Life Tables contd'

1
Life Tables contd.

• Questions?
• Review
• Today
• The essence of cohort survival
• Fertility/Births component

2
Mortality and Life Tables

• We need to determine s -- the sex-specific and
  age-cohort-specific survival rate
• For live births
• For each cohort to the next
• For the last, open-ended cohort
• Life Table is the key to determining s

3
Class Exercise

• Using the Alberta Male Life Tables 1995-1997,
  Compute
• S1,2 cohort 1 to cohort 2, 0-9, 10-19
• S0,1 birth to 0-9 cohort
• S8,8 open-ended 70 cohort
• Hint Use Table 5.2 of your text as a template

4
Cohort Survival

• We now know sn,n1 -- the sex-specific and
  age-cohort-specific survival rate
• Given the number of males in the 0-9 cohort in
  Lethbridge in 2002, how many will survive until
  2012?

5
We need to know
From the Lethbridge census
From the Alberta life table
6
Compute

• 3995 0.9970 3983
• An average of 12 0-9 year-olds will not survive
  the 10 year period t, t1
• Of course this is an expected number based on
  probabilities derived from 1995-97.
• This is the essence of the cohort survival
  technique

7
Fertility/Births Component

• Issues
• Computing average number of women per fecund
  cohort including a very important migration
  component!
• Computing fertility rate
• Sex assignment
• Survival into first cohort

8
Fertility and Births Model
9
Computing fertility
Age cohort of mother
Reminder that all births enter the n0 cohort
10
fn,0

• Fertility cultural variable
• Fecundity biological variable
• Fertility
• Spontaneous abortion
• Miscarriage (reported and unreported)
• Prenatal care and obstetric technology
• Environmental factors
• Abortion (reported and unreported)
• Fecund age group seems to be increasing
• We ignore these important factors!

11
Regional Fertility

• Varies by region
• More regional variance than survival rates?
• Thus we need a regional measure
• Small regions are subject to random variation
  from year-to-year in births
• Take a three year period centred on year in
  question to dampen variance

12
Births Data

• Need data based on place of residence of birth
  mother
• Provincial birth certificate data should show
  this
• Not births based on location of birth
• hospital location and residence pose issues

13
Potential Mothers The Population at Risk

• Mortality within fecund cohorts minor
• Migration within fecund cohorts - major
• Need average number of females in a fecund cohort
  over the time period t, t1
• note error p. 83

14
Potential Mothers The Population at Risk

• Corrections See text p. 83
• Must average by dividing by 2
• Female migrants also averaged
• We do not apply survival rate to net migrants
• Assume that PM is not a constraint

15
Now we model births

• Total births over the time period t,t1
• For fecund cohorts 2-5, over a ten year period to
  match cohort size

16
Modeling Sex Assignment

• Sex ratio is males/females per hundred
• Typically 105 in North America
• In China now, 120 and up to 130 in some regions,
  ultrasound and sex selection via abortion
• We express as a proportion for modeling
• Pf 0.485

17
Projecting Births by Sex
18
Projecting the first cohort
19
Putting it all together females
20
Putting it all together males