2.9 - Mistakes in Decision-making
The dual processing model of thinking
People often make decisions that seem illogical or unpredictable, and this can be explained through a model that describes two distinct ways of thinking. This dual processing model, proposed by Kahneman (2011), distinguishes between fast, intuitive thought processes and slower, more deliberate ones.
System 1 thinking
System 1 thinking is the fast, automatic mode of processing information. It relies on intuition and quick judgments, often leading to efficient but sometimes error-prone decisions. This occurs because System 1 draws on unconscious patterns and stored experiences without much effort.
System 2 thinking
In contrast, System 2 thinking is slow and deliberate, involving conscious effort and logical analysis. It uses all available information to make rational choices, which can override the initial impulses from System 1. This leads to more accurate decisions but requires time and mental energy.
Comparing System 1 and System 2 thinking
| System | Description | When it is used | Examples |
|---|---|---|---|
| System 1 | Fast, intuitive, automatic, unconscious, and effortless; often based on 'best guesses' | In automatic situations, especially when time is limited or information is overwhelming | Using non-compensatory strategies (deciding based on one key factor without weighing all options); relying on heuristics (mental shortcuts); following external anchors (influences like market trends); buying stocks when prices are rising and selling when falling |
| System 2 | Slow, deliberate, logical, conscious, rational, controlled, effortful, and statistical | To question or override System 1 impulses; involves fact-checking initial responses | Applying utility theory (weighing costs and benefits); using compensatory strategies (balancing pros and cons); relying on internal anchors (personal standards); buying stocks when falling and selling when rising |
Interaction between System 1 and System 2
System 1 and System 2 do not operate in isolation; they interact to shape decisions. For instance, System 2 might build on quick intuitions from System 1 by analysing them further. Conversely, System 1 can use knowledge accumulated through past System 2 processing to make expert-level snap judgments, such as those made by experienced chess players. This interaction explains why people sometimes make fast but informed choices.
Application to decision-making mistakes
System 1 thinking often leads to errors when people use heuristics like representativeness, where they ignore current evidence and over-rely on past trends or stored knowledge. This can result in dysfunctional behaviours, such as poor consumer choices. Shifting to more System 2 thinking, by deliberately questioning intuitions, can help overcome these mistakes and promote more rational decisions.
Choice blindness in decision-making
Choice blindness is a phenomenon where people fail to notice when their expressed preferences are manipulated or switched, revealing limitations in self-awareness during decision-making. This occurs because attention is not always focused on the details of choices, leading to unconscious acceptance of inconsistencies. The concept was explored in a key study by Hall et al. (2010), which extended previous research from visual to sensory domains like taste and smell.
Hall et al. (2010)
The aim was to investigate the effect of similarity, liking, and incentives on choice blindness in a naturalistic setting. Hypotheses predicted that choice blindness would be less common when shoppers sampled teas/jams with dissimilar tastes/smells, when they had a preference for one product over another, or when offered a free sample of their preferred option.
Method
- This was a field experiment using an independent measures design, conducted in a naturalistic supermarket setting in Sweden.
- Data were collected through structured interviews to assess participants' awareness of product switches.
- The sample consisted of 180 shoppers, with 66% female and ages ranging from 16 to 80 (mean age 40).
- Participants were selected using opportunity sampling, approached while shopping and asked to join a quality control test.
Procedure
- Participants tasted two jams and smelt two teas, rating their preferences on a ten-point scale.
- They were then asked to re-sample their preferred product, but the experimenter secretly switched it to their least preferred using a double-compartment 'magic' jar.
- Afterwards, participants discussed their preferences, rated discrimination difficulty on a ten-point scale, and indicated confidence in receiving the correct product.
- The order of jam and tea sampling was counterbalanced, and the jar was flipped consistently for switches.
- Controls included neutral ambient conditions, moderate noise and crowd levels, and standardised jar handling.
Variables
- Independent variables - Whether participants were told they would receive a free gift or not; whether products were similar versus dissimilar; whether the experimenter secretly switched the products or not.
- Dependent variables - Expression of concern (none, concurrent, retrospective, or sensory change detection).
Ethics
- Deception was used, so informed consent was not obtained initially.
- The study received approval from the Regional Swedish Ethics Board.
- Safeguards included debriefing with the right to withdraw, provision for written consent post-debrief, and ensured confidentiality.
Results
- 68% of participants showed choice blindness, failing to express any concerns.
- Detection was higher for dissimilar pairings than similar ones.
- Switches were more often detected when one product was disliked.
- Unexpectedly, in the tea condition, choice blindness was higher (80% vs. 54%) when a free gift was offered.
Conclusions
- Choice blindness occurs in real-world settings involving smell and taste, as well as vision.
- It is not affected by the consequences of decisions or the difficulty of differentiating between similar products.
Evaluation
- Strengths - High ecological validity from the supermarket setting; controlled variables like counterbalancing reduced order effects.
- Limitations - Potential social desirability bias in interviews, where participants might not voice concerns to appear cooperative; variability in supermarket noise could affect replicability.
Interference effects on advertising recall
Interference in memory occurs when new or old information disrupts the recall of other details, leading to mistakes in decision-making, such as forgetting product details from advertisements. Retroactive interference happens when later information hinders recall of earlier material, while proactive interference involves earlier information affecting later recall. This was investigated by Burke and Srull (1988) in the context of competitive advertising.
Burke and Srull (1988)
The aim was to investigate the retroactive and proactive effects of competitive advertising.
Method
- This laboratory experiment involved showing university students 12 text-only magazine advertisements.
- Participants rated either purchase likelihood or general interest.
- After a two-minute delay, they recalled products from either earlier (retroactive) or later (proactive) in the series.
- Advert variations included different brands/products, same products/different brands, or same brand/different models.
- A double-blind design minimised researcher bias.
Procedure
- Participants viewed the 12 advertisements.
- They rated each for either purchase intention or interest.
- After a delay, they recalled three products from specified positions in the sequence.
- Recall was tested under varying advert similarity conditions.
Results
- Recall doubled when rating for purchase intention compared to interest.
- Best recall occurred with varied products/brands; worst with similar ones.
- Retroactive interference was stronger for early adverts, not moderated by rating type; proactive effects were also noted but less emphasised.
Conclusions
- Retroactive interference impairs advert recall, but focusing on purchase intentions improves it.
- Proactive interference similarly reduces recall, highlighting how competing information leads to decision-making errors in consumer contexts.
Evaluation
- Strengths - Strong control through randomisation and double-blind procedures establishes cause and effect.
- Limitations - Low ecological validity as adverts were text-only, unlike real multi-media ones; potential demand characteristics if participants guessed the study's purpose.
Methodological issues in the research
Evaluating research methods helps understand the reliability and validity of findings on decision-making mistakes. Both Hall et al. (2010) and Burke and Srull (1988) used experimental designs, but each has strengths and weaknesses.
Strengths and weaknesses of the studies
| Strengths | Weaknesses |
|---|---|
| Experimental control: Both studies controlled confounding variables, allowing cause-and-effect conclusions. Burke and Srull used double-blind procedures to reduce bias; Hall et al. randomised order to counter order effects. | Validity issues: Burke and Srull's text-only adverts lack realism compared to image-based real ads. In Hall et al., interpersonal influences like talking to other shoppers could alter behaviour, and social desirability might suppress concerns. |
| Data collection: Structured interviews in Hall et al. provided detailed self-reports, though prone to bias. | Reliability concerns: Hall et al.'s supermarket setting had variable noise and crowding, reducing standardisation and replicability. |
Issues and debates in decision-making mistakes
Psychological theories often debate the influences on behaviour, and decision-making errors highlight key contrasts.
Individual versus situational explanations
Choice blindness in Hall et al. (2010) varied, with only half the sample affected by dissimilar products, suggesting individual factors like personality play a role. However, situational elements, such as product similarity or free gift incentives, increased blindness, showing environmental influences on decisions.
Determinism versus free will
The dual processing model leans deterministic, as unconscious System 1 thinking automatically shapes choices, reducing free will's role. Hall et al. (2010) showed behaviour can be determined by unnoticed factors, diverting attention from conscious control. Yet, engaging System 2 allows more deliberate, free choices by overriding automatic processes.