The Role of Neighborhood Origin in the Residential Satisfaction …
441
3.3.1 Effect of Neighborhood Origin on RS
In order to find out whether there is a difference in mean RS between residents of
informal origin neighborhoods and residents of formal origin neighborhoods, we ran
a Welch t-test because the assumption of homogeneity of variances was violated, as
assessed by Levene’s test for equality of variances (p ≤ 0.001). Data are mean ±
standard deviation, unless otherwise stated. There were 254 participants from formal
origin neighborhoods and 277 from informal origin neighborhoods. There were no
outliers in the data, as assessed via a boxplot, and the RS scores for each level of neighborhood origin were normally distributed, as assessed by the Shapiro-Wilk’s test (p
> 0.05). Residents of formal origin neighborhoods had higher residential satisfaction
(M = 10.55, SD = 1.82) than did those living in informal origin neighborhoods (M
= 9.58 SD = 2.61), which constitutes a statistically significant difference: M = 0.97,
95% CI [−1.35, −5.75], t(529) = −4.89, p ≤ 0.001.
Of the two kinds of neighborhoods under investigation, we expected the informal
origin neighborhoods to score higher on RS because we hypothesized that for lowincome people, community ties and social networks would be fundamental to the
needs of everyday life (Addo 2016; Amérigo and Aragones 1990). Residents of
informal origin neighborhoods usually share the same struggles, settle with extended
family nearby, and develop a strong sense of pride in and belonging to the residential
environment that they have built themselves (Lara 2012). However, we had expected
higher satisfaction with community and neighborhood to result in overall higher RS
for residents living in informal origin neighborhoods relative to those living in formal
origin neighborhoods. Yet, our results do not support this hypothesis.
3.3.2 Examining Built Environment Features that Contribute to RS
The t-test revealed a significant difference in RS between those living in formal origin
neighborhoods versus those living in informal origin neighborhoods, with residents
in the former group enjoying significantly higher RS. In the next step, we focused
on determining the features of the built environment that contribute to RS for the
residents of each kind of neighborhood. We ran a Multiple Linear Regression (MLR)
on informal origin neighborhoods and formal origin neighborhoods separately to
compare the models and identify the variables that contribute to RS in both formal
origin and informal origin neighborhoods. We used the backwards
1 variable selection
method, as this analysis was exploratory in nature. Through this method, we were
able to find the best fit for each group of participants (formal origin and informal
origin) by choosing from all the environmental features examined (PREQI variables).
1 Even though the backwards selection method is not usually considered valid because it relies
on computer algorithms rather than theoretical input, in this case all the variables had already
been validated in other studies as highly correlated with the dependent variable RS. The purpose
was not to find the best predictive model of RS. Instead, the models highlight the differences and
commonalities between the groups of interest.
441
3.3.1 Effect of Neighborhood Origin on RS
In order to find out whether there is a difference in mean RS between residents of
informal origin neighborhoods and residents of formal origin neighborhoods, we ran
a Welch t-test because the assumption of homogeneity of variances was violated, as
assessed by Levene’s test for equality of variances (p ≤ 0.001). Data are mean ±
standard deviation, unless otherwise stated. There were 254 participants from formal
origin neighborhoods and 277 from informal origin neighborhoods. There were no
outliers in the data, as assessed via a boxplot, and the RS scores for each level of neighborhood origin were normally distributed, as assessed by the Shapiro-Wilk’s test (p
> 0.05). Residents of formal origin neighborhoods had higher residential satisfaction
(M = 10.55, SD = 1.82) than did those living in informal origin neighborhoods (M
= 9.58 SD = 2.61), which constitutes a statistically significant difference: M = 0.97,
95% CI [−1.35, −5.75], t(529) = −4.89, p ≤ 0.001.
Of the two kinds of neighborhoods under investigation, we expected the informal
origin neighborhoods to score higher on RS because we hypothesized that for lowincome people, community ties and social networks would be fundamental to the
needs of everyday life (Addo 2016; Amérigo and Aragones 1990). Residents of
informal origin neighborhoods usually share the same struggles, settle with extended
family nearby, and develop a strong sense of pride in and belonging to the residential
environment that they have built themselves (Lara 2012). However, we had expected
higher satisfaction with community and neighborhood to result in overall higher RS
for residents living in informal origin neighborhoods relative to those living in formal
origin neighborhoods. Yet, our results do not support this hypothesis.
3.3.2 Examining Built Environment Features that Contribute to RS
The t-test revealed a significant difference in RS between those living in formal origin
neighborhoods versus those living in informal origin neighborhoods, with residents
in the former group enjoying significantly higher RS. In the next step, we focused
on determining the features of the built environment that contribute to RS for the
residents of each kind of neighborhood. We ran a Multiple Linear Regression (MLR)
on informal origin neighborhoods and formal origin neighborhoods separately to
compare the models and identify the variables that contribute to RS in both formal
origin and informal origin neighborhoods. We used the backwards
1 variable selection
method, as this analysis was exploratory in nature. Through this method, we were
able to find the best fit for each group of participants (formal origin and informal
origin) by choosing from all the environmental features examined (PREQI variables).
1 Even though the backwards selection method is not usually considered valid because it relies
on computer algorithms rather than theoretical input, in this case all the variables had already
been validated in other studies as highly correlated with the dependent variable RS. The purpose
was not to find the best predictive model of RS. Instead, the models highlight the differences and
commonalities between the groups of interest.