You are on page 1of 2

A study was conducted to determine the effects of using chicken manure at various levels on net fish

yield (NFY) per annum. It is hypothesized that increasing levels of chicken manure application weekly
leads to increased levels of NFY because of increased primary productivity and availability of natural
food. Curve fitting estimation using different methods of regression was performed to test this
hypothesis. These include Linear, Quadratic and exponential regressions in a bid to select the best fitted
curve that explains the relationship and ultimately gives credence or otherwise to the hypothesis.
Results
Table 1: Different Regression models
Equation
Model Summary
R
2
F df1 df2 Sig.
Linear .654 47.304 1 25 .000
Logarithmic .832 123.732 1 25 .000
Quadratic .914 127.923 2 24 .000
Cubic .919 86.453 3 23 .000
Exponential .652 46.748 1 25 .000

From Table 1, although the cubic equation has the highest R
2
, the quadratic method is the most ideal
since it has a high R
2
value among the remaining equation types which were all significant and also
because it is easier to explain.

Figure 1: Relationship between chicken manure use and net fish yield.
Without application of manure, 1 ton of NFY is achieved. Maximum production level of 7 tons can be
achieved at 668kg/ha/wk application of chicken manure. The inclusion of chicken manure explains 91%
of the variation in net fish yield (Figure 1). The regression curve is significant (Table 1, p = 0.000). The
relationship as shown by the equation in figure 1 means that for 52kg/ha/year addition of chicken
manure, there is an 18kg/ha/year gain in NFY up till the highest feasible CM level of 31 tons/ha/year
(p<0.05) but with increasing levels of chicken manure above this maximum level, the NFY will decrease
at a rate of 0.01kg/ha/year. It therefore follows that increasing chicken manure application will increase
NFY up to a maximum point after which a decline is experienced as a result of other factors which have
not been confounded in this experiment.
M
a
x
.

L
e
v
e
l

Data on mean annual concentration of SO
2
(g/m
3
) in 20 American cities was collected as index of
pollution. It is believed that several variables including two anthropogenic variables: number of
manufacturing enterprises employing 20 or more workers and population size, as well as four climatic
averages for weather stations at these cities, average annual temperature, average annual wind speed,
average annual precipitation, and average number of days with precipitation per year are responsible
for rising levels of SO
2
in these cities or alternatively only one or more of these factors do not contribute
to rising SO
2
levels. A multiple linear regression was carried out to determine which of these variables
is/are actually responsible or interact to cause the rising levels of SO
2
hence pollution.
Results
Forward elimination method revealed three models with significant regressions as shown in table 1.
Table 1: Possible regression models
Model R
R
Square
Adjusted
R
Square
Std.
Error of
the
Estimate
1 .811
a
.657 .638 16.09668
2 .855
b
.732 .700 14.65100
3 .891
c
.793 .755 13.25367

Table 2: Model Summary
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig. B
Std.
Error Beta
3 (Constant) 83.963 25.087 3.347 .004
Factory .027 .004 .740 6.262 .000
Temp -1.822 .561 -.516 -3.248 .005
Precip .852 .390 .341 2.185 .044

Model three as seen from table 2 has three variables confounded with significant effects (R
2
= 0.891,
F
3,16
= 20.47, p = 0.000) hence it is the model of choice. The full linear relationship can be written as:

This implies that a unit increase in the number of factories (with 20 employees) will lead to the addition
of 0.027 g/m
3
of SO
2
into the atmosphere if mean annual temperature and mean annual precipitation
are not changed. A unit Increase in temperature will reduce SO
2
concentration by 1.822 g/m
3
while an
increase in annual precipitation will increase SO
2
by 0.852 g/m
3
. However, the interplay of the three
factors: Number of factories, average annual temperature and average annual precipitation all
contribute to increasing levels of SO
2
in the atmosphere. Therefore three factors: number of factories
(with 20 employees), mean annual temperature and mean annual precipitation are critical in
determining SO
2
pollution in the cities examined.