Ace Linear Regression Practical Exam

Locate key statistics (coefficients, p-values, R-squared, F-statistic) in R's `summary()` and `anova()` outputs.

Interpret the intercept (β₀) and slope (β₁) coefficients for a Simple Linear Regression (SLR) model.

Write a precise, 2-line exam-style sentence to explain the meaning of the full SLR equation in context.

Define a residual as the difference between an observed and a predicted value (e = y − ŷ).

Interpret R-squared and the Residual Standard Error (RSE) to assess the fit of a single regression model.

Explain the purpose of Adjusted R-squared and use it to compare multiple regression models with different numbers of predictors.

Justify model preference by comparing the Adjusted R-squared and RSE of competing models.

Articulate the common exam trap of adding variables to a model solely to increase R-squared.

Deconstruct the ANOVA table, defining Sum of Squares Total (SST), Regression (SSR), and Error (SSE).

Calculate missing values (e.g., SSE, MSR, F-statistic) from a partial ANOVA table using their defining formulas.

State the null and alternative hypotheses for the overall F-test in a multiple regression context.

Interpret the F-statistic and its associated p-value to determine the overall significance of a regression model.

Assess the linearity assumption by identifying patterns (e.g., curvature) in a residual vs. fitted values plot.

Diagnose heteroscedasticity (non-constant variance) by identifying a funnel shape in a residual vs. fitted values plot.

Evaluate the normality of residuals by interpreting a Normal Q-Q plot.

Explain the independence assumption and how its violation (autocorrelation) can be detected.

Interpret an individual slope coefficient (βᵢ) in a Multiple Linear Regression (MLR) model, using the phrase 'holding all other variables constant'.

Assess the statistical significance of individual predictors in an MLR model using their t-statistics and p-values.

Interpret the coefficient of a dummy (binary) variable as the change in the predicted response for the specified category relative to the base category.

Write a precise sentence explaining the estimated difference in the response variable between the two categories of a dummy variable.

Interpret the coefficients in a model with an interaction term, explaining how the effect of one predictor depends on the level of another.

Calculate the slope of one predictor for different fixed values of another predictor in an interaction model.

Define multicollinearity and use the Variance Inflation Factor (VIF) to detect its presence and severity.

Describe the consequences of high multicollinearity on coefficient estimates and their standard errors.

Distinguish between a high-leverage point (unusual predictor value) and a high-influence point (a point that significantly changes the model).

Visually identify potential high-leverage or influential points on standard diagnostic plots.

Formulate precise, 2-3 line exam-style answers for interpreting coefficients, model fit, and assumption checks.

Identify and avoid common exam traps, such as confusing correlation with causation or misinterpreting dummy variables.

Complete a mock practical exam paper by interpreting outputs and deriving missing values under time pressure.