Geoscience Reference
In-Depth Information
Table 3.3 Summary of significant predictors of access to higher amount of water information
across all six water information maps
%
Latino
% School-aged
children
%
Renters
Adjusted
R
2
West Valley
(a) Count-based census
+
0.00
(b) Economic-proxy based census
+
+
0.43
(c) Count-based zip code
+
0.42
(d) Economic-proxy based zip code
+
+
0.27
(e) Count-based distance buffer
+
+
0.10
(f ) Economic-proxy based distance
buffer
*
+*
+*
0.15
Phoenix
(a) Count-based census
.
.
.
(b) Economic-proxy based census
+*
0.05
(c) Count-based zip code
+
0.14
(d) Economic-proxy based zip code
+
+
0.40
(e) Count-based distance buffer
+*
*
+*
0.18
(f ) Economic-proxy based distance
buffer
+
+
+*
0.11
East Valley
(a) Count-based census
+
+*
0.13
(b) Economic-proxy based census
+
+*
0.11
(c) Count-based zip code
+
*
+*
0.28
(d) Economic-proxy based zip code
+
+
0.29
(e) Count-based distance buffer
+*
+*
0.11
(f ) Economic-proxy based distance
buffer
+*
+
+*
0.14
Consistency map
*
+
+*
0.08
Logistic regression for each geographic unit
Note: + indicates a positive association with high water effort;
indication a negative association
with high water effort; * indicates a statistically significant relationship (p
0.05)
<
randomly in any representation. Collectively, the maps show variability in the
spatial arrangement of high and low amounts of water information. Most notably,
the distance buffer technique (Figs.
3.5e
, f) varies dramatically from both the
census blocks and zip code aggregation methods. There are also notable differences
in the ranking produced by using the program count and economic proxy technique.
The most notable differences occur in central Phoenix, where census tracts and
zipcodes are much smaller due to higher population densities, restricting the
geographic extent of some cultural resources. Table
3.3
contains descriptive statis-
tics. Notably, Phoenix has no areas classified as receiving high water information
availability ranking in either census tract aggregation. Mean Latino ranges from
13.34 to 31.17%, percent School-aged children ranges from 23.03 to 25.57%, and
percent renters ranges from 16.32 to 34.34%.
Separate logistic regressions for the East Valley, Phoenix, and the West Valley
tested the relationship between our predictor variables (% Latino, % renters, and the
