Skip to main content

Research, Policy and Practice

Table 3 Farm characteristics by feeding strategy (means include zeros unless specified in notes)

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

Variable   Cluster   Test statistic for differences between clusters#  
HH 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 Relationship with herd sizeα
  N 50 40 56 25 32 27 20 200 LRχ2 or F βlncattle
Livestock Cattle (head)† 9 122,3,5,6 171,5,6 221,5,6 165,6 831,2,3,4 891,2,3,4 33 223*** 1.00***
Range 1–39 3–34 6–31 5–80 7–34 18–395 32–340 3–395   
Sheep and goats (head) 20 315,6 505,6 325,6 495,6 4541,2,3,4 3851,2,3,4 132 145*** 1.01***
Livestock units†† 66 972,5,6 1551,5,6 1345,6 1385,6 9791,2,3,4 9661,2,3,4 330 273*** 0.99***
Proportion of cattle in LU 0.60 0.515 0.49 0.605 0.51 0.361,4 0.44 0.49 4.79*** −0.00
Land Cropland (ha)† 1.2 24,6 63,4,5,6 12,4,6 231,2,3,5 12,4,6 461,2,3,5 11 32.5*** 0.28***
Hayland (ha) 4.2 73 113 21,2,4,5 163 263 343 14 5.8*** 0.47***
Cropland/head cattle (ha) 0.28 0.194 0.424,5 0.094,6 1.401,2,3,5 0.032,4,6 0.623,5 0.46 12.3*** −0.06*
Hayland + cropland/head (ha) 0.99 0.913,5 0.993,5 0.201,2,4,6 2.343,5 0.351,2,4,6 1.053,5 1.01 12.2*** −0.09**
Pasture (ha) 6.4 212,5,6 741,3,5,6 112,5,6 365,6 3181,2,3,4 2791,2,3,4 103 31.5*** 1.35***
Pasture/LU (ha)¤ 0.52 0.54 0.63 0.32 0.46 0.40 0.46 0.33 1.27 −0.08
Fodder (kg/head/year) Total fodder kg/ha 2121 21752,6 13771 18995 19706 13303 9471,4 1647 24.3*** −0.21***
Roughage, total 1709 2047 1288 1622 15956 1265 7834 1477 20.7*** −0.22***
Concentrate, total 412 1284 893,4 2772,5 3761,2,5,6 653,4 1644 170 41.4*** −0.11
Concentrate, purchased† 369 1244,6 884,6 2704,6 51,2,3,5 654 181,3,4 94 56.9*** −0.22
Concentrate, self-produced† 43 44,6 14,6 74,6 3711,2,3 0 1461,2,3 76 66.6*** 0.01
Roughage, purchased† 701 425 142 15434,6 623,5 8774 773 454 18.7*** 0.01
Roughage, self-produced† 1008 16223,5 11463,5 791,2,4,5,6 15333,5 3881,2,3,4 7063 1023 54.7*** −0.47***
  Quality (protein/kg fodder) 90 99 108 103 1195 924,6 1225 107 3.83*** 0.00
Mobility Months on remote pasture† 2 12,3,4,5,6 81,3,4 51,2,5,6 51,2,5,6 111,3,4 101,3,4 6 159.9*** 0.37***
Remote pasture winter (1/0) 0.04 0-cv 0.36+cv 0.16 0.19 0.78+cv 0.70+cv 0.33 37.2*** 1.14***
Remote pasture summer (1/0) 0.28 0.20-cv 0.95+cv 0.64 0.75 0.96+cv 1.0+cv 0.73 77.6*** 1.13***
Any mobile stock (1/0) 0.28 0.20-cv 0.95+cv 0.60 0.78 0.89 0.95 0.72 79.3*** 0.95***
Sales Cattle (1/0) 0.44 0.33 0.55 0.64 0.38 0.74 0.60 0.52 16.3*** 0.77***
Milk (1/0) 0.10 0.10 0.09 0.20 0.03 0.07 0.05 0.09 5.6 0.09
Milk products (1/0) 0.04 0.07 0.11 0.08 0 0.11 0.20 0.09 6.6 0.62**
Beef (1/0) 0.10 0.07 0.09 0.12 0.03 0.15 0.30+cv 0.11 10.5* 0.61***
Carcass weight bullocks (kg)¤¤ 139 130 150 180 183 157 207 166 2.5** 28.8***
  1. #Continuous variables: Test statistic likelihood ratio (χ2) values for negative binomial regressions with cluster as the categorial predictor against the variables for livestock numbers, months on pasture, and fodder. For carcass weight, cattle as the proportion of LU, protein content and continuous land variables (log transformed), OLS regression (ANOVA) was used and the F ratio is reported. ***p < 0.01, **p < 0.05, *p < 0.1
  2. Superscripts 1, 2, 3, 4, 5 and 6 indicate clusters with which a significant difference exists at **p < 0.05 or lower using pairwise comparisons with Bonferroni correction
  3. #Binary variables: Pearson’s chi-squared test on variable listed against the cluster, ***p < 0.01, **p < 0.05, *p < 0.1
  4. Superscript +/-cv for binary variables indicates cells with Pearson’s residuals greater than the critical value with Bonferroni correction (+ or − 2.86), indicating contribution to χ2
  5. αSlope coefficients of regressions of log of cattle ownership against outcome variables. For carcass weight, cattle as the proportion of LU and land variables (the last of which were log transformed), OLS regression was employed; for counts of livestock numbers, months on pasture and kilogramme fodder, negative binomial regressions were used. For binary outcomes, logistic regression was employed, ***p < 0.01, **p < 0.05, *p < 0.1
  6. Variable used to create a cluster
  7. ††Livestock units based on Kazakh sheep units (sheep and goat = 1 LU, horse = 6 LU and cattle = 5 LU)
  8. ¤Those with titles or contracts for pastureland indicating area only, n = 132, no zeros
  9. ¤¤Sellers only, N = 48, no zeros