Geography Reference
In-Depth Information
Table 1.3.
GWR regression of railway station proximity and mean terrain elevation on
percent change in wheat acreage, 1881-1891
Type of multiple regression analysis
Ordinary least squares
Geographically weighted
Cases to fit (N)
283
283
Adjusted r-square
0.017
0.62
Parameters
P-values
P-values
Intercept
Not significant ( >0.10)
0.03
Distance from station (in meters)
0.01
0.00
Mean terrain elevation (in meters)
Not significant ( >0.10)
0.00
Parameters: five number summaries of regression coefficients
Parameter
Minimum
Lower quartile
Median
Upper quartile
Maximum
Intercept
-89.91
-35.61
-18.91
-0.55
102.03
Distance from
-0.02
0.00
0.00
0.00
0.02
station (in meters)
Mean terrain
-119.39
-1.63
0.19
1.81
49.14
elevation (in meters)
Note:
W heat acreage is a smoothed rate calculated as a geographically weighted average in a
moving window of ten parishes.
production stands out clearly. he intensiication of dairy- and beef-catle
farming is less pronounced than might be expected when the returns for
those two dates are used. The decline in wheat farming was dramatic.
In the parishes of central and upland Dorset, wheat acreage estimated
in 1871 had fallen by half or more in 1901, well in line with the national
average. In the same period, the density of beef and dairy catle remained
stable overall and increased in two clusters of parishes noted on figure 1.4.
In Dorset we can dig deeper into the decline of wheat growing. There
at the parish level, results from a geographically weighted regression
(GWR) model estimate the degree of change in wheat acreage, the inde-
pendent variable, explained by the interacting effects of the mean eleva-
tion of parish terrain and the distance from its center (centroid) to the
closest railway station.
32
The results show that 62 percent of the variation
in wheat acreage over the decade 1881 to 1891 can be accounted for by rail
accessibility and mean elevation (table 1.3).
The effect of proximate rail service was varied and complex, for it
carried a negative or positive influence on changes in wheat acreage de-
