Mental reference points provide a comparison basis when people lack an exact value. Individuals may judge a quantity relative to information already available in memory or to the surrounding context, then adjust the response through comparison or rounding. Because the selected reference point can differ across situations, the same quantity may receive different estimates under different conditions.
Anchoring shifts an estimate by making an initial value or comparison point psychologically influential, while framing changes how the same judgment problem is presented. Both effects can alter the information people emphasize when forming an answer. In psychology experiments, these manipulations help researchers examine how context influences reasoning rather than treating estimates as independent of presentation.
Rounding simplifies information by replacing a precise quantity with an easier approximate value. Comparison allows a person to judge an unfamiliar amount against a known or remembered reference, whereas magnitude scaling supports judgments about relative size. These strategies can make estimation practical, but they also reveal how people mentally represent quantity when precision is unavailable.
When information is incomplete, people draw on what they remember and combine those memories with available contextual cues. The resulting estimate reflects both the evidence at hand and the way quantity has been represented mentally. This makes numerical estimation useful for studying judgment under uncertainty, including how limited information can produce systematic shifts rather than purely random responses.
A researcher presents a quantity, likelihood, or related judgment for which an exact answer is not directly available, then records the participant's approximate response. The task can incorporate comparison, rounding, magnitude scaling, memory, or contextual information. Researchers can then examine how changes in framing or anchoring affect the estimates and use those patterns to evaluate models of reasoning.
These tasks provide evidence about numerical cognition, perception, decision-making, and cognitive bias. The estimates can show how people represent quantity and how judgment changes when contextual information or reference points shift. In everyday and experimental settings, the resulting patterns help connect observable responses with broader models of human reasoning and decision-making.