Which method is used to predict values of one variable from another?

Which method is used to predict values of one variable from another?

• A. Regression
• B. Correlation
• C. ANOVA
• D. t-test

Answer: (A)

Regression is used to predict values of one variable from another because it builds a model that expresses the dependent variable as a function of the independent variable, yielding a usable predictive equation. In simple linear regression, you fit a line Y = α + βX. The slope β tells you how much Y is expected to change for each unit increase in X, and you can plug an X value into this equation to obtain a predicted Y. The method finds α and β by minimizing the squared differences between observed Y and predicted Y, giving the best prediction under the linear assumption. Correlations measure how strongly and in what direction two variables move together, but they don’t provide a formula to predict Y from X. ANOVA tests whether there are differences in means across groups or whether a factor explains variance, rather than providing a predictive rule for a continuous outcome. The t-test compares means between two groups or tests a parameter, not a general method for predicting one variable from another.

Regression is used to predict values of one variable from another because it builds a model that expresses the dependent variable as a function of the independent variable, yielding a usable predictive equation. In simple linear regression, you fit a line Y = α + βX. The slope β tells you how much Y is expected to change for each unit increase in X, and you can plug an X value into this equation to obtain a predicted Y. The method finds α and β by minimizing the squared differences between observed Y and predicted Y, giving the best prediction under the linear assumption.

Correlations measure how strongly and in what direction two variables move together, but they don’t provide a formula to predict Y from X. ANOVA tests whether there are differences in means across groups or whether a factor explains variance, rather than providing a predictive rule for a continuous outcome. The t-test compares means between two groups or tests a parameter, not a general method for predicting one variable from another.