Statistical Analysis of Pig Growth Rates in Vietnam - Research Paper Example

1. Statistical Analysis of effects of Litter Mortality, Birth Weight, Farm of Origin, Year of Birth on Growth Performance of Pigs in Vietnam. 2. Introduction Climate and nutrition highly affect the growth rates of domestic animals specially pigs. In Vietnam, farmers lack the knowledge to provide adequate conditions of nutrition and weather protection to have a good growth rate average of pigs they raise (Phuoc & Ngoan 2005). Research is being conducted on different effects of weather, nutrition and other factors to determine the factors that mostly affect the growth rates of pigs. In this study, two western breeds of pigs from Landrace and Yorkshire were monitored in a number of farms in Vietnam. The data of litter mortality, birth weight, farm of origin, year of birth, and weight of big after twenty one days concerning a sample of one thousand pigs were collected and recorded. The purpose of the study is to compare growth performance of these two breeds of pigs by analyzing the effect of the different factors monitored in isolation and combined on the growth of pigs represented by their average weight after twenty one days of their birth. 3. Materials and Methods Data were analyzed by ANOVA, linear models, t-tests, regression, and other statistical methods using Office Excel 2003. 4. Empirical Results To determine the relationship between the Day 21 average piglet weight and the various influences monitored during the experiment such as breed, farm of origin, year of birth, litter mortality and birth weight, the following Ordinary Least Squares (OLS) regression analyses were examined: 4.1 Analysis I: Relationship between Day 21 Average Piglet Weight and Breed. The non-numerical value of Breed was coded numerically so that Breed equals one for Landrace (L) piglets and equals to zero for Yorkshire (Y). Regression analysis was applied with the numerical representation of breed as the independent variable and the dependent variable Day 21 average weight. Table 1 summarizes the results of the regression analysis. Table 1: Regression results for Day 21 Average Piglet Weight as a function of Breed Independent Variable Breed Dependent variable: Day 21 Average Piglet Weight Constant 4.846032 Coefficient 0.035295 T Stat 0.46674 R2 0.000218 Adjust R2 -0.00078 Standard Error 0.793569 Observations 1000 Table 1 shows adjusted R-squared of -0.00078 which means the independent variable breed predicts 0.07% of the dependent variable Day 21 average piglet weight. T-stat for this variable is 0.46674 so it is statistically insignificant. Both values indicate changes in breed do not affect the Day 21 average piglet weight. 4.2 Analysis II: Relationship between Day 21 Average Piglet Weight and Farm of Origin The non-numerical value of the farm-of-origin variable is numerically encoded to allow statistical analysis of the variable using the following code shown in table 2. Table 2: Encoding of Farm of Origin into Number Farm of Origin Numerical Code A 1 B 2 C 3 D 4 E 5 F 6 G 7 Regression analysis was applied with the numerical representation of farm of origin as the independent variable and the dependent variable Day 21 average weight. Table 3 summarizes the results of the regression analysis. Table 3:Regression results for Day 21 Average Piglet Weight as a function of Farm of Origin Independent Variable Enumerated Farm of Origin Dependent variable: Day 21 Average Piglet Weight Constant 5.038045 Coefficient -0.04348 T Stat -2.87992 R2 0.008242 Adjust R2 0.007248 Standard Error 0.790379 Observations 1000 Table 3 shows adjusted R-squared of 0.007248 which means the independent variable farm of origin predicts 0.7% of the dependent variable Day 21 average piglet weight which is still a small influence but with more effect than the breed variable. T-stat for this variable is -2.87992 so it is statistically significant. Both values indicate the limited effect of breed in Day 21 average piglet weight. The Day 21 weight can be predicted to some limited extent using the equation: Day 21 Average Weight of Piglet = 5.038045 - 0.04348 (farm of origin number) +E 4.3 Analysis III: Relationship between Day 21 Average Piglet Weight and Year of birth. Regression analysis was applied on the independent variable of year of birth in relation to the dependent variable Day 21 average weight. Table 4 summarizes the results of the regression analysis. Table 4: Regression results for Day 21 Average Piglet Weight as a function of Year of Birth Independent Variable Year of Birth Dependent variable: Day 21 Average Piglet Weight Constant -4.58144 Coefficient 0.004736 T Stat 0.219229 R2 0.006939 Adjust R2 4.82E-05 Standard Error -0.00095 Observations 0.793637 Table 4 shows adjusted R-squared of 0.000048 which means the independent variable year of birth predicts almost zero percent of the dependent variable Day 21 average piglet weight. T-stat for this variable is 0.219229 so it is statistically insignificant. Both values indicate the no effect of year of origin on Day 21 average piglet weight. 4.4 Analysis IV: Relationship between Day 21 Average Piglet Weight and Litter Mortality. Regression analysis was applied with the independent variable of litter mortality and the dependent variable Day 21 average weight. Table 5 summarizes the results of the regression analysis. Table 5: Regression results for Day 21 Average Piglet Weight as a function of Litter Mortality Independent Variable Litter Mortality Dependent variable: Day 21 Average Piglet Weight Constant 4.914733 Coefficient -1.0831 T Stat -3.0986 R2 0.009529 Adjust R2 0.008536 Standard Error 0.789866 Observations 1000 Table 5 shows adjusted R-squared of 0.008242 which means the independent variable litter mortality predicts 0.8% of the dependent variable Day 21 average piglet weight which is still a small influence similar to the farm of origin effect. T-stat for this variable is -3.0986 so it is statistically significant. Both values indicate the limited effect of litter mortality in Day 21 average piglet weight. The Day 21 weight can be predicted to some limited extent using the equation: Day 21 Average Weight of Piglet = 4.914733 - 1.0831 (litter mortality) +E Figure 1 demonstrates the inconsistency of the line drawn by the equation to predict Day 21 average weight of Piglet using litter mortality. Figure 1: Litter Mortality Line fit Plot (Excel 2003) 4.5 Analysis V: Relationship between Day 21 Average Piglet Weight and Birth Weight. Regression analysis was applied with the independent variable of birth weight and the dependent variable Day 21 average weight. Table 6 summarizes the results of the regression analysis. Table 6: Regression results for Day 21 Average Piglet Weight as a function of Birth Weight Independent Variable Birth Weight Dependent variable: Day 21 Average Piglet Weight Constant 3.570759 Coefficient 1.012168 T Stat 6.636745 R2 0.042269 Adjust R2 0.041309 Standard Error 0.776701 Observations 1000 Table 6 shows adjusted R-squared of 0.041309 which means the independent variable birth weight predicts 4.1% of the dependent variable Day 21 average piglet weight. T-stat for this variable is 6.636745 so it is statistically significant. Both values indicate that birth weight has the most effect out of measured variables in determining Day 21 average piglet weight. The Day 21 weight can be predicted using the equation: Day 21 Average Weight of Piglet = 3.570759 + 1.012168 (birth weight) +E Figure 2 demonstrates the level of approximation the line drawn by the equation provides while predicting Day 21 average weight of Piglet in relation to birth weight. Figure 2: Birth Weight Line fit Plot (Excel 2003) 4.5 Regression Analysis VI: Relationship between Day 21 Average Piglet Weight and Birth Weight, litter mortality, and farm of origin. These three variables: birth weight, litter mortality, and farm of origin were proven to be statistically significant in predicting the dependent variable Day 21 average weight. Table 7 summarizes the results of the regression analysis. Table 7: Regression results for Day 21 Average Piglet Weight as a function of Birth Weight, Litter Mortality, and Farm of Origin Independent Variables: Birth Weight, litter mortality, and Farm of Origin Dependent variable: Day 21 Average Piglet Weight Constant 3.693075 Coefficient of Farm of Origin -0.0523 T Stat of Farm of Origin -3.55772 Coefficient of Liter Mortality -1.30327 T Stat of Liter Mortality -3.8245 Coefficient of Birth Weight 1.102912 T Stat of Birth Weight 7.27243 R2 0.067439 Adjust R2 0.06463 Standard Error 0.767196 Observations 1000 Table 7 shows adjusted R-squared of 0.06463 which means the independent variables birth weight, litter mortality, and farm of origin predict 6.4% of the dependent variable Day 21 average piglet weight. T-stats for the three variables are respectively -3.55, -3.82, and 7.27 which indicate statistical significance of all three variables in predicting Day 21 average weight. The combined three variables have the most effect when grouped together to determine Day 21 average piglet weight. The Day 21 weight can be predicted using the equation: Day 21 Average Weight of Piglet = 3.693075 + 1.102912 (birth weight) - 1.30327 (litter mortality) - 0.0523 (farm of origin number) +E 5. Discussion In order to determine the effects of different factors such as breed, farm of origin, year of birth, litter mortality and birth weight (W0) in the average weight of piglets after twenty one days (W21), a number of regression analyses are completed. The results of the regression analysis of year of birth and breed as independent variables shows they have almost no effect on predicting the value of the dependent variable W21. The regression analysis of litter mortality and farm of origin shows the statistical significance yet limited ability of both independent variables in predicting and influencing 0.8% and 0.7% respectively of the value of W21. Regression analysis of W0 on W21, show the statistical significance and ability of W0 to influence and predict 4.1% of the values of W21. The ability to predict the value of W21 is further enhanced by regression analysis of three variables of litter mortality, farm of origin, and W0 together. The three variables influence and predict 6.4% of W21 all together. It is concluded that the factors of litter mortality, farm of origin and birth weight have the most effect on pigs' future weight. Vietnam needs to concentrate its pig farms in farms proved to provide better growing conditions. They also need to research methods to maintain zero litter mortality as to enhance weight of their grown pigs. Further research is also need to determine the conditions that provide highest weights at birth as it was statistically proven to affect the future weight of pigs. References Excel (2003). Microsoft Office Professional. Phuoc, L.V. & Ngoan, L.D. (2005). Effect of the Environmental Factors on Physiological Parameters, feed intake and Growth of Mong Cai and Landrace Pigs in Central Vietnam. Hue University of Agriculture, Vietnam. Read More