Wednesday, July 17, 2013

Shonky birth statistics

An article appears in Stuff today:


 The number of children born outside wedlock is fast approaching the number born to married parents, and debate is raging over what it will mean for society.

Debate raging? I hadn't noticed. Examples of "raging debates" are  asset sales or same-sex marriage.

Anyway at the end of the article these statistics appear:

BIRTH STATISTICS
Births to married and unmarried couples
Fifty years ago, 8 per cent of births were ex-nuptial.
Last year, 48 per cent of births were ex-nuptial.
About 80 per cent of Maori women give birth ex-nuptially.
About 50 per cent of Pakeha women give birth ex-nuptially.
About 13 per cent of Asian women give birth ex-nuptially.
15 in every 1000 ex-nuptial births is to a woman aged between 40 and 44.
25 in every 1000 ex-nuptial births occur to women aged between 15 and 19.

The third, fourth and fifth are wrong. 80 percent of births to Maori women are ex-nuptial. That's not the same as "80 percent of Maori women give birth ex-nuptially." Not all women give birth.

The last two are really shonky. I'll use the last because I'm familiar with the data.

25 in every 1000 ex-nuptial births is 2.5 percent. Yet in 2012 15-19 year-olds had 6.2 percent of all births (nuptial and ex-nuptial). So a large majority of teenagers are giving birth inside marriage? No.

The current teenage birth rate is around 24.89 per 1000 15-19 females. 23.8 per 1000 unmarried 15-19 year-old females gave birth ex-nuptially.

There were 3,768 births to teenagers last year; that means 3,602 were ex-nuptial which represents 12.4 percent of the 29,159 ex-nuptial births.

Expressed as per above then:

124 in every 1000 ex-nuptial births occur to women aged between 15 and 19



5 comments:

Kokila Patel said...

This was a story in the Daily Telegraph, so it looks like someone has decided to look if the same applies in NZ.

Anonymous said...

Lindsay, I'm afraid you haven't yet got the figures right. In the case of the ethnic statistics, the figures refer to the ethnicity of the child, not the woman - eg, 80 percent of Maori newborn children, 50 percent of European/Other newborns and 13 percent of Asian newborns were born ex-nuptially [in the three-year period centred on census year 2006, I expect].

Secondly, your recalculation of the last two statistics is quite wrong (and equally nonsensical, if you think about it). I've checked the age-specific ex-nuptial fertility rates on Statistics New Zealand's online Infoshare database and found comparable statistics to those misdescribed in the Stuff article. In 2011, the age-specific ex-nuptial birth rate (ex-nuptial births per 1,000 estimated not-married women in each age group) was 25 births per 1,000 for not-married women aged 15-19 years and 16 births per 1,000 for not-married women aged 40-44 years (24 and 17 per 1,000, respectively, in 2012). In the case of 15-19 year olds, this is almost exactly the same as the overall fertility rate, since few are married at that age.

Lastly, you missed a glaring error near the beginning of the article:

"The figures showed there were just 1000 more children born to married parents than unmarried in the March quarter.

"That is well down on 2000, when children born to unwed parents were outnumbered by more than SEVEN TO ONE." [my emphasis]

In fact, the ratio of nuptial to ex-nuptial births was 1.3 in 2000, down from 3.8 in 1980. The last time the ratio was anything like seven to one was in 1970, when it was 6.7. [SNZ annual data for March years.] The ratio has been 1.1 since 2007, it is not really news.

No matter, TVNZ grabbed the Stuff article and ran it complete with all the errors!

Lindsay Mitchell said...

"Lindsay, I'm afraid you haven't yet got the figures right. In the case of the ethnic statistics, the figures refer to the ethnicity of the child, not the woman - eg, 80 percent of Maori newborn children, 50 percent of European/Other newborns and 13 percent of Asian newborns were born ex-nuptially [in the three-year period centred on census year 2006, I expect]."

Birth statistics are based on mother's primary ethnicity. Go back to Infoshare and check that out. You can select "Female" or "Maori Female". That descriptor defines the mother, not the child. These figures have nothing to do with the Census.

"Secondly, your recalculation of the last two statistics is quite wrong (and equally nonsensical, if you think about it). I've checked the age-specific ex-nuptial fertility rates on Statistics New Zealand's online Infoshare database and found comparable statistics to those misdescribed in the Stuff article. In 2011, the age-specific ex-nuptial birth rate (ex-nuptial births per 1,000 estimated not-married women in each age group) was 25 births per 1,000 for not-married women aged 15-19 years and 16 births per 1,000 for not-married women aged 40-44 years (24 and 17 per 1,000, respectively, in 2012). In the case of 15-19 year olds, this is almost exactly the same as the overall fertility rate, since few are married at that age."

Yes. That is exactly what I am saying. What's your point of difference? And why is my re-calculation "equally non-sensical"?

23.8 births per 1000 unmarried 15-19 year-olds does not equate to 25 out of 1000 ex-nuptial births being to 15-19 year-olds.

You are correct that I missed the first "glaring error".

Anonymous said...

On the first matter, we are both right. The birth statistics I downloaded are titled Live births by nuptiality, Maori and Total Population. The headings relating to Maori are, respectively, Exnuptial - Maori child; Nuptial - Maori child; Total - Maori child.
Yes, there are also birth statistics based on the ethnicity of the birth mother. It is not clear which stats the writers are using. I accept that you made a good point about their error in referring to all women (not just women giving birth).

Population data from the census is used as the denominator for ethnic fertility rates, because there are no annual ethnic population estimates for populations other than Maori. Also, the census is the only reliable source of marital status data by ethnic group, from which to calculate nuptial and ex-nuptial birth rates.

On a more careful read of your last point, I can now see your reasoning. I thought the main issue was that the writers had misdescribed the exnuptial fertility rate for 15-19 year olds. A different emphasis, I guess.

I'm done!

Anonymous said...

I think the issue here is to look at outcomes using these figures. For example, Maori children appear far less likely to be born into the home of a (stable?) married couple. So, how do Maori youth turn out in other comparisons and is there a negative pattern across generations?

That consideration will allow a debate about some of the supposed benefits of marriage.

3:16