Curve estimation
Curve estimation fits several simple one-predictor models at once (linear, logarithmic, inverse, quadratic, cubic, compound, power, S, growth, exponential) and draws the fitted curves over the data, so you can see which shape describes the relationship best.
Menu: Analyze → Regression → Curve Estimation…
When should I use Curve estimation?
- One scale outcome and one scale predictor, and you suspect the relationship is not a straight line.
What are the assumptions?
- Compound, power, S, growth and exponential models are fitted on ln(outcome): the outcome must be positive. Logarithmic and power models need a positive predictor.
- R² of log-scale models refers to ln(outcome), so compare it with care.
How do I run it in StatLab?
- Choose the dependent and the independent variable.
- Tick the models to compare (Linear and Quadratic are ticked by default).
How do I read the output?
| Output | Meaning |
|---|---|
| Model Summary and Parameter Estimates | R², F, df and Sig. for each model, then the constant and b1–b3. |
| Fitted curves | The data with every fitted curve, one colour per model. |
Worked example
Model score by age in the bundled sample dataset (File → Sample dataset: 150 survey respondents with gender, programme group, age, pre/post test scores, five satisfaction items, income, smoking and education): linear R² = .14, quadratic R² = .14.
How would I report it in APA style?
A linear model of wellbeing on age gave R² = .14, F(1, 148) = 23.84, p < .001.
Numbers above were computed by StatLab’s validated engine from the sample dataset when this page was built.
What are common mistakes?
- Choosing the model with the highest R² without looking at the plot: cubic curves can chase noise.
- Comparing R² between a log-scale model and a model on the original scale as if they were the same measure.