yu issn 1450-863 x godina xiii broj 1 / 2011. - horizonti.ekfak.kg.ac.rs
yu issn 1450-863 x godina xiii broj 1 / 2011. - horizonti.ekfak.kg.ac.rs
yu issn 1450-863 x godina xiii broj 1 / 2011. - horizonti.ekfak.kg.ac.rs
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
EKONOMSKI HORIZONTI, 2011, 13, (1)<br />
A form of the model of preferences for the entire study group is obtained<br />
by calculating the arithmetic mean of the paramete<strong>rs</strong>’ values for e<strong>ac</strong>h of the<br />
respondents.<br />
Main advantage of conjoint analysis is also the possibility to establish the<br />
importance of certain aspects of a given product. It is determined with a use of a<br />
following formula (Hair, Ande<strong>rs</strong>on 1998 p. 608):<br />
where:<br />
s<br />
s<br />
max { U<br />
jl<br />
} − min { U<br />
jl<br />
}<br />
l<br />
j l<br />
j<br />
s<br />
j<br />
j<br />
W =<br />
× 100 % (2),<br />
j m<br />
⎛<br />
s<br />
s<br />
max { } min { }<br />
⎞<br />
∑ ⎜ U<br />
jl<br />
− U<br />
jl ⎟<br />
⎝<br />
l<br />
j l<br />
j<br />
j<br />
j ⎠<br />
j = 1<br />
s<br />
U<br />
jl j<br />
- partial utility for l-level j-th variable for respondent s,<br />
s<br />
W<br />
j<br />
- importance of e<strong>ac</strong>h variable for respondent s,<br />
lj<br />
- number of level for variable Z j<br />
Similarly as in the case of the paramete<strong>rs</strong>, the importance of the attributes<br />
for the total sample is calculated as the arithmetic mean of the values for individual<br />
respondents.<br />
The obtained model provides a basis for the further analysis. Depending on<br />
the purpose of the study, the results can be applied to determine the price and<br />
product concept, market segmentation, competitive analysis, advertising and<br />
selection of the distribution channel (Wittink, Cattin 1989). In the next sections of<br />
this article an example analysis will be provided.<br />
2. CHARACTERISTICS OF THE RESEARCH SAMPLE<br />
In order to determine the structure of preferences for particular aspects of a<br />
bank <strong>ac</strong>count, in April 2010 a survey among students of Cr<strong>ac</strong>ow Unive<strong>rs</strong>ity of<br />
Economics was conducted. Out of 206 questionnaires received, 188 were<br />
completed correctly. Othe<strong>rs</strong> were excluded from the analysis.<br />
Respondents were char<strong>ac</strong>terized by the following features:<br />
a) gender,<br />
b) pl<strong>ac</strong>e of residence,<br />
c) work status,<br />
d) amount of money allowed to spend e<strong>ac</strong>h month,<br />
129