Compromise effect
The middle option feels safest, so it wins.
What it means
The compromise effect is the tendency, when choosing among options arrayed on a quality–price spectrum, to disproportionately pick the intermediate one and avoid the extremes. Its mechanism is extremeness aversion coupled with the search for a defensible, low-regret choice: the middle option looks like a safe hedge that is neither recklessly cheap nor extravagantly expensive, and it is easy to justify to oneself and others. This reveals, like the decoy effect, that preferences are context-dependent — the same option's appeal rises simply because more extreme alternatives flank it. Sellers exploit this by adding a high-end 'premium' tier whose main job is not to sell but to convert the previously top option into a comfortable middle, lifting its share. A boundary condition is that the effect depends on the options being hard to evaluate in isolation and on the chooser's uncertainty; experts with clear preferences are less swayed. It matters in product-line design, pricing tiers, menus, and subscription plans, where the careful placement of extremes shapes which middle option people 'freely' choose.
Why the middle wins
Simonson and Tversky grounded the effect in loss aversion. Within a set, each option is weighed by its advantages and disadvantages relative to its neighbors, and because losses loom larger than equivalent gains, an option at either extreme carries one conspicuously large disadvantage — dearest on price, or weakest on quality. The middle option has smaller disadvantages on both sides, so its relative losses net out lower. A second, complementary account is reason-based choice: people gravitate to options they can defend with an explicit reason, and 'it is a sensible compromise' is a ready-made justification that settles an otherwise hard trade-off. Both accounts predict the same thing — the intermediate option wins precisely when the attributes pull against each other and no single option is clearly best.
What the evidence shows
Simonson first demonstrated the effect in 1989, and Simonson and Tversky generalized it in 1992. The strongest current summary is Neumann, Bockenholt and Sinha's 2016 meta-analysis of 142 experimental observations, which confirms that middle options are chosen significantly more often but finds the magnitude varies up to threefold with method: it is smaller under price-quality trade-offs and binary-to-trinary comparisons, and larger for utilitarian products, more attribute dimensions, and high-quality extensions. Crucially, three common ways of measuring share can yield contradictory conclusions about the same moderator. Field and incentive-compatible tests corroborate the lab work — Pinger and colleagues found compromise options gained roughly five percent of share on a real restaurant menu, and studies with real payments for durable goods reproduce the effect.
Where it breaks down
Several limits temper the effect. First, measurement fragility: because it is defined as a change in relative share when the set is altered, the metric chosen — absolute-share change, relative shift, or middle-option proportion — can enlarge, shrink, or even reverse the apparent result, so a lone headline number should be read with care. Second, the three-option layout is not magic; adding further options can push choosers onto other heuristics and weaken the pull toward the middle. Third, other-regarding or altruistic framings can blunt or reverse it. Fourth, it needs attributes that line up on shared, comparable dimensions; when options resist a common ranking, there is no obvious middle to prefer. Even where it holds, the field shift is measured in single-digit percentage points, not landslides.
Related but distinct
The compromise effect is often confused with the decoy, or attraction, effect. Both are context effects Simonson catalogued, but they operate differently. The decoy effect needs an added option that is asymmetrically dominated — clearly worse than one target on every attribute — which lifts that specific target by making it look like a bargain. The compromise effect needs no dominance at all: the added option simply has to be more extreme, repositioning a former extreme as the new middle. So a decoy boosts one named option, while a compromise reshapes the geometry of the whole set. Extremeness aversion is the underlying tendency; the compromise effect is one of its manifestations, not a synonym for it, and the same design can trigger both at once.
Examples
Faced with small, medium, and large coffees, most buy medium — and adding an extra-large lifts medium's share.
Few diners order the cheapest bottle on the wine list or the dearest; add one very expensive bottle at the bottom and the second-dearest starts to look like sensible middle ground.
Broadband sold at 100, 500, and 900 Mbps sends most households to 500 — not because they know what they need, but because the middle is the easiest choice to defend to yourself.
A charity's donation page shows $25, $100, and $250 buttons; the $100 middle draws the most gifts, and adding a $1,000 tier quietly turns $250 into the new comfortable compromise.
During open enrollment, employees choosing among bronze, silver, and gold health plans cluster on silver, which reads as adequate cover without overpaying, whether or not it matches their actual usage.
First described in Simonson (1989).
Key references
- Neumann, N., Bockenholt, U., & Sinha, A. (2016). A meta-analysis of extremeness aversion. Journal of Consumer Psychology, 26(2), 193-212. doi.org/10.1016/j.jcps.2015.05.005
- Pinger, P., Ruhmer-Krell, I., & Schumacher, H. (2016). The compromise effect in action: Lessons from a restaurant's menu. Journal of Economic Behavior & Organization, 128, 14-34. doi.org/10.1016/j.jebo.2016.04.017
- Chang, C.-C., Chuang, S.-C., Cheng, Y.-H., & Huang, T.-Y. (2012). The compromise effect in choosing for others. Journal of Behavioral Decision Making, 25, 109-122. doi.org/10.1002/bdm.720
- Simonson, I., & Tversky, A. (1992). Choice in context: Tradeoff contrast and extremeness aversion. Journal of Marketing Research, 29(3), 281-295. doi.org/10.1177/002224379202900301
- Simonson, I. (1989). Choice based on reasons: The case of attraction and compromise effects. Journal of Consumer Research, 16(2), 158-174. doi.org/10.1086/209205