Multiple Causes of Food Insecurity Multiple Regression Analysis

  • Slides: 23
Download presentation
Multiple Causes of Food Insecurity: Multiple Regression Analysis Reference: Gujarati (2004)

Multiple Causes of Food Insecurity: Multiple Regression Analysis Reference: Gujarati (2004)

THE THREE-VARIABLE MODEL

THE THREE-VARIABLE MODEL

THE THREE-VARIABLE MODEL

THE THREE-VARIABLE MODEL

THE THREE-VARIABLE MODEL

THE THREE-VARIABLE MODEL

Maternal education and community characteristics as indicators of nutritional status of children – application

Maternal education and community characteristics as indicators of nutritional status of children – application of multivariate regression Nutritional Status of Children 5

Main results • • • Step 1: Estimating the coefficients of the model(Table 10.

Main results • • • Step 1: Estimating the coefficients of the model(Table 10. 2, 10. 3). Step 2: Examining how good the model predicts(Table 10. 4, 10. 5). Step 3: Hypotheses testing. Tests about the equation(Table 10. 6, 10. 7). Tests about individual coefficients(Table 10. 8, 10. 9). Part and partial correlation coefficients. Step 4: Checking for violations of regression assumptions. Checking normality of the errors(Table 10. 10, Figure 10. 1, 10. 2). Checking for homogeneity of variance of the residuals(Figure 10. 3, 10. 4). Nutritional Status of Children 6

Empirical analysis Nutritional Status of Children 7

Empirical analysis Nutritional Status of Children 7

Empirical analysis Data description and methodology Nutritional Status of Children 8

Empirical analysis Data description and methodology Nutritional Status of Children 8

Table 10. 1 Means and standard deviations of variables: Malawi sample Mean Standard deviation

Table 10. 1 Means and standard deviations of variables: Malawi sample Mean Standard deviation BFEEDNEW 1. 57 0. 49 DIARRHEA 0. 16 0. 37 CALREQ 0. 32 0. 47 ATTCLINI 1. 32 0. 59 AGEMNTH 27. 29 16. 28 CLINFEED 2. 29 4. 57 DRINKDST 1. 98 1. 39 EDUCSPOUS 2. 23 1. 30 LATERINE 0. 40 0. 49 PXFD 9. 00 9. 85 Variables Nutritional Status of Children 9

Table 10. 2 Determinants of weight for age Z-scores Variables Model 1 Model 2

Table 10. 2 Determinants of weight for age Z-scores Variables Model 1 Model 2 Coefficients Std. error Constant -3. 03 0. 504 -2. 93 0. 492 EDUCSPOUS 0. 173 0. 058 0. 156 0. 057 ATTCLINI 0. 439 0. 144 0. 439 0. 142 DRINKDST -0. 143 0. 051 -0. 147 0. 05 LATERINE 0. 134 0. 152 0. 148 PXFD 0. 009 0. 008 AGEMNTH -0. 084 0. 021 -0. 088 0. 021 AGESQ 0. 001 0. 0003 CLINFEED 0. 717 0. 181 0. 699 0. 179 DIARRHEA -0. 608 0. 201 -0. 620 0. 198 BFEEDNEW 0. 664 0. 226 0. 625 0. 224 0. 566 0. 212 -0. 06 0. 072 CALREQ HEALTDST -0. 094 0. 071 Nutritional Status of Children 10

Table 10. 3 Determinants of height for age Z-scores Variables Model 1 Model 2

Table 10. 3 Determinants of height for age Z-scores Variables Model 1 Model 2 Coefficients Std. error Constant -2. 159 0. 677 -2. 564 0. 685 EDUCSPOUS 0. 048 0. 072 0. 106 0. 074 ATTCLINI 0. 514 0. 18 0. 402 0. 177 DRINKDST -0. 024 0. 064 -0. 018 0. 062 LATERINE 0. 072 0. 192 0. 114 0. 185 PXFD 0. 014 0. 01 AGEMNTH -0. 091 0. 034 -0. 094 0. 034 AGESQ 0. 001 0. 0004 0. 0009 0. 0004 CLINFEED 0. 686 0. 227 0. 761 0. 224 DIARRHEA -0. 829 0. 261 -0. 932 0. 260 BFEEDNEW 0. 149 0. 305 0. 174 0. 299 0. 252 0. 085 -0. 127 0. 091 INSECURE HEALTDST -0. 106 0. 092 Nutritional Status of Children 11

Equations to predict the weight for age Z-scores for model 1 Nutritional Status of

Equations to predict the weight for age Z-scores for model 1 Nutritional Status of Children 12

Table 10. 4 Summary of the model for determinants of weight for age Z-scores

Table 10. 4 Summary of the model for determinants of weight for age Z-scores Adjusted R 2 Std. error of the estimate Models R R 2 1 0. 488 0. 238 0. 203 1. 09 2 0. 506 0. 256 0. 222 1. 079 Nutritional Status of Children 13

Table 10. 5 Summary of the model for determinants of height for age Z-scores

Table 10. 5 Summary of the model for determinants of height for age Z-scores Models 1 2 Adjusted R 2 Std. error of the estimate R R 2 0. 418 0. 174 0. 13 1. 285 0. 448 0. 201 0. 157 1. 265 Nutritional Status of Children 14

Table 10. 6 Analysis of variance table for weight for age Z-scores Sum of

Table 10. 6 Analysis of variance table for weight for age Z-scores Sum of Squares df Mean Square F Sig Regression 89. 139 11 8. 104 6. 797 0 Residual 284. 936 239 1. 192 Total 374. 075 250 Regression 95. 757 11 8. 705 7. 475 0 Residual 278. 318 239 1. 165 Total 374. 075 250 Models I II Nutritional Status of Children 15

Table 10. 7 Analysis of variance table for height for age Z-scores Sum of

Table 10. 7 Analysis of variance table for height for age Z-scores Sum of Squares df Mean Square F Sig Regression 70. 948 11 6. 45 3. 901 0. 00 Residual 335. 649 203 1. 653 Total 406. 597 214 Regression 81. 525 11 7. 411 4. 628 0. 00 Residual 325. 072 203 1. 601 Total 406. 597 214 Models I II Nutritional Status of Children 16

Table 10. 8 Tests of individual coefficients for determinants of weight for age Z-scores

Table 10. 8 Tests of individual coefficients for determinants of weight for age Z-scores Variables Model 1 Model 2 t-stat P value Constant -6. 016 0. 000 -5. 956 0. 000 EDUCSPOUS 2. 987* 0. 003 2. 70* 0. 007 ATTCLINI 3. 047* 0. 003 3. 102* 0. 002 DRINKDST -2. 817* 0. 005 -2. 938* 0. 004 LATERINE 0. 882 0. 378 1. 03 0. 304 PXFD 1. 185 0. 237 AGEMNTH -4. 041* 0. 000 -4. 282* 0. 00 AGESQ 3. 096* 0. 002 3. 386* 0. 001 CLINFEED 3. 966* 0. 000 3. 911* 0. 00 DIARRHEA -3. 018* 0. 003 -3. 126* 0. 002 BFEEDNEW 2. 935* 0. 004 2. 791* 0. 006 2. 668* 0. 008 -0. 851 0. 396 CALREQ HEALTDST -1. 32 0. 188 Note: * denotes at 1 per. Status centoflevel of significance. Nutritional Children 17

Table 10. 9 Tests of individual coefficients for determinants of height for age Z-scores

Table 10. 9 Tests of individual coefficients for determinants of height for age Z-scores Variables Model 1 Model 2 t-stat P value Constant -3. 187 0. 002 -3. 745 0. 000 EDUCSPOUS 0. 661 0. 509 1. 437 0. 152 ATTCLINI 2. 863* 0. 005 2. 273** 0. 024 DRINKDST -0. 37 0. 712 -0. 286 0. 775 LATERINE 0. 375 0. 708 0. 615 0. 539 PXFD 1. 479 0. 141 AGEMNTH -2. 642* 0. 009 -2. 78* 0. 006 AGESQ 1. 949** 0. 053 2. 022** 0. 044 CLINFEED 3. 018* 0. 003 3. 394* 0. 001 DIARRHEA -3. 171* 0. 002 -3. 584* 0. 00 BFEEDNEW 0. 489 0. 625 0. 580 0. 563 2. 977* 0. 003 -1. 395 0. 164 INSECURE HEALTDST -1. 15 0. 252 Note: * denotes at 1 per cent level. Nutritional of significance, Status** ofdenotes Childrenat 5 per cent level of significance. 18

Table 10. 10 Part and partial correlation coefficients for weight for age and height

Table 10. 10 Part and partial correlation coefficients for weight for age and height for age Variables Weight for age Part correlation Partial correlation Height for age Part correlation Partial correlation EDUCSPOUS 0. 169 0. 19 0. 042 0. 046 ATTCLINI 0. 172 0. 193 0. 183 0. 197 DRINKDST -0. 159 -0. 179 -0. 024 -0. 026 LATERINE 0. 057 0. 024 0. 026 PXFD 0. 067 0. 076 0. 094 0. 103 AGEMNTH -0. 228 -0. 253 -0. 168 -0. 182 AGESQ 0. 175 0. 196 0. 124 0. 136 CLINFEED 0. 224 0. 248 0. 192 0. 207 DIARRHEA -0. 17 -0. 192 -0. 202 -0. 217 BFEEDNEW 0. 166 0. 187 0. 031 0. 034 HEALTDST -0. 075 -0. 085 -0. 073 -0. 08 Nutritional Status of Children 19

Figure 10. 1 Histogram of standardized residuals of weight for age Nutritional Status of

Figure 10. 1 Histogram of standardized residuals of weight for age Nutritional Status of Children 20

Figure 10. 2 Normal P-P plot of regression standardized residuals Nutritional Status of Children

Figure 10. 2 Normal P-P plot of regression standardized residuals Nutritional Status of Children 21

Figure 10. 3 Residuals plotted against predicted values for weight for age Nutritional Status

Figure 10. 3 Residuals plotted against predicted values for weight for age Nutritional Status of Children 22

Figure 10. 4 Residuals plotted against predicted values for height for age Nutritional Status

Figure 10. 4 Residuals plotted against predicted values for height for age Nutritional Status of Children 23