Separate detecting a signal from deciding it is there
Use hits and false alarms to distinguish response policy from improved sensory discrimination.
A question to investigate
If a listener reports more tones, must their hearing have improved?
Calculate two response rates and explain why more detections alone do not establish better sensitivity.
Mechanism and evidence
Sensation concerns detecting and converting physical stimulation; perception concerns organizing and interpreting sensory information. Attention, expectations and context affect reported experience. An absolute threshold is often operationalized as detection on half the trials under specified conditions. In signal detection, observed responses depend on both information quality and a decision criterion. A more liberal criterion can increase both hits and false alarms. The linked text introduces these distinctions.
In these invented trials, a yes response has two possible meanings. On a tone trial it is a hit; on a noise-only trial it is a false alarm. A no on a tone trial is a miss, while a no on a noise-only trial is a correct rejection. Label all four cells before calculating percentages. Total yes responses combine two denominators and therefore obscure the mechanism.
Condition A produces 24 hits, 16 misses, 4 false alarms and 36 correct rejections. Condition B produces 34 hits, 6 misses, 14 false alarms and 26 correct rejections. A’s hit rate is 24/40 = 60%; its false-alarm rate is 4/40 = 10%. B’s corresponding rates are 85% and 35%. Both rise by 25 percentage points. This pattern is compatible with a criterion change; it does not by itself quantify unchanged sensitivity.
Make a second prediction for a genuinely stronger signal with unchanged instructions. One might expect better separation of signal and noise, but the actual response distribution must be measured. To estimate a sensitivity statistic, one must state its model and assumptions; subtracting the two rates is not the standard normal-distribution sensitivity measure. This lesson deliberately stops at the observable rates.
For transfer, design a comparison that includes noise-only trials and records the instructions. Do not test classmates with loud sounds or flashing stimuli. The supplied records allow the reasoning without any sensory exposure. A diagram of the response categories plus your written interpretation is the required artifact.
Inspect the relationships
| Reality and response | Says yes | Says no |
|---|
| Signal present | Hit | Miss |
| Signal absent | False alarm | Correct rejection |
Your table: Replace these labels with A’s counts, then B’s counts. Keep the two 40-trial denominators visible.
Worked example
Condition A: hit rate 60%, false-alarm rate 10%. Condition B: 85% and 35%. Reporting more signals also generated more false alarms; an improved hearing claim is not established.
Independent application
A new detector has 18 hits in 20 signal trials and 9 false alarms in 10 noise trials. Which report is accurate? Explain the denominator, process or rule before checking the response. Record a revision after feedback.
Reading
OpenStax Psychology 2e, sensation versus perception. The numerical records and tasks are original fictional teaching examples; they are not findings reported by this source.
