MKTG3001 Chap.10 Regression Analysis and Prediction
Regression Analysis and Prediction
This chapter covers the advanced level of survey analysis, where the goal shifts from describing data or spotting relationships to predicting an outcome. Regression checks whether one or more independent (predictor) variables can forecast a dependent variable.
You learn to set up the regression null and alternative hypotheses and apply the p < 0.05 decision rule, to read the unstandardised coefficient B as the direction and size of predicted change per one-unit rise in a predictor, and to read R² as the proportion of variation in the outcome that the model accounts for, on a scale from 0 to 1. The lectures teach a fixed interpretation sequence: p-value first, then B, then R².
The chapter then contrasts simple regression (a single predictor) with multiple regression (two or more predictors of one Likert-type outcome) and shows how basic, intermediate and advanced analyses combine in a group research project.
What this chapter covers
- 01
Regression as prediction rather than association
- 02
Dependent variable versus independent (predictor) variable
- 03
Regression hypotheses and the p < 0.05 decision rule
- 04
Unstandardised coefficient B: direction and amount of change
- 05
R² and explanatory power from 0 to 1
- 06
Simple versus multiple regression
- 07
Combining basic, intermediate and advanced analyses in a project
Worked example · free
Interpreting a fitted line: app ease predicting renewal intention
- 2Significance first: H₀ says app ease does not significantly predict renewal intention. Since p = 0.004 falls below 0.05, H₀ is rejected in favour of H₁, so app ease is a statistically significant predictor.
- 2Coefficient next: B = 0.37 is positive, so every one-point increase in app ease is predicted to raise renewal intention by 0.37 units. For a rating of 3 the prediction is 1.6 + 0.37 × 3 = 1.6 + 1.11 = 2.71.
- 1Explanatory power last: R² = 0.33 tells us app ease accounts for 33% of the variation in renewal intention, leaving 67% to other influences such as price or dock availability.
Key terms
- Predictor variable
- An independent variable whose values are used in a regression model to forecast the dependent variable.
- Unstandardised coefficient B
- A regression statistic giving the direction and size of predicted change in the outcome for each one-unit increase in a predictor.
- Coefficient of determination
- The statistic R², a value between 0 and 1 showing the proportion of variation in the dependent variable explained by the regression model.
Regression Analysis and Prediction FAQ
What is the difference between correlation and regression in marketing research?
Correlation measures whether two variables move together and sits with the intermediate techniques. Regression is treated as an advanced technique because it designates one variable as the outcome and estimates how far that outcome is expected to shift when a predictor changes, which lets a manager forecast results such as purchase or loyalty scores.
How do I know whether to use simple or multiple regression?
Count the independent variables in the research question. One predictor of one outcome calls for simple regression; two or more predictors of a single outcome call for multiple regression. In both cases the outcome stays a single variable, and the course examples measure every variable on Likert-type scales, reporting p-value, B and R².
Is a low R² a sign that a regression result is useless?
Not necessarily. A significant predictor with a modest R², say 0.20, still forecasts the outcome reliably; it simply means the model accounts for a fifth of the variation and other drivers explain the rest. Report the figure honestly and suggest further predictors that might raise explanatory power in a multiple model.
Exam move
Build a one-page sheet that lists the three outputs in the taught order, p-value, B and R², with a single sentence template for each, and rehearse it until you can write it from memory, since the exam is closed book and handwritten. Then invent two quick scenarios of your own: one with a single predictor and one with three.
For each, decide aloud whether simple or multiple regression fits and justify it using the number of independent variables, the single dependent variable and the Likert-type measurement. Practise turning a printout into a 200 to 250 word paragraph that reaches a managerial recommendation, and check every direction word against the sign of B.
Finally, revise how regression connects to the basic and intermediate techniques, because an essay may ask you to choose among all three levels.
Working through Regression Analysis and Prediction in MKTG3001? Sia is AskSia’s AI Marketing tutor — ask any MKTG3001 Regression Analysis and Prediction question and get a clear, step-by-step explanation grounded in how MKTG3001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.