BUSI 614 Week 7 SPSS Quantitative Assignment 6 Example

What Drives Purchase Intention? A Multiple Regression in SPSS for a Snack Brand

Research Question and Variables

A composite snack company is preparing to relaunch its protein bar and wants to know which consumer perceptions most strongly relate to intention to buy it. The dependent variable is purchase intention, the mean of three items on a seven-point scale. The four predictors, each the mean of multi-item seven-point scales, are perceived quality, trust in the brand, fairness of the price and recall of the brand's recent advertising. Because the question asks how several interval-level predictors together and separately relate to one interval-level outcome, multiple regression is the appropriate procedure. The illustrative file contains 250 consumers who had seen the product in stores.

Descriptive Statistics and Correlations

Purchase intention was modest on average. It correlated moderately with quality and trust, weakly with price fairness and hardly at all with ad recall. Quality and trust were correlated with each other at .58, which means their unique contributions must be separated in the regression. Ad recall was nearly unrelated to every other variable, so whatever the regression found for it would closely match its simple correlation with intention.

SPSS Procedure

The model was estimated through Analyze, Regression, Linear, with purchase intention as the Dependent variable and the four perceptions entered together as Independents using the Enter method. Under Statistics, Estimates, Confidence intervals, Model fit, Descriptives, Collinearity diagnostics and Durbin-Watson were selected. Under Plots, standardized residuals were plotted against standardized predicted values, and a histogram and normal probability plot of residuals were requested. Under Save, Cook's distances and standardized residuals were added to the data file.

Results

The model was significant, F(4, 245) = 20.76, p < .001, with R = .50, R² = .253 and adjusted R² = .241. Taken together, the four perceptions accounted for roughly 25% of the variance in purchase intention, and the standard error of the estimate was 0.93 scale points. Expressed as Cohen's f², the model's effect was .34, just under the .35 that Cohen (1992) labeled a large effect for regression.

Brand trust had the largest unique relationship with intention, β = .30: holding the other perceptions constant, a one-point increase in trust was associated with a 0.34-point increase in intention. Price fairness, β = .22, and perceived quality, β = .20, also contributed significantly. Ad recall did not, B = 0.01, p = .845, and its confidence interval was narrow around zero.

The VIF values, all below 1.6, show that collinearity is not a serious problem. Mason and Perreault (1991) showed, however, that even moderate collinearity combined with a modest sample and R² can make coefficient estimates less precise, which is why the quality coefficient, whose predictor shares variance with trust, has a wider interval relative to its size than price fairness.

Assumption Checks

The plot of standardized residuals against predicted values showed no curved pattern or funnel shape, supporting linearity and constant variance. The residual histogram was roughly normal, with slight positive skew of 0.36, partly because seven respondents gave the lowest possible intention score. No standardized residual exceeded 3 in absolute value, and the largest Cook's distance was .047, far below the common threshold of 1, so no single case unduly influenced the results. Rerunning the model without the seven floor responses left quality, trust and price fairness significant and ad recall close to zero. The Durbin-Watson statistic of 2.06 indicated no autocorrelation among residuals, which is expected for cross-sectional survey data.

Interpretation for Management

For this brand, earning trust moves intention more than buying recall. The results suggest that the relaunch should emphasize trust and price fairness, for example through visible ingredient sourcing, third-party certifications or a satisfaction guarantee, and through a price that consumers see as reasonable for the quality. Advertising recall had no unique relationship with intention, suggesting that awareness alone does not drive interest; advertising may be more effective if it carries the trust-building message rather than merely keeping the brand visible. In practical terms, the coefficients suggest that a campaign raising average trust by half a point, holding other perceptions steady, would be associated with an increase of about 0.17 points in intention, while a similar gain in price fairness would be associated with about 0.10 points. These are modest shifts for an individual, but across a market of potential buyers they can be the difference between a relaunch that gains shelf space and one that loses it.

Limitations

The data are cross-sectional and correlational, so the coefficients show associations, not causes; consumers who already like the brand may report higher trust and higher intention. A controlled test of trust-building messages would be needed to establish effects. Purchase intention is also an imperfect proxy for behavior. Morwitz et al. (2007) reported that stated intentions track later purchases more closely for established products than for new ones, and for durables more than for nondurables, which means intention for a relaunched snack should be validated against trial and repurchase data once the bar is back on shelves. Finally, the model leaves three-quarters of the variation unexplained, so factors such as taste, availability and promotions deserve attention in future research.

References

Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159. https://doi.org/10.1037/0033-2909.112.1.155

Mason, C. H., & Perreault, W. D., Jr. (1991). Collinearity, power, and interpretation of multiple regression analysis. Journal of Marketing Research, 28(3), 268-280. https://doi.org/10.1177/002224379102800302

Morwitz, V. G., Steckel, J. H., & Gupta, A. (2007). When do purchase intentions predict sales? International Journal of Forecasting, 23(3), 347-364. https://doi.org/10.1016/j.ijforecast.2007.05.015