Calculator Form
Example Data Table
This sample shows a binary predictor and a continuous score. A higher group 1 mean will usually produce a positive point biserial relationship.
| Observation | Binary Group | Score |
|---|---|---|
| 1 | 1 | 88 |
| 2 | 1 | 84 |
| 3 | 1 | 79 |
| 4 | 1 | 76 |
| 5 | 1 | 91 |
| 6 | 0 | 71 |
| 7 | 0 | 69 |
| 8 | 0 | 62 |
| 9 | 0 | 58 |
| 10 | 0 | 66 |
Formula Used
M1 is the mean score for observations coded 1.
M0 is the mean score for observations coded 0.
sy is the overall standard deviation of the continuous scores.
p is the proportion of cases in group 1, and q = 1 - p is the proportion in group 0.
n is the total sample size used for significance testing and confidence intervals.
How to Use This Calculator
- Select Summary statistics if you already know the group means, total standard deviation, group 1 percentage, and sample size.
- Select Raw paired data if you have line-by-line observations coded as 0 or 1 with a numeric score.
- Enter the values carefully. In raw mode, each line must begin with 0 or 1 followed by the score.
- Click Calculate Correlation to show the result section below the header and above the form.
- Review the correlation, variance explained, confidence interval, p value, and group means.
- Use Download CSV or Download PDF to save your results for reports, coursework, or audits.
Frequently Asked Questions
1) What does point biserial correlation measure?
It measures the relationship between one binary variable and one continuous variable. It shows whether the two binary groups differ meaningfully on the numeric outcome and how strongly they differ.
2) When should I use this calculator?
Use it when your predictor has only two valid codes, usually 0 and 1, and your outcome is numeric. Common examples include pass versus fail, treatment versus control, and selected versus not selected.
3) What is considered a strong result?
Interpretation depends on context, but absolute values near 0.10 are small, near 0.30 are moderate, and near 0.50 or above are often considered strong. Always judge effect size with domain knowledge.
4) Can I use percentages for the binary variable?
In summary mode, yes. Enter the percentage belonging to group 1. In raw mode, use exact codes of 0 and 1 for each observation so the calculator can separate the two groups correctly.
5) Why does the calculator need overall standard deviation?
The formula scales the mean difference by the spread of all continuous scores. Without the overall standard deviation, the size of the group difference cannot be converted into the correlation coefficient.
6) What does a negative point biserial value mean?
A negative value means the group coded as 1 has a lower average score than the group coded as 0. The sign reflects direction, while the absolute size reflects strength.
7) Does a significant p value guarantee importance?
No. Statistical significance only suggests the observed relationship is unlikely under a null model. Practical importance depends on the effect size, confidence interval width, and the real decision context.
8) Can I export the results for documentation?
Yes. The page includes CSV and PDF download options after calculation. These exports are useful for reporting, teaching materials, audits, quality checks, and research summaries.