Hall & Johansson (2008) — Using choice blindness to study decision making and introspection
What is choice blindness, and how does it relate to (or differ from) change blindness?
Choice blindness is the phenomenon in which a person fails to notice a discrepancy between the choice they made and a manipulated outcome (e.g. they are presented the alternative), yet they often rationalize the outcome as if it were their own. It relates to change blindness (failure to detect visual changes) insofar as both reflect limits of conscious access, but it differs by operating in the domain of decision and introspection rather than pure perception.
What is the “self-transforming survey” that Hall et al. (2012) developed? Explain the mechanisms and purpose of that trick.
The “self-transforming survey” is a paper survey designed so that a respondent’s earlier answers can be covertly replaced (flipped to the opposite) mid-survey (via sleight-of-hand in the paper). The trick ensures the participant doesn’t suspect the reversal and then asks them to explain or defend their (now reversed) position, testing whether they detect the change and whether they endorse it.
Summarize Maier’s (1931) “solution of a problem and its appearance in consciousness” article. What is the basic problem (e.g. two cords), and what phenomena does Maier emphasize (suddenness, change in representation, lack of conscious intermediate steps)?
Maier’s (1931) paper Reasoning in Humans. II. The Solution of a Problem and its Appearance in Consciousness investigates how people solve insight problems (notably the two-string problem). In that paradigm, two cords hang from the ceiling out of reach, and subjects must tie them together, but cannot grasp both simultaneously. The solution typically involves tying a weight to one string, swinging it, then catching it while holding the other. Maier emphasizes phenomena such as the suddenness of insight (solutions appear abruptly), the absence of conscious intermediate stages, and the idea that representational relationships change (i.e. the meaning of parts is restructured).
Describe the canonical experimental paradigm in Hall & Johansson (2008) to demonstrate choice blindness. How do they covertly manipulate the choice, and how do they question participants afterward?
In the canonical paradigm, participants choose between two stimuli (e.g. which of two faces is more attractive). Immediately afterward, the experimenter covertly swaps the chosen item with the unchosen item (via sleight-of-hand), so the participant is presented with what they did not choose. The participant is then asked to explain why they prefer the presented item. Because subjects often do not detect the swap, they provide reasons for a choice they never made.
Describe the experimental procedure: how they collected participants’ moral judgments, effected the reversal, and probed their responses and detection.
Participants first fill out moral/dilemma items (e.g. attitudes on principle statements). Then in a reversal condition, one or more of their earlier answers are covertly changed to the opposite. The participant is presented with the “reversed” answer and asked to comment, explain, or justify it. After completing the survey, participants are asked whether they noticed any changes. Detection rates and qualitative explanations are recorded.
What manipulations or hints did Maier provide in his experiments (e.g. bumping a string, giving pliers) to facilitate insight solutions? What did those manipulations show about the process of insight?
Maier provided hints, such as bumping one string to make it swing (a cue to use it as a pendulum) or providing pliers (to suggest weight). These manipulations facilitated insight: some participants solved the problem immediately following the hint, even though they reported not consciously noticing the hint’s effect. The results show that cues can trigger restructuring without conscious awareness, supporting the view that insight involves latent reorganization rather than stepwise trial and error.
What are the main empirical findings (detection rates, confabulated reasons) of that work? What do these results imply about introspection and self-knowledge?
Empirically, many manipulated trials go undetected (often a majority), and participants often produce plausible, detailed justifications for the “choice” they are shown (which was not their actual choice). This suggests that introspection is limited, that verbal reports do not necessarily reflect the real causes of choice, and that people readily construct post hoc rationalizations to explain outcomes they did not originate. It raises doubts about the trustworthiness of self-reports of reasons for decisions.
What were the principal results (e.g. detection rates, proportion endorsing reversed positions)? What do these results suggest about stability, accessibility, and construction of moral attitudes?
Results: many participants (a majority) failed to detect at least one reversal (e.g. 69% on some analyses). Many participants accepted and defended the reversed positions without recognizing they had changed their earlier stance. These results suggest that moral attitudes, when expressed on surveys, may not be deeply held or stable—they can be constructed or rationalized on the spot, and individuals may have poor introspective access into their own moral attitudes.
What predictions or accounts from Maier about the development of solution in consciousness does he offer (e.g. nucleus vs whole, experience before solution)? What is his view of whether the restructuring is conscious or unconscious?
Maier speculated whether solutions emerge gradually from a “nucleus” or emerge as a whole. He leans toward the latter: solutions appear as sudden gestalt restructurings. Just before the solution, subjects often experience tension or impasse; the restructuring itself occurs unconsciously, then “breaks into” awareness as a fully formed solution. In his view, the reorganization is not accompanied by conscious transitional steps; rather consciousness only catches the final reorganized percept.
What alternative accounts (e.g. memory lapse, inattention, social compliance, perceptual confusion) might explain the choice blindness phenomenon, and how do the authors respond to or preempt those critiques?
Possible alternative accounts include memory failure (they forgot their original choice), inattention or distraction, social compliance (they go along with the experimenter), or perceptual confusion (they misperceived the item). The authors respond by showing that detection is not purely explained by memory lapse (control trials, detection checks), that participants give specific (not generic) justifications (arguing against mere compliance), and that manipulations survive attention controls. They argue that the phenomenon is not trivial to explain by those alternatives and that it points to fundamental introspective gaps.
What interpretative or methodological concerns do Hall et al. raise about survey research and polling in light of their findings? What recommendations do they (explicitly or implicitly) offer to survey designers?
Hall et al. warn that responses in polling and surveys may reflect momentary constructions rather than stable attitudes; that expressed moral opinions may be more malleable than assumed; and that survey instruments may overestimate people’s introspective certainty. They recommend that survey designers incorporate consistency checks, multiple measures, and prompts for justification (so as to reduce confabulation), and to treat isolated survey responses cautiously as measures of stable belief.
What theoretical implications does choice blindness have for models of decision making, introspection, and the status of verbal reports of reasons?
Choice blindness implies that decision processes are often opaque to introspection: our reasons are constructed after the fact rather than directly accessed. Models of decision making should treat verbal reasons skeptically; introspective reports are not necessarily faithful to underlying processes. It suggests that part of our “self-knowledge” is narrative construction rather than direct perception, and that preference and reasons may be more fluid and retroactively rationalized than we ordinarily assume.
Identify at least one moderator (individual difference or situational factor) that affected whether participants noticed reversals, and explain how it influenced detection.
One moderator is attitude strength / issue involvement: participants with stronger prior convictions are more likely to notice reversals. Another is cognitive load or time pressure / complexity of the item: when items are subtle or the participant is under time stress, detection declines. Also, eliciting immediate reasons at initial response also reduces later acceptance of a reversal.