Do We Judge Products on Too Little
Daniel Kahneman's acronym WYSIATI says judgment treats what is in view as the whole case. A mock-jury study, two consumer experiments and a dinnerware pricing study test pieces of that, and they also show what shrinks the effect.
Answer to the question
Do people judge products on too little information?
Often, in experiments. People asked to judge products from a partial description rated the attributes it mentioned as more important, and gave more extreme evaluations, than people who also saw a comparison product. Mock jurors given one side of a case predicted one-sidedly, and rating the unseen side first cut that gap from 7.1 jurors to 2.7.
Key takeaways
- Brenner and colleagues (1996): students who read only the plaintiff's side of a mock case guessed 12.3 of 20 jurors for the plaintiff, those who read only the defendant's side guessed 9.5, and a group that saw both guessed 11.3.
- In the same paper, rating the unseen side's strength before predicting cut the average gap between one-sided readers from 7.1 jurors to 2.7.
- Sanbonmatsu and colleagues (2003): 171 students rated attributes mentioned in four comments about a bicycle as more important than unmentioned ones, 2.25 against 1.23, except participants with high bicycle knowledge.
- In their third experiment, 98 students gave three mentioned car attributes 80.8 per cent of importance from a pruned list and 34.7 per cent from a full list.
- Hsee (1998): shown one dinnerware set, people priced a 24-piece set at 32.69 dollars and a 40-piece set with broken items at 23.25. Shown both, they priced the larger set higher. A 2023 replication of 400 people reproduced the first result and not significantly the second.
The acronym is WYSIATI, short for "what you see is all there is". Daniel Kahneman coined it in chapter 7 of Thinking, Fast and Slow for the way judgment treats the information at hand as the whole case.
A shopper meets a product through a page, a shelf or a pitch, and each of those is a selection. People judge products from the selection in front of them, and the open question is whether they register what it leaves out. Three sets of experiments bear on that, one of them replicated.
A Label for a Family of Findings
In Kahneman's account, the mind builds the most coherent story it can from whatever information happens to be available, and confidence follows the coherence of that story rather than the quantity or quality of the evidence behind it. A thin set of facts is easier to make coherent than a rich one, so less information can produce more confidence. Missing evidence does not register as missing.
He treats WYSIATI as an umbrella over findings that were studied separately, and names overconfidence, framing effects and base-rate neglect as consequences. That makes it a label for a family of results rather than one measured effect, so any test of it is a test of a piece. The three below are pieces about judgment from partial evidence: in a mock courtroom, in two consumer experiments and in a pricing study. The same gap-filling shows up when one visible trait stands in for the unseen ones, which is the territory of [How Far the Halo Effect Spreads](https://thewhyweproject.com/articles/how-far-the-halo-effect-spreads).
Mock Juries Saw Half the Case
Kahneman's chapter describes a legal demonstration, and the matching paper is more detailed than the summary. Students at San Jose State University read six adapted disputes, covering civil, criminal and child custody cases. Lyle Brenner and colleagues, Amos Tversky among them, published the results in the Journal of Behavioral Decision Making. Some saw only the plaintiff's side, some only the defendant's, and a jury group saw both. Each person predicted how many of 20 jurors would favour the plaintiff and gave a range they were 90 per cent sure the answer fell within.
The averages followed what people had read. Those who saw the plaintiff's side alone guessed 12.3 jurors for the plaintiff, those who saw the defendant's side alone guessed 9.5, and the jury group guessed 11.3. The one-sided readers also gave narrower ranges, 10.4 against 11.1, and were less accurate than the jury group. The authors concluded that people do not compensate enough for missing information.
A later experiment in the same paper tested a remedy. Participants who first rated how strong the unseen side's case was, and then predicted, cut the average gap between plaintiff-side and defendant-side readers from 7.1 jurors to 2.7. They rated the unseen side as only slightly weaker than the one they had read, 2.8 on a five-point scale where 3 meant equal. The shortfall lay in never consulting the other side, since participants did not think it weak.
Four Comments on a Bicycle
Four comments about a touring bicycle called "Brand H", from a reputable firm at an average price, were all that 171 undergraduates had to go on in the first of three experiments by David Sanbonmatsu and colleagues, reported in the Journal of Consumer Psychology in 2003. Each comment concerned one attribute, such as comfort or frame strength, and was favourable, middling or unfavourable. Participants then listed the attributes they thought most important when they judge products like bicycles. Attributes that had appeared in the comments were rated as more important than attributes that had not, 2.25 against 1.23 on average. The pattern did not appear in participants with high bicycle knowledge, though that group had only 18 people.
The second experiment used a compact car. Eighty-three students read a radio transcript for "Brand A", priced under 12,000 dollars and credited with three good attributes. Some read it alone, others alongside a Brand B described on six different good attributes. Alone, Brand A drew the more extreme evaluation, 1.29 against 0.66, and the weight given to the attributes it mentioned statistically accounted for the difference. Confidence did not differ significantly between the two conditions.
In the third, 98 business students allocated 100 per cent of importance across attributes of a car described on three. Given a full list of nine attributes plus an "all other" category, they gave the three mentioned attributes 34.7 per cent. Given a pruned list of only those three plus "all other", they gave them 80.8 per cent. The authors report that three things reduced the overweighting: knowing the product category, thinking about one's own criteria before evaluating, and having a comparison product described on other attributes.
A product page is a selection by construction, because it shows the attributes the seller chose to show. Participants in these experiments did what a shopper does with a short list: they gave the listed attributes more weight than a full list produced. A shopper who wants to judge products evenly can borrow the mock-jury remedy and ask what the unlisted attributes would say. A single review is the visible fact in a related case, covered in [Why Does One Bad Review Hurt So Much](https://thewhyweproject.com/articles/why-does-one-bad-review-hurt-so-much).
Why Did the Smaller Dinner Set Win?
A set of 24 intact dinner pieces was priced above a set of 40 with some broken, but only when nobody could compare them. Christopher Hsee ran the study and reported it in the Journal of Behavioral Decision Making in 1998. One set had 24 intact pieces. The other, the higher-value set, had 40 pieces with some of the cups and saucers broken. People who saw only one set, and so had to judge products with no comparison, put a higher price on the 24-piece set. People who saw both put a higher price on the 40-piece set.
A preregistered close replication, published in Collabra: Psychology in 2023, used 400 participants. It reproduced the separate-evaluation result and found the joint-evaluation reversal pointing the same way as before without reaching statistical significance. Across the three studies it ran, it supported "less is better" in all three and gave weaker support to "more is better".
| Prices put on the sets | 24 intact pieces | 40 pieces, some broken |
|---|---|---|
| Hsee (1998), shown one set | $32.69 | $23.25 |
| Hsee (1998), shown both | $29.70 | $32.03 |
| Replication, shown one set | $33.70 | $23.96 |
| Replication, shown both | $30.35 | $31.67 |
In the replication's joint condition, 59 per cent of participants would pay more for the 40-piece set and 41 per cent more for the 24-piece set. The first number seen setting the scale of a judgment is a neighbouring case, covered in [Why the First Price You See Decides What You Pay](https://thewhyweproject.com/articles/why-the-first-price-you-see-decides-what-you-pay).
Our reading is that the label is useful and the tests are narrower than the label. When shoppers judge products from a page or a shelf, the experiments suggest the description gets read as the product, and that a comparison, or a rating of the missing side, is what loosens that. Each experiment shows partial evidence moving an evaluation, and in each the movement shrank once the missing side was in view.
Common questions
Do people judge products on too little information?
Often, in experiments. People asked to judge products from a partial description rated the attributes it mentioned as more important, and gave more extreme evaluations, than people who also saw a comparison product. Mock jurors given one side of a case predicted one-sidedly, and rating the unseen side first cut that gap from 7.1 jurors to 2.7.
What does WYSIATI mean?
It stands for "what you see is all there is", Daniel Kahneman's term in Thinking, Fast and Slow for judgment that treats the information at hand as the whole case. Confidence follows how coherent the available story is, so a thin set of facts can feel as certain as a rich one. In consumer terms, a shopper may judge products from a handful of listed attributes without noticing the unlisted ones.
Does seeing both options change the price people pay?
It did in Hsee's dinnerware study. Shown one set, people priced the 24-piece set at 32.69 dollars and the 40-piece set at 23.25. Shown both, they priced the larger set higher, 32.03 against 29.70. A 2023 replication reproduced the first pattern and found the second in the same direction but not significant.
Does knowing more about a product reduce the effect?
In the bicycle experiment, participants with high bicycle knowledge did not rate the mentioned attributes as more important, though that group had only 18 people. Across the three experiments the authors also found less overweighting when people considered their own criteria first or saw a comparison product. These are ways to judge products on more than the description.
Sources
Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux, chapter 7. Used for the definition and the account of confidence following coherence.
Brenner, L. A., Koehler, D. J., & Tversky, A. (1996). 'On the evaluation of one-sided evidence.' Journal of Behavioral Decision Making, 9(1), 59-70. Full text at bear.warrington.ufl.edu/brenner/papers/brenner-jbdm96.pdf
Sanbonmatsu, D. M., Kardes, F. R., Houghton, D. C., Ho, E. A., & Posavac, S. S. (2003). 'Overestimating the importance of the given information in multiattribute consumer judgment.' Journal of Consumer Psychology, 13(3), 289-300.
Hsee, C. K. (1998). 'Less is better: When low-value options are valued more highly than high-value options.' Journal of Behavioral Decision Making, 11(2), 107-121.
Vonasch, A. J., et al. (2023). 'Less is better in separate evaluations versus more is better in joint evaluations: Mostly successful close replication and extension of Hsee (1998).' Collabra: Psychology, 9(1). doi:10.1525/collabra.77859