TECHNICAL NOTE 2 ADJUSTMENT FOR AGE HEAPING IN THE NORTHERN TERRITORY ABORIGINAL AND TORRES STRAIT ISLANDER POPULATION


INTRODUCTION

1 Age heaping describes the phenomenon of uneven population age distribution in Census or survey data. When age heaping occurs, data show systematic spikes on particular ages such as those ending in 0 or 5. This happens as a result of 'digit preference' or rounding when respondents are unsure of their age or the age of others they are reporting on behalf of. Age heaping is a relatively common demographic issue in developing countries and is generally not observed in the Australian population, except in the Northern Territory, where it is evident in Census counts and population estimates for the Aboriginal and Torres Strait Islander population.

2 The purpose of this technical note is to present the findings from an investigation undertaken by the ABS to determine the occurrence and the extent of age heaping in the Aboriginal and Torres Strait Islander population in the Northern Territory. This technical note also presents data for this population that have been adjusted to reduce the distributional impact of age heaping.


PRESENCE OF AGE HEAPING IN THE NORTHERN TERRITORY

3 The graph below shows the estimated Aboriginal and Torres Strait Islander population for the recent Census years in the Northern Territory. It shows some noticeable population spikes at ages ending in 0 and 5.

4 While the ABS has not traditionally applied an adjustment for age heaping, the increasing demand for data by single year of age for key reporting and population projection purposes has necessitated a methodological response to ensure the best possible estimates by age for the Northern Territory.


METHODOLOGY

5 Five methods were examined to adjust for age heaping in the Northern Territory Aboriginal and Torres Strait Islander population. While the Sprague method was considered the most robust, a brief discussion of each method is included below.


Moving averages (3 year and 5 year) method

6 The ‘Moving Averages’ method involves calculating the average of the population in the three or five consecutive ages, and regards this average as the population at the central age of the 3 or 5 consecutive ages. The advantages of the method are that it is relatively easy to calculate and is commonly used. However, the method fails to maintain consistency with state-level totals published elsewhere, which makes it ineffective in the consideration of population estimates.


Griffith Feeney method

7 The ‘Griffith Feeney’ method distributes the implausible number of people at ages ending in 0 and 5 to surrounding ages by linear interpolation using an established set of formulae (Feeney, 1979). The method has the relative advantage of maintaining state-level totals, but is resource-intensive and only gives the mid-point of the five-year age group. Age heaping is not effectively resolved at the single year of age.


Estimates based directly on births

8 This method applies the age structure of the annual series of births in the cohorts corresponding to the Census age to the five-year age group total of the original population. The method maintains state totals, but requires a consistently complete set of birth registrations across at least 70 years, which is not currently available by Indigenous status.


Estimates based on survivors of births

9 This method is similar to ‘Estimates based directly on births’, but incorporates deaths registrations to consider the impact of deaths on populations. The same advantages and disadvantages apply.


Sprague method

10 The 'Sprague' method involves applying a set of established multipliers (Judson & Popoff, 2004) to the population totals at the 5 year age group level. The method is easy to implement, uses available population data and maintains state level totals and age group sex ratios. For these reasons, the Sprague method was applied to adjust the age heaping in the Aboriginal and Torres Strait Islander population in the Northern Territory.


STEPS TO CALCULATE THE ADJUSTED NORTHERN TERRITORY ABORIGINAL AND TORRES STRAIT ISLANDER POPULATION

11 The Sprague method was applied as follows:

Step 1: The Sprague multipliers were applied to the five-year age group totals to estimate the population at each single-year age within that age group. This produced the adjusted total Aboriginal and Torres Strait Islander population by single year of age.
Step 2: The average proportion of Aboriginal and Torres Strait Islander males to the total Aboriginal and Torres Strait Islander population from the three most recent Censuses (2001, 2006 and 2011) was calculated for each age. This proportion was then applied to the single year of age totals calculated in Step 1 to obtain the male population at each age.
Step 3: The female population in each single year of age was then calculated by subtracting the adjusted male population (from Step 2) from the adjusted total Aboriginal and Torres Strait Islander population (from Step 1).
Step 4: The sex ratios (SRs) for each age were then compared with the average SRs for 2001, 2006 and 2011, to ensure that they were consistent.
Step 5: The adjusted population totals by five-year age groups for both male and female were compared with that of the unadjusted population. Very minor adjustments were made to the adjusted population to maintain age group totals and sex ratios for each age group.

DATA ADJUSTED FOR AGE HEAPING

12 The Northern Territory Aboriginal and Torres Strait Islander population estimates that have been adjusted for age heaping are provided in Datacube: Aboriginal and Torres Strait Islander population estimates (adjusted for age heaping), Northern Territory. This datacube is provided for reference purposes and includes original and adjusted estimates. To maintain the effective relationship between all population estimates (Aboriginal and Torres Strait Islander, non-Indigenous and total population) across all levels of geography, the adjusted data have not been used in the official series. Consultation will occur ahead of the 2016 rebasing cycle to determine whether these adjustments should be applied in the official series.

13 It is important to note that the ABS does not generally release Aboriginal and Torres Strait Islander population estimates by single year of age due to quality concerns. Users are advised to use the single year of age data provided here with caution. For most reporting and analysis purposes only, population estimates in five-year age groups is recommended.


DATA AVAILABLE ON REQUEST

14 The ABS will continue to use the unadjusted Aboriginal and Torres Strait Islander population estimates for publications and customised requests for consistency. Where users require age heaping-adjusted data, the method presented herein is available upon request.


REFERENCES

15 Judson, D. & Popoff, C. (2004) Selected General Methods in Siegel, J. & Swanson, D. (Eds), The Methods and Materials of Demography. Elsevier Academic Press: San Deigo.

16 Feeney, G., (1979). A Technique for Correcting Age Distributions for Heaping on Multiples of Five, Asian and Pacific Census Forum, 5(3).