Course contentsShow
← Mathematics
Lesson 3069 of 3,12160. Statistics — ModelingFree lesson

Inference for the Slope

Conduct hypothesis tests and construct confidence intervals for the true slope parameter.

Inference for the Slope

You'll learn how to test hypotheses and build confidence intervals for the true slope in linear regression, moving beyond just estimating it.

The Core Idea

When you fit a simple linear regression line to sample data, the slope you calculate is just an estimate of the true population slope. Like any estimate from a sample, it has uncertainty. Inference for the slope lets you answer two critical questions:

  1. Hypothesis testing: Is there really a linear relationship, or could the observed slope just be due to random chance? (Testing if the true slope β₁ equals zero)
  2. Confidence intervals: What's a plausible range for the true slope value?

How It Works

The sample slope follows a t-distribution (under the regression assumptions). The test statistic is:

t = (b₁ - β₀) / SE(b₁)

where:

  • b₁ is your calculated sample slope
  • β₀ is the hypothesized slope (often 0)
  • SE(b₁) is the standard error of the slope

For a confidence interval, you use:

b₁ ± t × SE(b₁)*

where t* is the critical value from the t-distribution with (n - 2) degrees of freedom.

Why This Matters

Testing whether the slope is zero tells you if your predictor variable actually matters. A 95% confidence interval for the slope gives you a range of plausible effect sizes—not just "yes, there's a relationship" but "each unit increase in X is associated with between ___ and ___ change in Y."

Key Takeaway: Inference for the slope uses t-tests and confidence intervals to quantify uncertainty about the true linear relationship, helping you determine whether an observed association is statistically significant and estimate its magnitude.