Not really sure why the first one is viewed as "less is better" at all. The design doesn't rule out the idea that someone would prefer a $150 coat to a $45 scar

Not really sure why the first one is viewed as "less is better" at all. The design doesn't rule out the idea that someone would prefer a $150 coat to a $45 scarf. To me all it demonstrates is that generosity is at least partially anchored in the quality of a gift relative to the category it was selected in -- I'd rather a really nice scarf than a really crummy coat, even if coats cost more than scarfs. I can experiment this way on myself: if someone got me a gorgeous, $25 pocket square, I'd derive much more utility than if they got me a hideous $300 Men's Warehouse cotton-poly suit. And you could even read this a third way: some objects you want one of (I don't need a second microwave, for instance, even if it's better than the one I own) and some objects you want many of (I would be happy with any bottles of Scotch!) The second study also lends itself to more rational / less "behavioural" explanations. First, it's possible there's diminishing marginal returns between 24 and 31 pieces to the point that f(31) ~= f(24) and f(31) - disposal_cost < f(24). Second, perhaps utility for a dinner set depends on its quality as in the first study. I have a relatively diffuse prior about the quality of the dinner sets, but by learning that Option A has 9 broken pieces, I update my belief so that I now think Option A has 31 crummy pieces and Option B has 24 decent pieces. Maybe the brokenness of the 9 pieces is orthogonal to the quality of the various sets, but in general this would seem to be a good heuristic and a smart use of updating. To take either of these things as a conclusion about expectation setting or less-is-more seems silly to me. More is more, but only in utility terms, not in raw input dollars.