Ranchers or pastoralists? Farm size, specialisation and production strategy amongst cattle farmers in south-eastern Kazakhstan

Pastoralism

Research, Policy and Practice

Table 5 Geographical and historical factors associated with feeding strategy cluster

Variable		1: small sedentary	2: medium mobile	3: medium fodder purchaser	4: medium fodder producer	5: large mobile fodder purchaser	6: large mobile fodder producer	Total farms	*F or Pearson’s χ*^2#	*βlncattle*^α
N		40	56	25	32	27	20	200		β
Worked on sovkhoz (1/0)		0.5	0.54	0.36	0.59	0.22^-cv	0.65	0.48	13.3**	−0.38**
Agricultural education (1/0)		0.03	0.2	0	0.13	0.07	0.25	0.12	14.1**	0.29
District of residence	Enbek (1/0)	0.15	0.09	0.44^+cv	0.25	0.26	0.20	0.20	14.7**	0.30
	Kegen (1/0)	0.15	0.07^-cv	0.24	0.13	0.56^+cv	0.25	0.20	29.4***	0.40**
	Raiymbek (1/0)	0.7	0.84	0.32	0.63	0.19	0.55	0.59	44.7***	−0.50***
Population of home village†		47073	48433	915212	6254	7530	5849	6043	4.3***	660*
Distance from Almaty (km)^††		213	22953	1722	211	1722	200	205	5.6***	−12.88***

^†From 2009 census
^††Generated by straight line GIS distance tool—as crow flies
^#Continuous variables: OLS regressions with group as the categorial predictor. Statistic given is the F ratio for the overall model (***p < 0.01, **p < 0.05, *p < 0)
Superscripts 1, 2, 3, 4, 5 and 6 indicate groups to which difference is significantly different from zero at **p < 0.05 or lower using pairwise comparisons with Bonferroni correction
^#Binary variables: Pearson’s chi-squared test: ***p < 0.01, **p < 0.05, * p < 0
Superscript +/-cv for binary variables indicates those cells which have Pearson’s residuals greater than the critical value with Bonferroni correction (+ or − 2.86), indicating contribution to χ²
^αContinuous variables: OLS regression on log of cattle ownership; for binary variables: logistic regression