BUSI 820 Week 5 Discussion Forum 5 Main Thread Example

Discussion Forum 5: Main Thread and Forum 4 Reply

BUSI 820 Quantitative Research Methods, Discussion Forum 5 Main Thread (with the Forum 4 reply)

Thread by [Student Name]

The data set is illustrative, generated for this page with a fixed random seed and analyzed with the statistics SPSS reports for correlation and bivariate regression; the bank, the classmate and the prompt are composites standing in for the course's own; the books are real; the week, points and readings follow Liberty's public course guide for BUSI 820.

The second research question in my turnover study asks whether supervisor support is associated with intention to leave. Both are summated scales, and in Forum 2 their skewness values were 0.21 and 0.09, comfortably inside the band that Morgan et al. (2020) accept for parametric work. A scatterplot of the 186 respondents showed a loose, roughly linear cloud sloping downward, with no curve or extreme outliers, so Pearson's r is appropriate.

The correlation was negative and statistically significant, r(184) = -.18, p = .015, with a 95 percent confidence interval from -.31 to -.04. Spearman's rho, computed on ranks and therefore sturdier against skewed scores, gave nearly the same value, rho = -.17, p = .020, so the normality assumption is not carrying the conclusion. Employees who rate their supervisors as more supportive tend to report less intention to leave.

A bivariate regression expresses the same relationship in the units of the scales. The equation was predicted intention to leave = 16.73 minus 0.16 times supervisor support. An employee whose support score is ten points higher, for example 25 rather than 15 on the 6-to-30 scale, is predicted to score about 1.6 points lower on the 5-to-25 intention scale. That is a real difference, but a modest one.

The effect size is where interpretation matters most. Squaring r gives about .03, so supervisor support shares roughly 3 percent of the variance in intention to leave. Cohen (1988) would call a correlation of this size small, and the confidence interval allows anything from a near-zero relationship to a small-to-medium one. With 186 respondents, a weak relationship clears the significance bar; the p value reflects sample size as much as strength. For the bank, this means supervisor support is worth attention but is unlikely to solve turnover on its own; 97 percent of the variation in intention to leave lies elsewhere.

Finally, the result is correlational. Employees who plan to leave may also judge their supervisors more harshly, so the direction of influence could run the other way, and both variables may reflect a third factor such as branch workload. I will describe support as associated with lower intention to leave, not as reducing it. The multiple regression planned for Forum 7 will show whether the association holds once branch size, work arrangement and tenure are controlled.

Reply to Victor (Forum 4)

Victor, your Forum 4 thread used chi-square to test whether customer complaints differ across your four distribution centers, with complaint type, late or damaged, as the second variable. Your table showed that the smallest center recorded only 9 complaints in total, so two of its cells have small counts. Morgan et al. (2020) advise checking expected counts and warn that the test becomes unreliable once roughly a fifth of the cells expect fewer than five cases. With eight cells, even two small cells would reach that threshold. Could you report the expected counts from the SPSS output? If they fall below 5, combining the two smallest centers or collecting a second quarter of data would make the test more trustworthy, and adding Cramér's V would show how strong the association is.

References

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates.

Morgan, G. A., Barrett, K. C., Leech, N. L., & Gloeckner, G. W. (2020). IBM SPSS for introductory statistics: Use and interpretation (6th ed.). Routledge.