Home » Topics » Fertility » Calculating Period Parity Progression Ratio
Calculating Period Parity Progression Ratio [message #7074] |
Wed, 19 August 2015 06:17 ![Go to next message Go to next message](/theme/default/images/down.png) |
|
Hello everyone,
I'm having issues sorting out the data required for calculating Period PPR using synthetic parity cohort method.
I've been able to work out the year each woman had a specific birth order and the number of women involved.
I'm working with the Nigeria DHS 2013.
gbemtrol
|
|
|
Re: Calculating Period Parity Progression Ratio [message #7076 is a reply to message #7074] |
Wed, 19 August 2015 15:09 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Liz-DHS
Messages: 1516 Registered: February 2013
|
Senior Member |
|
|
Dear User,
We don't do Period Parity Progression Ratios standardly in DHS, however Dr. Tom Pullum is willing to share Quote:something relevant to this that I wrote a few years ago. This is Chapter 16 from a book:
Pullum, T. W. 2004. Natality: Measures based on censuses and surveys. Chapter 16 of The Methods and Materials of Demography, D. Swanson and J. Siegel (eds.), Academic Press.
Pages 426-428 are relevant.
I am attaching a file as well.
Thank you!
-
Attachment: MandM16.pdf
(Size: 229.40KB, Downloaded 1351 times)
[Updated on: Wed, 19 August 2015 15:15] Report message to a moderator
|
|
|
|
|
Re: Calculating Period Parity Progression Ratio [message #7080 is a reply to message #7079] |
Thu, 20 August 2015 11:48 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Liz-DHS
Messages: 1516 Registered: February 2013
|
Senior Member |
|
|
Here is Dr. Pullum's response:
Quote:To get that specific table in Stata, you could use the following lines. Note that in that survey, year of birth is coded with only two digits. For example, "1995" is coded "95". More recent surveys generally use four digits for the year.
The main complication is that the births in the birth histories are sequenced in reverse temporal order. That is, birth 1 is the most recent birth. You need to work from bord, with gives the birth order of each birth. One way to do that, with local notation, is shown here. There are other ways.
Another complication is that the indexes for the births in the birth history are 01, 02,....09, 10, 11, etc. I recommend that the leading 0 for 01 through 09 be removed with a rename command, as shown.
Let me know if you have other questions.
* Construction of table 16.29 in Chapter 16 of Methods and Materials of Demography
* open PHIR3BFL.dta
set more off
* get the names of the b variables
describe b*01
describe bord*
* there are up to 20 births in the birth histories
* remove the unnecessary 0's in the index for the b variables
rename b*_0* b*_*
* The births in the birth history are numbered in reverse order;
* it will be convenient to resequence them by birth order
* This approach can be used for other years and parities
gen year_1=.
gen year_2=.
local li=1
while `li'<=20 {
replace year_1=b2_`li' if bord_`li'==1
replace year_2=b2_`li' if bord_`li'==2
local li=`li'+1
}
tab year_1 year_2 if year_1>=90 & year_2>=90
|
|
|
|
|
Re: Calculating Period Parity Progression Ratio [message #8691 is a reply to message #8674] |
Wed, 02 December 2015 13:23 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Liz-DHS
Messages: 1516 Registered: February 2013
|
Senior Member |
|
|
Dear User,
Here is a response from Dr. Tom Pullum:
Quote:I don't have the original programs, but the Stata lines to reproduce table 16.28, which gives the number of births by birth order and calendar year, are simple:
* To reproduce table 16.28 in Methods and Materials of Demography, 2nd edition
* Philippines 1998 survey; use the BR file
* The table in M&M only goes to order 8
* Note that year of birth, b2, in this survey, gives only the last two digits
*open PHBR5BFL.dta
set more off
* unweighted, as in table 16.28
tab bord b2 if b2>=90
* weighted, which would be typical for analysis
tab bord b2 [iweight=v005/1000000] if b2>=90
I do not have time to re-write the Stata code to reproduce table 16.5, which is woman-years of exposure in each combination of five-year age intervals and five-year time intervals. DHS does not normally go back in time for more than 15 years, and the time intervals are generally 0-4 years before the survey, 5-9 years, before the survey, etc., rather than time intervals defined by calendar years such as 1995-99. You can look for a link to a fertility rates program elsewhere on the forum or the website. Such programs will produce the numerators and the denominators of the rates, as well as the rates themselves.
The approach I generally take for such a table is to construct, for each woman in the IR file, the first and last cmc when she was in each age interval, and the first and last cmc when she was in each time interval. (For calendar years of time, the first and last months will be the same for all woman. For years ago, the first and last months will differ somewhat from one woman to another, depending on the month in which she was interviewed.) You then count up the number of months each woman lived in each combination of age and time. Then add up across all women and divide by 12 to go from months to years. You can do the addition with the collapse command in Stata.
[Updated on: Wed, 02 December 2015 14:22] Report message to a moderator
|
|
|
|
Re: Calculating Period Parity Progression Ratio [message #13942 is a reply to message #12421] |
Sun, 28 January 2018 04:56 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
rajaram
Messages: 3 Registered: March 2013 Location: mumbai
|
Member |
|
|
Dear sir/madam,
I'm having issues sorting out the data required for calculating Period PPR using synthetic parity cohort method.
I've been able to work out the year each woman had a specific birth order and the number of women involved.
I'm working with the India NFHS 2016. I am using the below syntax to calculate year of birth for each birth order.
local n=1
while `n'<=15 {
gen year_`n'=.
local n=`n'+1
}
local li=1
while `li'<=20 {
replace year_1=b2_`li' if bord_`li'==1
replace year_2=b2_`li' if bord_`li'==2
replace year_3=b2_`li' if bord_`li'==3
replace year_4=b2_`li' if bord_`li'==4
replace year_5=b2_`li' if bord_`li'==5
replace year_6=b2_`li' if bord_`li'==6
replace year_7=b2_`li' if bord_`li'==7
replace year_8=b2_`li' if bord_`li'==8
replace year_9=b2_`li' if bord_`li'==9
replace year_10=b2_`li' if bord_`li'==10
replace year_11=b2_`li' if bord_`li'==11
replace year_12=b2_`li' if bord_`li'==12
replace year_13=b2_`li' if bord_`li'==13
replace year_14=b2_`li' if bord_`li'==14
replace year_15=b2_`li' if bord_`li'==15
local li=`li'+1
}
recode year_1 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_1) //**for parity of 1995-1999 all births from 19985 to 1995 has been taken***\\
recode year_2 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_2)
recode year_3 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_3)
recode year_4 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_4)
recode year_5 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_5)
recode year_6 (1985/1999=1)(2000/2004=2)(2005/2009=3)(2010/2014=4)(else=.) ,gen(parity_6)
cummulative parity by year. 0 order parity is taken as women in the age group 30-49 for year 1995-99 and age group 25-49 for year 2000-2004.
Year parity0 parity1 parity2 parity3 parity4 parity5 parity6
1995-99 29,655 23607 18294 12,382 7,167 3,484 1,476
2000-04 41,739 34600 28451 20,896 13,481 7,767 4,093
2005-09 55,584 45999 39062 29,290 19,453 11,724 6,588
2010-2014 67,106 58636 50037 36860 24,022 14,523 8,318
please let me know where i am doing mistake. Please let me the procedure to calculate pariry progression ratio from NFHS data.
I will be thankful to you.
[Updated on: Sun, 28 January 2018 04:58] Report message to a moderator
|
|
|
|
Goto Forum:
Current Time: Fri Aug 16 22:50:55 Coordinated Universal Time 2024
|