This article describe the **t-test effect size**. The most commonly used measure of effect size for a t-test is the **Cohen’s d** (Cohen 1998).

The `d`

statistic redefines the difference in means as the number of standard deviations that separates those means. The formula looks like this (Navarro 2015):

\[

d = \frac{\mbox{(mean 1)} - \mbox{(mean 2)}}{\mbox{std dev}}

\]

In this article, you will learn:

**Cohen’s d formula**to calculate the effect size for one-sample t-test, for independent t-test (with pooled standard deviation or not) and for paired samples t-test (also known as repeated measures t-test).**Effect size interpretation**describing the critical value corresponding to**small, medium and large**effect sizes.**Calculation of the Cohen’s d in R**

Contents:

#### Related Book

Practical Statistics in R II - Comparing Groups: Numerical Variables## Prerequisites

Load the required R package for computing the Cohen’s d:

`library(rstatix)`

Demo dataset:

`head(ToothGrowth, 3)`

```
## len supp dose
## 1 4.2 VC 0.5
## 2 11.5 VC 0.5
## 3 7.3 VC 0.5
```

## Effect size interpretation

T-test conventional effect sizes, poposed by Cohen, are: 0.2 (small efect), 0.5 (moderate effect) and 0.8 (large effect) (Cohen 1998, Navarro (2015)). This means that if two groups’ means don’t differ by 0.2 standard deviations or more, the difference is trivial, even if it is statistically significant.

d-value | rough interpretation |
---|---|

0.2 | Small effect |

0.5 | Moderate effect |

0.8 | Large effect |

## Cohen’s d for one-sample t-test

To calculate an effect size, called `Cohen's d`

, for the one-sample t-test you need to divide the mean difference by the standard deviation of the difference, as shown below. Note that, here: `sd(x-mu) = sd(x)`

.

**Cohen’s d formula**:

\[

d = \frac{m-\mu}{s}

\]

- \(m\) is the sample mean
- \(s\) is the sample standard deviation with \(n-1\) degrees of freedom
- \(\mu\) is the theoretical mean against which the mean of our sample is compared (default value is mu = 0).

**Calculation**:

`ToothGrowth %>% cohens_d(len ~ 1, mu = 0)`

```
## # A tibble: 1 x 6
## .y. group1 group2 effsize n magnitude
## * <chr> <chr> <chr> <dbl> <int> <ord>
## 1 len 1 null model 2.46 60 large
```

## Cohen’s d for independent t-test

The independent samples t-test comes in two different forms:

- the standard
*Student’s t-test*, which assumes that the variance of the two groups are equal. - the
*Welch’s t-test*, which is less restrictive compared to the original Student’s test. This is the test where you do not assume that the variance is the same in the two groups, which results in the fractional degrees of freedom.

### Cohen’s d for Student t-test

There are multiple version of Cohen’s d for Student t-test. The most commonly used version of the Student t-test effect size, comparing two groups (\(A\) and \(B\)), is calculated by dividing the mean difference between the groups by the pooled standard deviation.

**Cohen’s d formula**:

\[

d = \frac{m_A - m_B}{SD_{pooled}}

\]

where,

- \(m_A\) and \(m_B\) represent the mean value of the group A and B, respectively.
- \(n_A\) and \(n_B\) represent the sizes of the group A and B, respectively.
- \(SD_{pooled}\) is an estimator of the pooled standard deviation of the two groups. It can be calculated as follow :

\[

SD_{pooled} = \sqrt{\frac{\sum{(x-m_A)^2}+\sum{(x-m_B)^2}}{n_A+n_B-2}}

\]

**Calculation**. If the option `var.equal = TRUE`

, then the pooled SD is used when compting the Cohen’s d.

`ToothGrowth %>% cohens_d(len ~ supp, var.equal = TRUE)`

```
## # A tibble: 1 x 7
## .y. group1 group2 effsize n1 n2 magnitude
## * <chr> <chr> <chr> <dbl> <int> <int> <ord>
## 1 len OJ VC 0.495 30 30 small
```

Note that, for small sample size (< 50), the Cohen’s d tends to over-inflate results. There exists a **Hedge’s Corrected version of the Cohen’s d** (Hedges and Olkin 1985), which reduces effect sizes for small samples by a few percentage points. The correction is introduced by multiplying the usual value of d by `(N-3)/(N-2.25)`

(for unpaired t-test) and by `(n1-2)/(n1-1.25)`

for paired t-test; where N is the total size of the two groups being compared `(N = n1 + n2)`

.

```
ToothGrowth %>% cohens_d(
len ~ supp, var.equal = TRUE,
hedges.correction = TRUE
)
```

```
## # A tibble: 1 x 7
## .y. group1 group2 effsize n1 n2 magnitude
## * <chr> <chr> <chr> <dbl> <int> <int> <ord>
## 1 len OJ VC 0.488 30 30 small
```

### Cohen’s d for Welch test

The Welch test is a variant of t-test used when the equality of variance can’t be assumed. The effect size can be computed by dividing the mean difference between the groups by the “averaged” standard deviation.

**Cohen’s d formula**:

\[

d = \frac{m_A - m_B}{\sqrt{(Var_1 + Var_2)/2}}

\]

where,

- \(m_A\) and \(m_B\) represent the mean value of the group A and B, respectively.
- \(Var_1\) and \(Var_2\) are the variance of the two groups.

**Calculation**:

`ToothGrowth %>% cohens_d(len ~ supp, var.equal = FALSE)`

```
## # A tibble: 1 x 7
## .y. group1 group2 effsize n1 n2 magnitude
## * <chr> <chr> <chr> <dbl> <int> <int> <ord>
## 1 len OJ VC 0.495 30 30 small
```

## Cohen’s d for paired samples t-test

The effect size for a paired-samples t-test can be calculated by dividing the mean difference by the standard deviation of the difference, as shown below.

**Cohen’s d formula**:

\[

d = \frac{mean_D}{SD_D}

\]

Where `D`

is the differences of the paired samples values.

**Calculation**:

`ToothGrowth %>% cohens_d(len ~ supp, paired = TRUE)`

```
## # A tibble: 1 x 7
## .y. group1 group2 effsize n1 n2 magnitude
## * <chr> <chr> <chr> <dbl> <int> <int> <ord>
## 1 len OJ VC 0.603 30 30 moderate
```

## Summary

This article shows how to compute and interpret the t-test effect using the Cohen’s d statistic. We describe the formula of the Cohen’s d for one-sample, two-samples and paired samples t-test. Examples of R codes are provided for the calculations.

## References

Cohen, J. 1998. *Statistical Power Analysis for the Behavioral Sciences*. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates.

Hedges, Larry, and Ingram Olkin. 1985. “Statistical Methods in Meta-Analysis.” In *Stat Med*. Vol. 20. doi:10.2307/1164953.

## Recommended for you

This section contains best data science and self-development resources to help you on your path.

### Coursera - Online Courses and Specialization

#### Data science

- Course: Machine Learning: Master the Fundamentals by Stanford
- Specialization: Data Science by Johns Hopkins University
- Specialization: Python for Everybody by University of Michigan
- Courses: Build Skills for a Top Job in any Industry by Coursera
- Specialization: Master Machine Learning Fundamentals by University of Washington
- Specialization: Statistics with R by Duke University
- Specialization: Software Development in R by Johns Hopkins University
- Specialization: Genomic Data Science by Johns Hopkins University

#### Popular Courses Launched in 2020

- Google IT Automation with Python by Google
- AI for Medicine by deeplearning.ai
- Epidemiology in Public Health Practice by Johns Hopkins University
- AWS Fundamentals by Amazon Web Services

#### Trending Courses

- The Science of Well-Being by Yale University
- Google IT Support Professional by Google
- Python for Everybody by University of Michigan
- IBM Data Science Professional Certificate by IBM
- Business Foundations by University of Pennsylvania
- Introduction to Psychology by Yale University
- Excel Skills for Business by Macquarie University
- Psychological First Aid by Johns Hopkins University
- Graphic Design by Cal Arts

### Amazon FBA

#### Amazing Selling Machine

### Books - Data Science

#### Our Books

- Practical Guide to Cluster Analysis in R by A. Kassambara (Datanovia)
- Practical Guide To Principal Component Methods in R by A. Kassambara (Datanovia)
- Machine Learning Essentials: Practical Guide in R by A. Kassambara (Datanovia)
- R Graphics Essentials for Great Data Visualization by A. Kassambara (Datanovia)
- GGPlot2 Essentials for Great Data Visualization in R by A. Kassambara (Datanovia)
- Network Analysis and Visualization in R by A. Kassambara (Datanovia)
- Practical Statistics in R for Comparing Groups: Numerical Variables by A. Kassambara (Datanovia)
- Inter-Rater Reliability Essentials: Practical Guide in R by A. Kassambara (Datanovia)

#### Others

- R for Data Science: Import, Tidy, Transform, Visualize, and Model Data by Hadley Wickham & Garrett Grolemund
- Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems by Aurelien Géron
- Practical Statistics for Data Scientists: 50 Essential Concepts by Peter Bruce & Andrew Bruce
- Hands-On Programming with R: Write Your Own Functions And Simulations by Garrett Grolemund & Hadley Wickham
- An Introduction to Statistical Learning: with Applications in R by Gareth James et al.
- Deep Learning with R by François Chollet & J.J. Allaire
- Deep Learning with Python by François Chollet

Version: Français

Dividing the mean difference by the standard deviation of the differences does not conform with the definition of Cohen’s d. With this computation, if each person in the sample would increase or decrease with the exact same amount (however small or large), then this would yield an effect size of (minus) infinity. The proposed ‘effect size’ says something about the variability of the effect, but nothing about the strength of the effect. To obtain an actual effect size, one should divide the mean difference by the pooled pre- or post-test standard deviation.