Question $7$

$[22]$

HBAT management has long been interested in more accurately predicting the satisfaction level of its customers. If successful, it would provide a better foundation for their marketing efforts. To this end, researchers at HBAT proposed that multiple regression analysis should be attempted to predict the customer satisfaction based on their perceptions of HBAT's performance. In addition to finding a way to accurately predict satisfaction, the researchers also were interested in identifying the factors that lead to increased satisfaction for use in differentiated marketing campaigns. Thus, explanation was a critical

objective since this analysis was intended to provide management with the key elements necessary to improve customer satisfaction. To apply the regression procedure. researchers selected customer satisfaction $(x_{19})$ as the dependent variable $(Y)$ to be predicted by $13$ independent variables representing perceptions of HBAT's performance $(x_6$ to $x_{18}).$ Bnef descriptions of the independent variables are included in Table $1.$ The relationship among the $13$ independent variables and customer satisfaction was assumed to be statistical, not functional, because it inwolved perceptions of performance and may include levels of measurement error.

HBAT's management team is inferested in any differences in perceptions between those customers located and served by their Gauteng$-$based salesforce versus those outside the South Africa who are served mainly by independent distributors. Their specific focus is to identity any perceptions that difler in terms of their primary areas of operations $($product line. proing. elc $)$ between these two sets of customers. One objective is to identify which perceptionis$)$ provide dstinctive differences to develop differentated strategles between the customer aroups But a second objective, class fication, is equally important as a means of deselophing support for the modet among operationar units. The abilty of the management team to easity distinguish between customers even withouf knowing theif geographic lacation makes for improved salesforce coorctination. Moreower, identifing customers which are dissimiter to ther contemporanes thate ailow for even more differentated services

Swhich $0.024$ Now promel L

STAlB$3$O $\mathrm{.}70$M fougnext

Ind tosd or loab

differentiated services.

Table $7a$

$\backslash$table$[[$Owerali Model Fn: Goodness$-$di$-$Ft Me$3$sures$],[,$CHANGE IN $-2$LL$],[,,$Froms Rase Model,From Prior Step,$],[,$Vatae,Chance,Ser.,Chane,$31.,],[2$ Log Likelhood$(\mathrm{.}2$L$),39\mathrm{.}95,42,148,0,17,011,0\mathrm{.}000,],[$Cox and Sneb $N^r,0,505,,,,,],[$Nrociterter $0^2,0,577,,,,-,],[$Pseuso $R^2,0\mathrm{.}512,,,,,],[,$value,Sic.,,,,$],[$Fosmer and Lemeshow $x^2,5,306,6722,,,,],[$variaties in the reyution,,,,,,$],[$independern Vetimble,$6,59$d$.$ ellou,Wald,$,5$ s$,$explib$)],[K_{12}$ Oampethuefroing,$1\mathrm{.}079,0\mathrm{.}357,$$$,135,1,0,013,2\mathrm{.}992]]$

Ky Price FexibilityConstantVarables not in the equationindependet VariableX ProductQualityX,ECommetce ActiutiosX, Techrical SupportX, Complaint ResclutionAya AvetisingB$-$logstie cowiciant, ExpB$)$sponntiated coeicientKa ProductineNy Salastorce ImaeXWaranty & ClaimsX New ProductsX Dalreey SpeedClascication atraclualGroupiHampshGouteng$1,844$Cutsde Gateng$-14,192$AN$6$LYSSSAMPLE$0,639$Gautute$(96\mathrm{.}213,71214,614$Sette Statist$0,6563,5010\mathrm{.}0060,6030,0913,409$O$,$B$4912\mathrm{.}327$PredctadGecup Mombershipo,$02629028,331$OutsidGauten$20$Total$188\mathrm{.}31$ellulena$0,0046,321$E$920,000$Sigficance$0,4180,0610\mathrm{.}9370,4050,7620\mathrm{.}0050,3670\mathrm{.}122$D$,8730,919$HOLDOUT SAMPLE$0\mathrm{.}088$CutsdnGaecg Total$82610\mathrm{.}000413$

$\backslash$table$[[$Actial Croup,Qoutens,$\backslash$table$[[$Dutside$],[$Guytang$.]],$Total,Gaulehy,$\backslash$table$[[$Oucside$],[$Gautens$.]],$Total$],[$Memhership$,$Segtar:,$26,9,4,13],[$Gaident$,196\mathrm{.}2,,,(692)21,,],[$Cusside Gauteng,$6,28,34,2,25,27$

$18507$