# Estimating population size of under 2s

This question was posted the Assessment and Surveillance forum area and has 26 replies.

### Alex Mokori

Nutrition consultant

Normal user

16 May 2012, 11:52

### Mark Myatt

Consultant Epidemiologist

Frequent user

23 May 2012, 10:20

**(1) Rough and ready general method.**We usually assume that, in low income countries, about 18% of the population is under 5 years of age. If we assume uniformity then we have 3.6% of the population in each year under the age of five years. This means that we can expect about 7.2% of the population to be below 2 years. The formula would be:

```
N(children < 2 years) = population * 0.072
```

**(2) Using country specific data.**This involves you finding data on the local population structure. This can be found in census reports and may also be available in DHS / MICS survey reports and from other sources (e.g. UNICEF websites, World Bank website, CIA World Factbook). You may be able to find the exact data that you are looking for. If not, you will usually be able to find data that can calibrate the the rough and ready approach (1) above. For example, the UNICEF country statistics for Bangladesh website reports:

```
Total population (1000s) : 148,692
Population (1000s) under 5 : 14,707
```

Using this:
```
N(children > 2 years) = 148,692 * (14,707 / 148,692) * (2 / 5)
= 5,883 THOUSAND (5.9 million)
```

If your were working in a district in Bangladesh then you would use:
```
N(children > 2 years) = District population
* (14,707 / 148,692) * (2 / 5)
N(children > 2 years) = District population * 0.04
```

I hope this helps.### Chantal Autotte Bouchard

AAH

Normal user

24 May 2012, 02:05

### Mark Myatt

Consultant Epidemiologist

Frequent user

24 May 2012, 10:19

```
Age-group n %
------------ ----- -------
6-17 months 121 24.6%
18-29 months 108 22.0%
30-41 months 106 21.6%
42-53 months 110 22.4%
53-59 months 46 9.4%
------------ --- -------
491 100.0%
```

showing the general pattern.
If you see this pattern then you can take you 11% and say that there are about 2.2% in each year. If you don't see this pattern then you could (if you are sure of the survey) use the survey data to distribute your population accordingly.
### Mark Myatt

Consultant Epidemiologist

Frequent user

24 May 2012, 11:34

### Anonymous 3692

team leader

Normal user

6 May 2016, 16:00

in pakistan first we do survey of under 5 year children and then calculate under 2 year children

by raking 32 percent of the 5 year children date

under 2 year children= 0.32 *5500

if a district has population of 5500 under fiver children then under 2 year children will be 1760

### Muhammad Asim

Normal user

4 Jan 2017, 17:31

Dear fellows,

I am going to conduct PhD research entitled "Mother child Undernutrition in Socio Cultural context: A study in Sahiwal Division, Punjab Pakistan. My target population is mothers having less than two years children. I am worrying about to calculate the target population (Less than two years children) in a given population to develop the sample framework. Please help me to know about the less than two years children in a given population. Looking for help...

### Mark Myatt

Consultant Epidemiologist

Frequent user

5 Jan 2017, 08:33

Population data from the United States Census Bureau’s International Data Base might help. National census data will usually offer finer detail (e.g. down to small areas). The issue with Pakistan is that the most recent census was in 1998 and much can change on 17 years. Local authorities may have more recent data. Demographic data from DHS might be useful. The last DHS was in 2012.

I hope this is of some use.

### Nitush Fikir

Nutritionist

Normal user

5 Jan 2017, 11:57

Hi Mark,

Thank you for your continuous technical guidance and help through the years. I couldn’t properly grasp the idea on your guidance on 24 May 2012, 10:19 above ("If you see this pattern then you can take you 11% and say that there are about 2.2% in each year").

How can I calculate the 11% and 2.2% per group from the example you have given. I would be thankful if you could further explain it.

Thanks for your usual help

### Mark Myatt

Consultant Epidemiologist

Frequent user

6 Jan 2017, 10:50

Chantal gave the example of the under five years population being 11%. The 11% is given. If we assume that the age distribution is pretty uniform them we would expect there to be about 11 / 5 = 2.2% of the under five years population to be in each single year group.

Given the example table from SMART data above:

Age-group n % ------------ ----- ------- 6-17 months 121 24.6% 18-29 months 108 22.0% 30-41 months 106 21.6% 42-53 months 110 22.4% 53-59 months 46 9.4% ------------ --- ------- 491 100.0%

We can see that the assumption of uniformity is reasonable. We would expect slightly larger numbers in the younger groups and you may want to account for that. Here we see 24.6% in the younger children so we might use 11 * 0.25 = 2.75% in year 1 and the same in year 2.

These are informed guessed. Try to use the best information that you can get and as many sources as you can get to make the best guess you can.

I hope this helps.

### Nitsuh Fikir

Public Health

Normal user

6 Jan 2017, 14:33

Thank you very much Mark. It's clear and very helpful.

### Idris Muhammad Abdullahi

NSLO at CDC NSTOP AFENET

Normal user

12 Oct 2017, 02:06

Please, I am looking for help.

My question is how many % should be used to calculate 24 months to 59 months of age out of total population.

In Nigeria we use 4% for 0-11mnth, 5% for pregnant women, while 17% for 9-59 months and 22% for WCBA.

So please help me with 24-59 month % in the total population?

### Andrew Hall

Nuritionist and parasitologist

Normal user

12 Oct 2017, 17:33

You could look at the data for Nigeria on the United States Census International Database, here, which can present the population of any country in the world by year of age for any year between 1950 and 2050.

You select the Report: Population by year of age; select Nigeria; select the year 2017, or up to 25 years by holding down the control key when you click on years; and select Show individual country data only. One of the tabs has a link to the methods used.

Even if the total population is different from the official government population, you could calculate the percentage in each age group on the assumption that the underlying distribution by age is close to the actual distribution. So for children in the running years 2, 3 and 4 in Nigeria the estimated population in 2017 is 9.3% of the total.

### ngakani nyongolo delvaux

nutritionniste. msf Hollande

Normal user

26 Oct 2017, 16:23

24-59 months age group makes up 19% in a population

### Muhammad Asim

Normal user

26 Oct 2017, 16:33

Dear Group members,

I am conducting PhD research on Child Malnutrition in Pakistan. My target population 6-23 Months children. I am very confuse, how to determine the sampling framework of my study. How can I determine, how many children are in age group of 6-23 months of age. Please help me in This regard.

Looking for answer and reply

Regards

Muhammad Asim

Pakistan

### Dr. Hussein M. Aden

Consultant

Normal user

26 Oct 2017, 18:19

Dear Experts,

Kindly let me know how to calculate in <2, and <5 in Somalia.

Thanks,

Hussein

### Edward K

Normal user

26 Oct 2017, 18:26

Hi Asim,

This question can be objectively answered by disaggregating recent census data in Pakistan by age group. Another method especially at sub national level is to obtain anthropometry data from population level surveys eg Nutrition SMART surveys and compute % of 6-23 months age group from 6-59 months.

### Muhammad Asim

Normal user

27 Oct 2017, 18:56

Thank you Dr. Edward for suggestions to estimate the population 6-23 months children

### ljeanneEH

Environmental Health Scientist

Normal user

28 Mar 2019, 15:46

Dear Colleagues,

I am trying to estimate the population of children under 2 years in a specific district of Peru. I have the birth rates and <4 yr population from several years. The current replies are very helpful, but I would like to know if these are standard, published approaches. Could you provide references (especially for the first reply) so that I can present and/or publish my estimations based on these approaches?

Many thanks.

### Anonymous 39006

Ghana Health Service

Normal user

12 Feb 2020, 21:55

I would be grateful if I can be assisted with the percentage of the entire population that gives children 21-59 months?

### Mark Myatt

Consultant Epidemiologist

Frequent user

13 Feb 2020, 12:12

That "21-59" months seems very specific. Why would you need this?

It is a rough an ready approach ... we can estimate the proportion of the under five years population that is aged between 21 and 59 months as:

(59 - 21) / 59 = 0.6440678 = 64.41%

You can apply this "correction" to the population aged under five years. If (e.g.) 17.3% of the population is aged under five years then we might expect about:

0.173 * 0.6440678 = 0.1114237 = 11.14%

to be aged between 21 and 59 months.

I hope this is of some use.

### Hamdard

Social worker

Normal user

20 Jul 2020, 07:03

Under 5 year= 655

How will be total population

How will be under 15 year

How will be under 2 year

### Anonymous 40398

Student

Normal user

25 Aug 2020, 14:46

Can anyone tell what are different population percentages that we assume to calculate under 1 year, under 5 year and under 15 year?

Reference links will help too.

Cheers!!

### Noreen M Mucha

Independent Consultant

Normal user

25 Aug 2020, 18:19

It is recommended to use the latest Census numbers which you can obtain from the statistics department. UNICEF or UN agencies also generally estimate.

### Mark Myatt

Consultant Epidemiologist

Frequent user

26 Aug 2020, 10:22

I think Noreen is right. There is no simple one-size-fits-all formulae that can be applied. The best estimates will be from census numbers. These are often regularly updated to account for fertility and mortality and population movements. Data are available from a number of sources. I have found the US Census Bureau's Internation Database to be useful.

If you have any problem posting a response, please contact the moderator at post@en-net.org.