Effect Size (Difference) Equation, Derivation & Rearrangements

Core idea

Overview

The raw effect size represents the absolute difference between the means of two distinct groups, such as an experimental and a control group. Unlike standardized metrics, this calculation preserves the original units of measurement, allowing for direct interpretation of the magnitude of change.

When to use: Apply this formula when comparing two groups on a continuous dependent variable where the measurement scale has intrinsic, well-understood meaning. It is ideal for metrics like reaction time in milliseconds, blood pressure in mmHg, or scores on a validated psychometric scale.

Why it matters: This calculation shifts the focus from simple statistical significance (p-values) to practical significance. It helps researchers and practitioners determine if the size of an intervention's impact is large enough to justify its implementation in real-world settings.

Symbols

Variables

D = Difference, M_E = Exp. Mean, M_C = Ctrl. Mean

D

Difference

V a r iab l e

M_{E}

Exp. Mean

V a r iab l e

M_{C}

Ctrl. Mean

V a r iab l e

Walkthrough

Derivation

Formula: Effect Size (Difference)

Basic calculation of the raw difference between two group means.

Means have been pre-calculated.

1

Subtract means:

Simple subtraction of the control mean from the experimental mean shows the raw effect impact.

D = M_{E} - M_{C}

Result

D = M_{E} - M_{C}

Source: GCSE Psychology — Research Methods / Statistics

Free formulas

Rearrangements

Solve for $D$

Make D the subject

D = M_{E} - M_{C}

The difference (D) is calculated by subtracting the control mean ( $M_{C}$ ) from the experimental mean ( $M_{E}$ ).

Difficulty: 1/5

Solve for $M_{E}$

Make $M_{E}$ the subject

M 1 = D + M_{C}

To find the experimental mean ( $M_{E}$ ), add the difference (D) to the control mean ( $M_{C}$ ).

Difficulty: 2/5

Solve for $M_{C}$

Make $M_{C}$ the subject

M 2 = M_{E} - D

To find the control mean ( $M_{C}$ ), subtract the difference (D) from the experimental mean ( $M_{E}$ ).

Difficulty: 2/5

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

Visual intuition

Graph

Graph unavailable for this formula.

The graph is a horizontal line because the difference between the two means remains constant regardless of the independent variable. This results in a constant value for D, represented by a flat line parallel to the X-axis.

Graph type: constant

Why it behaves this way

Intuition

Imagine two distinct distributions of scores on a number line, representing the experimental and control groups; the effect size D is the direct horizontal distance between the centers (means) of these two distributions.

D

The raw, unstandardized difference between the average outcomes of two groups.

A larger absolute value of D indicates a greater practical difference between the groups' average results.

M_{E}

The arithmetic mean (average) score or value for the experimental group.

Represents the typical outcome observed when the intervention or treatment is applied.

M_{C}

The arithmetic mean (average) score or value for the control group.

Represents the typical outcome observed under standard conditions or without the intervention, serving as a baseline.

Signs and relationships

- M_C: The subtraction indicates a comparison of the experimental group's mean relative to the control group's mean. A positive D means $M_{E}$ is greater than $M_{C}$ , while a negative D means $M_{C}$ is greater than $M_{E}$ , thereby

Free study cues

Insight

Canonical usage

The effect size D will always carry the same units as the original measurements from which the means $M_{E}$ and $M_{C}$ were derived.

Common confusion

A common mistake is to confuse this raw difference effect size with standardized effect sizes (e.g., Cohen's d), which are dimensionless because they involve division by a standard deviation, thus canceling units.

Dimension note

If the dependent variable being measured is a score, count, or other quantity without inherent physical units (e.g., a rating on a Likert scale or a number of correct answers), then the effect size D will also be

Unit systems

$M_{E}, M_{C}, D$ Varies (e.g., milliseconds, mmHg, arbitrary score units) · All variables in this equation (experimental mean, control mean, and their difference) must share the same units and dimension, which are determined by the dependent variable being measured.

One free problem

Practice Problem

A clinical psychologist evaluates a new anxiety treatment. The experimental group (M1) has a mean post-treatment score of 22, while the control group (M2) has a mean score of 35. Calculate the raw effect size (D).

Exp. Mean22

Ctrl. Mean35

Solve for: $D$

Hint: Subtract the control group mean from the experimental group mean.

The full worked solution stays in the interactive walkthrough.

Where it shows up

Real-World Context

Group A (Sleep) score 8, Group B (No Sleep) score 5. Difference = 3.

Study smarter

Tips

Ensure both groups are measured using the exact same units before subtracting.
Interpret the difference in the context of the scale's total range.
Report raw differences alongside standard deviations to provide a sense of data spread.

Avoid these traps

Common Mistakes

Assuming a large difference is always statistically significant.

Keep going

Related Formulas

Common questions

Frequently Asked Questions

Basic calculation of the raw difference between two group means.

Apply this formula when comparing two groups on a continuous dependent variable where the measurement scale has intrinsic, well-understood meaning. It is ideal for metrics like reaction time in milliseconds, blood pressure in mmHg, or scores on a validated psychometric scale.

This calculation shifts the focus from simple statistical significance (p-values) to practical significance. It helps researchers and practitioners determine if the size of an intervention's impact is large enough to justify its implementation in real-world settings.

Assuming a large difference is always statistically significant.