Checked, found and fixed, in that order: GB701's Unit 9 memo tests six regression assumptions and shows which conclusions survive robust errors and a log outcome. Searches like "gb 701 unit 9 assignment example", "gb701 unit 9 sample" and "gb701 unit 9 example" land here.
What a finished GB701 Unit 9 assumption diagnostics memo looks like
Five pages with four figures: residuals against fitted values, a normal Q-Q plot, a scale-location plot and a Cook's distance chart. A table lists each assumption, the check used, the result and the response. Variance inflation factors run from 1.02 to 1.21. A squared tenure term adds nothing, p of .729. Residual spread grows across thirds of the fitted values, standard deviations of 9.0, 12.7 and 15.5 days, and the Koenker form of the Breusch-Pagan test returns 25.33 on 5 degrees of freedom, p below .001. Shapiro-Wilk gives W of .976, p of .001, with residual skew of 0.59. Fourteen accounts exceed the 4/n Cook's threshold, the largest at .078. HC3 errors keep e-invoicing at -6.44, interval -10.19 to -2.69, and a log model estimates a 15.1 percent reduction.
How a GB701 Unit 9 example is structured
The memo's logic runs from check to consequence, so that each failed assumption is tied to the specific inference it threatens. Heteroscedasticity leaves the coefficients unbiased but makes ordinary standard errors unreliable, which is why the remedy is HC3 standard errors rather than a new model; the e-invoicing interval barely moves. Non-normal residuals matter less for coefficient tests at n of 216 than for prediction intervals, and the memo says that plainly rather than calling the result harmless or fatal. A log-transformed outcome serves as a sensitivity model: its residuals pass both checks, Breusch-Pagan p of .464 and Shapiro-Wilk p of .376, and it puts the e-invoicing association at -15.1 percent, interval -22.9 to -6.5. Influence is handled by refitting without the fourteen flagged accounts, which moves the coefficient to -5.25 days. The raw-days model stays primary because its units match the business question.
Collinearity, cleared quickly
Every variance inflation factor sits below 1.25, so the predictors carry distinct information and the standard errors are not inflated by overlap. The memo notes that credit limit, dropped earlier for correlating at .97 with purchases, would have failed this check badly.
The fan in the residual plot
Residual spread rises with predicted days, from a standard deviation of 9.0 in the lowest third to 15.5 in the highest. The Breusch-Pagan result confirms what the plot shows, and the memo explains why slower-paying accounts plausibly vary more.
Normality, handled honestly
Shapiro-Wilk rejects normality, yet with 216 cases the sampling distribution of each coefficient is close to normal. The memo accepts the coefficient tests, warns that prediction intervals for single accounts would be too narrow, and avoids declaring the violation either harmless or fatal.
Fourteen influential accounts
None has a Cook's distance near 0.5, and the largest leverage belongs to an account buying $481,000 a year. Refitting without all fourteen changes the e-invoicing coefficient from -6.44 to -5.25 days, a shift the memo reports without deleting anyone from the primary analysis.
Two remedies, one primary model
HC3 standard errors address unequal variance directly, and the log model tests whether conclusions depend on the outcome's scale. Both agree on direction and approximate size, so the memo keeps days as the reporting unit and presents the log model as a check.
Where marks go in GB701 Unit 9
Diagnostics memos in GB701 draw criticism when they list tests without consequences. A Breusch-Pagan statistic reported with no statement of what it changes shows the procedure was followed and not understood. Declaring normality violated and then abandoning the model, or declaring it satisfied because a histogram looked roughly bell-shaped, both miss the doctoral standard of proportionate judgment. Deleting influential cases from the primary analysis without justification is questioned sharply, since it lets the analyst choose the answer. Instructors also expect remedies matched to problems: robust standard errors for unequal variance, a transformation or a different model for a scale problem, and a sensitivity analysis for influence. A memo showing which conclusions survive every remedy, and which do not, tends to earn the strongest comments. Plots must be labeled well enough to read without the text.
Get a GB701 Unit 9 example written to your instructions
Send the model output or data, the checks the assignment requires and the Unit 9 rubric. Once the custom diagnostics memo is done, inside 24-48h, every check has been run, each failure reported in plain terms and every remedy tied to the inference it protects. Your first sample is free.
GB701 Unit 9 questions, answered
Which normality test should a GB701 diagnostics memo use?
Shapiro-Wilk is widely used and appears in most statistical software, but at larger sample sizes it flags departures too small to matter. Pair it with a Q-Q plot and a statement of what the regression's inferences actually depend on. For coefficient tests with a few hundred cases, the plot and the reasoning usually carry more weight than the p-value.
Are robust standard errors enough to fix heteroscedasticity?
For inference about coefficients, heteroscedasticity-consistent errors such as HC3 generally suffice, and R, Stata and SPSS extensions all provide them. They do not repair prediction intervals, and they do not address a misspecified model. If the fan pattern suggests a scale problem, adding a transformed outcome as a sensitivity check is worth the extra page.
Should influential cases be removed?
Rarely from the primary analysis, unless they are errors. Report them, check whether each reflects a data problem, and refit without them as a sensitivity analysis. If conclusions hold either way, say so; if they change, the influential accounts become a finding worth discussing, since they may represent a distinct kind of customer.