PsychologyPerceptionA-Level

AQAIBAbiturBachilleratoCambridgeCCEA

Signal Detection Theory (d')

Quantifies the ability to distinguish signal from background noise.

Understand the formulaSee the free derivationOpen the full walkthrough

d^{'} = Z (H R) - Z (F A R)

Open Full Walkthrough Try Calculator

This public page keeps the free explanation visible and leaves premium worked solving, advanced walkthroughs, and saved study tools inside the app.

Core idea

Overview

Signal Detection Theory (SDT) provides a mathematical framework for quantifying the ability to discern between information-bearing patterns and random noise. The sensitivity index, dPrime, represents the distance between the means of the noise and signal distributions in units of standard deviation, allowing for an assessment of sensory capability independent of an observer's decision criteria.

When to use: Use this equation when assessing perceptual performance in tasks where participants must distinguish a stimulus from background noise. It is ideal for experiments where response bias, such as a tendency to say 'yes' regardless of the stimulus, might otherwise skew raw accuracy scores.

Why it matters: It allows researchers to separate an observer's actual sensory capability from their psychological decision-making strategy. This is crucial in high-stakes fields like medical diagnostic imaging, air traffic control monitoring, and forensic eyewitness identification.

Symbols

Variables

d' = Sensitivity (d'), HR = Hit Rate, FAR = False Alarm Rate

d^{'}

Sensitivity (d')

V a r iab l e

H R

Hit Rate

V a r iab l e

F A R

False Alarm Rate

V a r iab l e

Walkthrough

Derivation

Definition: Signal Detection Theory (d')

Defines sensitivity d' as the difference between z-scores of hit and false alarm rates.

Assumes underlying signal and noise distributions with equal variance (classic SDT).
Hit rate and false alarm rate are treated as probabilities (0–1) and may be bounded away from 0 and 1 in practice.

Compute sensitivity from z-scores:

Z(·) converts a probability to a standard-normal z-score; larger d' indicates better discriminability.

d^{'} = Z (H R) - Z (F A R)

Result

d^{'} = Z (H R) - Z (F A R)

Source: A-Level Psychology — Research Methods / Perception

Free formulas

Rearrangements

Solve for $d^{'}$

Make dPrime the subject

d^{'} = Z (H R) - Z (F A R)

dPrime is already the subject of the formula.

Difficulty: 1/5

Solve for $H R$

Make HR the subject

H R = Φ (d^{'} + Z (F A R))

Rearrange the Signal Detection Theory (d') formula to solve for HR (Hit Rate).

Difficulty: 2/5

Solve for $F A R$

Make FAR the subject

F A R = Φ (Z (H R) - d^{'})

Rearrange the Signal Detection Theory equation for sensitivity (d') to make the False Alarm Rate (FAR) the subject.

Difficulty: 2/5

The static page shows the finished rearrangements. The app keeps the full worked algebra walkthrough.

Visual intuition

Graph

The graph of Hit Rate (HR) against signal strength or sensitivity (d') follows a sigmoid (S-shaped) curve, beginning with a shallow slope at low sensitivity levels, transitioning to a steep increase, and leveling off as it approaches an asymptote near 1.0. This shape represents the psychophysical reality that changes in signal detection are most pronounced at moderate levels of sensitivity, while floor and ceiling effects limit performance at extreme values. It illustrates the probabilistic nature of signal detection where the likelihood of a 'hit' is constrained by the overlap of noise and signal distributions.

Graph type: sigmoid

Why it behaves this way

Intuition

A statistical picture showing two overlapping normal distribution curves, one representing responses to noise alone and the other representing responses to signal plus noise.

A quantitative measure of an observer's sensitivity to a signal, independent of their response bias.

How well someone can truly differentiate between a real event and just background noise, without being influenced by their tendency to say 'yes' or 'no'.

The probability of correctly detecting a signal when it is present.

The rate at which the observer correctly identifies a target.

FAR

The probability of incorrectly identifying a signal when only noise is present.

The rate at which the observer mistakenly identifies a target when there is none.

Z(x)

The inverse of the standard normal cumulative distribution function, converting a probability (like HR or FAR) into a Z-score.

It transforms a percentage (e.g., 75% hits) into a point on a standard bell curve, allowing comparison of how 'far apart' the signal and noise distributions are.

Signs and relationships

- Z(FAR): The subtraction of Z(FAR) means that as the false alarm rate increases, the calculated sensitivity d' decreases. This reflects that if an observer frequently makes false alarms, their ability to truly distinguish signal

Free study cues

Insight

Canonical usage

The Signal Detection Theory sensitivity index, d', is a dimensionless measure derived from dimensionless probabilities (Hit Rate and False Alarm Rate) transformed into z-scores.

Common confusion

A common confusion is to treat Hit Rates or False Alarm Rates as raw counts rather than proportions, or to expect d' to have a physical unit. All components of the d' calculation are dimensionless statistical measures.

Dimension note

d' is a dimensionless index that quantifies sensitivity by measuring the distance between the means of the signal and noise distributions in standard deviation units.

Unit systems

$H R$ none · Hit Rate is a proportion (0 to 1) representing the probability of correctly detecting a signal.

$F A R$ none · False Alarm Rate is a proportion (0 to 1) representing the probability of incorrectly reporting a signal when only noise was present.

$Z (H R)$ none · The Z-transform of the Hit Rate, representing the z-score corresponding to the cumulative probability HR.

$Z (F A R)$ none · The Z-transform of the False Alarm Rate, representing the z-score corresponding to the cumulative probability FAR.

$d^{'}$ none · The sensitivity index, d', is a dimensionless measure of an observer's ability to discriminate between signal and noise.

Ballpark figures

Quantity:

One free problem

Practice Problem

A radiologist correctly identifies a tumor (Hit Rate) 84.1% of the time and incorrectly identifies a tumor in healthy tissue (False Alarm Rate) 15.9% of the time. Calculate the sensitivity index d'.

Hit Rate0.841

False Alarm Rate0.159

Solve for: $d P r im e$

Hint: Convert the percentages to decimals and find the corresponding Z-scores; note that Z(0.841) is approximately 1.

The full worked solution stays in the interactive walkthrough.

Where it shows up

Real-World Context

A doctor's ability to correctly identify a disease on an X-ray (Hit) versus misidentifying a healthy scan (False Alarm).

Study smarter

Tips

Always convert Hit and False Alarm rates to decimal proportions between 0 and 1 before finding Z-scores.
A dPrime value of 0 indicates the observer cannot distinguish signal from noise at all.
Adjust extreme rates of 0 or 1 slightly (e.g., using 1 divided by 2N) to avoid mathematically infinite Z-scores.

Avoid these traps

Common Mistakes

Assuming a high hit rate always means high sensitivity (ignoring the false alarm rate).

Common questions

Frequently Asked Questions

Defines sensitivity d' as the difference between z-scores of hit and false alarm rates.

Use this equation when assessing perceptual performance in tasks where participants must distinguish a stimulus from background noise. It is ideal for experiments where response bias, such as a tendency to say 'yes' regardless of the stimulus, might otherwise skew raw accuracy scores.

It allows researchers to separate an observer's actual sensory capability from their psychological decision-making strategy. This is crucial in high-stakes fields like medical diagnostic imaging, air traffic control monitoring, and forensic eyewitness identification.

Assuming a high hit rate always means high sensitivity (ignoring the false alarm rate).