Statistical modelling/Numbers of cattle
In order to test whether the homesteads, for which livestock provides the major source of income, possess more cattle on
average than those for which livestock does not, a linear mixed model (REML), with sub-region and dip-tank area within
sub-region assumed to be random effects and primary source of income as a fixed effect, was fitted to the transformed cattle numbers in
CS11Data1. To perform the analysis, the relevant
GenStat input is entered in the dialog box obtained by clicking Stats →
Mixed Models (REML) → Linear Mixed Models and putting the variable livst_inc2 in the Fixed model: window and subregion/diptank in the Random model: window.
There is a significant difference (P<0.05) between the two groups of homesteads.
**** REML Variance Components Analysis ***** |
Response Variate : log_cat_no
|
|
*** Estimated Variance Components ***
|
Random term |
Component |
S.e.
|
subregion |
0.0154 |
0.0173
|
subregion.diptank |
0.0341 |
0.0198
|
*** Residual variance model *** |
Term |
Factor |
Model |
Parameter |
Estimate
|
S.e.
|
Residual |
|
Identity |
Sigma2 |
0.442
|
0.0341
|
*** Wald tests for fixed effects *** |
Fixed term |
Wald statistic |
d.f. |
Wald/d.f. |
Chi-sq prob |
|
livst_inc2 |
6.03 |
1 |
6.03 |
0.014 |
|
|