ZEMCH 2015 - International Conference Proceedings | Page 500

levels. Both the size of the extensions and the presence of changes to internal partitions were
strongly correlated to satisfaction with acoustic privacy levels, while the presence of self-built
external roofs was positively correlated to results on natural light perception. No statistically significant correlations were found between personalisation strategies and winter or summer temperatures, safety, ventilation, neither with visual privacy satisfaction levels. However, income decile was correlated to satisfaction levels with natural light (rs=-0.366, n=699, p=.002) and moisture
conditions (rs=-0.218, n=96, p=.002), while the users’ former residential situation was correlated to
satisfaction with summer temperatures (rs=-0.228, n=105, p=.019).
Table 6. Spearman’s Correlation results between satisfaction and personalisation variables
Independent Variable

Dependent Variable

Coef.

Sig.

n

n bedrooms

extension size

-0.783

<.001

16

lot size

extension size

+0.660

=.005

16

extension possibilities

extension size

+0.510

=.044

16

modification capabilities

kitchen partitions

+0.213

=.035

98

kitchen size

kitchen partitions

+0.249

=.012

101

finishing works

no modification

-0.449

<.001

94

bathroom location

extension size

-0.475

=.063

16

moisture levels

extension end use

-0.202

=.065

84

acoustic privacy

extension size

-0.703

=.002

16

-

change partitions

+0.200

=.044

102

natural light

external roofs

-0.389

<.001

109

After segmentation by dwelling type, developer, personalisation levels, and general demographic characteristics, Kuskal-Wallis H test evidenced that the largest number of significant differences
in satisfaction levels resulted from the two first grouping variables (Figure 6). As shown in Table 7, when dwelling type was used as segmenting variable, satisfaction ranks for dwelling size,
modification capabilities, kitchen size, bathroom location, moisture levels, visual privacy, safety,
and winter temperatures were significantly different. When segmented by developer, differences in satisfaction ranks for number of bedrooms, modification capabilities, bathroom location,
and acoustic privacy were also significant. In contrast, when these satisfaction levels were segmented by income decile, age of the oldest female and number of users, only satisfaction with
current number of bedrooms showed significant differences among groups (H=12.275, p<.001),
whilst when segmented by personalisation extent and workmanship quality only differences in
finishing works were significant (H=71.163, p<.001). Although no significant differences were found
in finishing works when segmented by developer or by dwelling type, the differences were significant when segmented by income decile (e.g. H=17.270, p=.008 for floors and H=13.001, p=.043 for
wall finishing works).
Interestingly, when expectation levels were segmented by developer, architectural quality expectations were significantly higher for the NGO group (H=42.516, p<.001) whilst no significant
differences were found in terms of environmental quality expectations. Nonetheless, when environmental quality expectations were used as segmenting variable, satisfaction levels with both
moisture (H=11.773, p=.003) and winter temperatures (H=8.958, p=.011) were significantly different,
whilst differences in satisfaction with kitchen sizes (H=12.750, p=.002) were only significant when
segmented by expected architectural quality.

498

ZEMCH 2015 | International Conference | Bari - Lecce, Italy