
Experimental Perspective
A rational perspective
Careful and logical integration of information about a product
Weigh up the pluses and minuses of each alternative
Arriving at a satisfactory decision (?)
Highly involved
Behavioural influence perspective
Decisions are a learned response to environmental cues, e.g. buying a ‘special offer’ on impulse in a shop; buying chocolate at the checkout.
Decisions influenced by cues such as bright colours, effective packaging, easy of visual search
Low involvement
Experiential perspective
Selection made when highly involved but not easily explained rationally
Elaboration, Likelihood Model
Petty, Cacippo & Goldman (1981)
Students told that the University was instituting a comprehensive examination which must be passed to allow them to graduate.
High-involvement group – told this would happen before they graduated
Low-involvement group – told this would happen in 10 years’ time
Students told that the University was instituting a comprehensive examination which must be passed to allow them to graduate.
High-involvement group – told this would happen before they graduated
Low-involvement group – told this would happen in 10 years’ time
Varied: -
The quality of the arguments: strong vs weak
(affect attitudes of high-involvement group?)
Expertise of the ‘source’: from within the University vs outside the University
(affect attitudes of low-involvement group?)
Use of heuristics Gigerenzer & Gaissmaier (2011)
Heuristics rely on reducing effort by:
“A heuristic is a strategy that ignores part of the information, with the goal of making decisions more quickly, frugally, and … [sometimes] … more accurately than complex methods”
Shah & Oppenheimer (2008) Using fewer cues Simplifying the weighting of cues Integrating less information Examining fewer alternatives
Recognition heuristic Gigerenzer & Gaissmaier (2011)
If one of two alternatives is recognised and the other is not, then infer that the recognised alternative has the higher values with respect to the criterion
Recognition Heuristic
Coates et al. (2004)
Hoyer & Brown (1990)
Priming a familiar brand increases the probability it will be considered for purchase
In a blind test, most people preferred a high quality peanut butter in preference to 2 alternative, lower quality, jars.
When the brand label of the high quality peanut butter was changed to another jar, most now preferred the ‘branded’ jar. - Sensation Transference
Use of heuristics - processing fluency Alter & Oppenheimer (2009)
“Processing fluency, the ease with which people process information, reliably influences people’s judgements across a broad range of dimensions. … Because every cognition falls along a continuum from effortless to demanding it generates a corresponding fluency experience”
Argued that fluency experiences create a cue, or heuristic, for judgments and decision-making.
Heuristics in complex choices Hauser et al (2009)
Found that sequential heuristics predict consumer choices well.
Heuristic decision rules are more likely when:
there are more products
there are more features to be evaluated
quantifiable features are more salient
there is more time pressure
the consumer is in an early phase of the decision process
the effort required to make a decision is more salient
the above are combined forming a complex choice.
Choice Architecture Thaler & Sunstein (2008)
Many ways to present choice to a decision maker
What is chosen depends on how the choice is presented
Choice Architecture Johnson et al (2012)
Claim there is no neutral architecture (allowing free choices)
All choice presentations have a (usually implicit) default).
Tools for architecture fall into two categories.
Tools used in structuring the choice task (what to present decision making)
Tools used in describing the choice options (how to present it).
Structuring Choice
- Number of Alternatives
- Number of alternatives
‘Tyranny of choice’ ‘Choice overload’
Need to balance 2 criteria
more options create greater cognitive burden
More options increase the chances of offering a preference match
- Technology & Decision aids
Help us to identify attractive choice alternatives..and to filter out ones that are not of interest.
Manipulating the set of alternatives offered (creating a default)
Making some product attributes more salient, e.g. Booking.com
Filters can be helpful of shape choice, e.g. Airbnb (helpful?)
Structuring choice
- Technology & Decision aids
- Technology & Decision aids
Help us to identify attractive choice alternatives..and to filter out ones that are not of interest.
Manipulating the set of alternatives offered (creating a default)
Making some product attributes more salient, e.g. Booking.com
Filters can be helpful of shape choice, e.g. Airbnb (helpful?)
Structuring Choice
- Defaults
One of the most powerful and popular choices
Default settings – determine initial encounters with products
Re-use defaults - come into play with subsequent product uses
Choice option default – use of pre-checked boxes
Persistent defaults – where past choices are remembered
Reverting defaults – where past choices are ‘forgotten’/deleted
There are obvious ethical risks associated with defaults.
Structuring Choice.
- Choice over time
Early positive outcomes are preferred … so we yield to immediate temptations/choices and heavily discount later outcomes
Uncertainty about the future …. tend to focus on salient or desirable future outcomes (we see a rosy future)
Over optimism about the future … assume we will accomplish more than we actually do.
Tools that translate choice into immediate salient out are more successful.
Structuring choice
- Task Structure affects the search process
Simple choices – one from a small set of alternatives
Large choice sets – need to consider the role of search costs
Search costs can be reduced by:
Tools that translate choice into immediate salient out are more successful.
