27.10.2012 Views

Analysis of Sales Promotion Effects on Household Purchase Behavior

Analysis of Sales Promotion Effects on Household Purchase Behavior

Analysis of Sales Promotion Effects on Household Purchase Behavior

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

7.5 Results Drivers <str<strong>on</strong>g>of</str<strong>on</strong>g> <str<strong>on</strong>g>Promoti<strong>on</strong></str<strong>on</strong>g> Resp<strong>on</strong>se<br />

In this secti<strong>on</strong>, the hypotheses as derived in Secti<strong>on</strong> 3.2 and 3.3 will be empirically tested<br />

using 17,144 promoti<strong>on</strong>al purchase records from 156 households for the period (1996(II) –<br />

1997(IV)). The structure <str<strong>on</strong>g>of</str<strong>on</strong>g> this secti<strong>on</strong> is as follows. The results using the singlepromoti<strong>on</strong><br />

data, the decomposed multiple-promoti<strong>on</strong> data and the combined data are<br />

compared in sub-secti<strong>on</strong> 7.5.1. Subsequently, the obtained results are dealt with in secti<strong>on</strong><br />

7.5.2 in detail. Each hypothesis is tabulated al<strong>on</strong>g with the estimated coefficients regarding<br />

the variable tested. Finally, the implicati<strong>on</strong>s <str<strong>on</strong>g>of</str<strong>on</strong>g> the findings are discussed in Secti<strong>on</strong> 7.5.3.<br />

7.5.1 Data Legitimacy Check<br />

Before discussing the findings in detail, first the validity <str<strong>on</strong>g>of</str<strong>on</strong>g> the findings is checked by<br />

comparing the estimates for the combined-record data format (where all promoti<strong>on</strong><br />

shopping trips are decomposed in such a way that each data record deals with <strong>on</strong>ly <strong>on</strong>e type<br />

<str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong>) with the single-promoti<strong>on</strong> data (where <strong>on</strong>ly those records are selected where<br />

<strong>on</strong>ly <strong>on</strong>e type <str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong> is present) and the decomposed multiple-promoti<strong>on</strong> data<br />

(where those records are selected where more than <strong>on</strong>e different type <str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong> was<br />

present during a specific shopping trip, these shopping trips are decomposed such that<br />

every record deals with <strong>on</strong>ly <strong>on</strong>e type <str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong>) (see Secti<strong>on</strong> 7.3 for details about these<br />

three data formats). The results <str<strong>on</strong>g>of</str<strong>on</strong>g> the binary logistic regressi<strong>on</strong> analyses for these three<br />

different data inputs can be found in Table A7.2. The significance levels reported for the<br />

categorical variables indicate whether the variables itself are significant or not.<br />

Overall, the parameter estimates for the three groups are reas<strong>on</strong>ably in agreement.<br />

When we look at the parameter estimates for the possible drivers <str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong> resp<strong>on</strong>se,<br />

the overall picture that emerges <str<strong>on</strong>g>of</str<strong>on</strong>g> drivers <str<strong>on</strong>g>of</str<strong>on</strong>g> promoti<strong>on</strong> resp<strong>on</strong>se is also c<strong>on</strong>sistent (most<br />

significant estimates have the same sign and about the same size). Although some<br />

differences are present, the relative ranking <str<strong>on</strong>g>of</str<strong>on</strong>g> the parameter estimates for the different<br />

categories <str<strong>on</strong>g>of</str<strong>on</strong>g> a variable lead to the same findings.<br />

125

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!