Table2 Table4

Table 3

Table 3:
Estimates (log odds) from discrete-time hazard models for first and higher-order births in Zimbabwe 1990-94, with interaction terms between individual and aggregate education1.
 
 
  Model 1 Model 2 Model 3 Model 4
 
FIRST BIRTHS
 
WOMAN'S EDUCATION  
  none or < 3 years of primary2  0  0  0  0
  3-6 years of primary  0.13  0.03  0.16  0.14
  complete primary (7 years) or < 2 years of secondary -0.23 -0.29 -0.19 -0.23
  2-3 years of secondary -0.54** -0.63* -0.49** -0.53**
  4 or more years of secondary -0.89*** -0.97*** -0.83*** -0.88***
 
ALTERNATIVE MAIN AND INTERACTION TERMS INVOLVING AGGREGATE EDUCATION3
  Average length -7  0.11      
  (Average length-7) *
(length of individual education-1)/11
-0.12      
 
  Proportion illiterate-0.16     -1.05  
  (Proportion illiterate-0.16) *
(length of individual education-1)/11
     0.79  
 
  Average length among 3+ - 8        0.15
  (Average length among 3+ - 8) *
(length of individual education-1)/11
      -0.21
 
  none or < 3 years of primary * (average length - 7)    0.13    
  3 or more years of primary education and < 2 years of secondary * (average length - 7)    0.04    
  2 or more years of secondary * (average length - 7)    0.02    
 
 
HIGHER-ORDER BIRTHS
 
WOMAN'S EDUCATION  
  none or < 3 years of primary2  0  0  0  0
  3-6 years of primary  0.01  0.00 -0.01  0.02
  complete primary (7 years) or < 2 years of secondary -0.10 -0.08 -0.11 -0.06
  2-3 years of secondary -0.20* -0.20* -0.22** -0.17
  4 or more years of secondary -0.38*** -0.36** -0.40*** -0.34**
 
ALTERNATIVE MAIN AND INTERACTION TERMS INVOLVING AGGREGATE EDUCATION3
  Average length -7 -0.01      
  (Average length-7) *
(length of individual education-1)/11
 0.02      
 
  Proportion illiterate-0.16     -0.02  
  (Proportion illiterate-0.16) *
(length of individual education-1)/11
     0.10  
 
  Average length among 3+ - 8       -0.07
  (Average length among 3+ - 8) *
(length of individual education-1)/11
       0.09
 
  none or < 3 years of primary * (average length - 7)   -0.01    
  3 or more years of primary education and < 2 years of secondary * (average length - 7)   -0.02    
  2 or more years of secondary * (average length - 7)    0.05    
 

1 Also age and religion were included in the first-birth models, and age, parity, duration since previous birth and religion were included in the higher-order birth models. See also notes 4 and 5.
2 Arbitrarily chosen reference category.
3 Length of education is defined as in the the text (i.e. it is set to 1, 4.5, 7.5, 9.5 and 12 years in the 5 categories), but to facilitate interpretation, it is subtracted by 1 year and divided by 11. Thus, 0 corresponds to the lowest educational level and 1 to the highest. `Proportion illiterate' is the proportion with less than 3 years of education, as in Table 1, and `Average length among 3+' is the average among those with 3 or more years of education. 7 or 8 years are subtracted from the averages and 0.16 from the proportion illiterate to get main effects close to those reported in Table 1.
4 In addition, urban/rural character of the current place of residence was included as a main effect and in interaction with length of individual education. Because rural is defined as 0 and urban as 1 also in these interactions, the effects of individual education shown in the table are for rural women.
5 In addition, urban/rural character of the current place of residence was included in interactions with the three educational groups, just as average length of education. Because rural is defined as 0 and urban as 1 also in these interactions, the effects of individual education shown in the table are for rural women.
* significant at the 0.10 level ** significant at the 0.05 level *** significant at the 0.01 level

 

Table2 Table4

A Search for Aggregate-Level Effects of Education on Fertility, Using Data from Zimbabwe
Øystein Kravdal
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871
http://www.demographic-research.org/Volumes/Vol3/3