January 31, 2006

One less gestation

Figures say a lot. There's one in particular that I ran across some years ago and then lost in the confusion on my book shelves. The little publication it is in (Radical Journal of Health, April-June 1995) came to light a day or two ago; unaccountably and mysteriously, it has found its way to our dining table.

What figure am I talking about? There's a table at the back titled "Mean Age at Marriage, 1981". It lists the mean age at which Indian women got married, rural and urban separately, and then breaks those numbers down by level of education. The figures are from the 1981 Census, yes, but hold on, there's something interesting in there.

First, what was the mean age at marriage for all women? Rural: 16.5 (that's 16 years and 6 months) Urban: 17.6 (17 years and 7 months).

Stunningly young, of course. Think of the number of women -- girls, I should say -- who were younger than that when they got married, if the mean was 16.5 and 17.6.

But there's something more interesting still.

For illiterate women, the mean age at marriage was: rural, 16.3; urban 16.8.

But what if the woman had been to primary school? Those numbers jump to: rural, 17.1; urban, 17.4.

The difference: rural, 0.8 years (9+ months); urban, 0.6 years (7+ months).

And it was that 9+ months that struck me. Why? You guessed it: 9 months in pregnancy.

So what these figures are saying is this -- put a rural Indian woman through primary school, and that immediately means she will likely be 9 months older when she marries. That means one less potential baby she will give birth to, just by virtue of being that many months older when she marries. (And the urban jump of 7+ months is close enough to 9 that we can say the same).

Apologies for putting this so bluntly and baldly. But this one figure has always put in perspective for me the value of education, and especially educating girls.

31 comments:

Anoop Sundaram said...

pretty nice post!! and a real neat emphasis at the end.......

Ashish Gupta said...

While I don't like to start talking like liberatarians, but as I am preparing for my GMAT, I couldn't help but notice some "logical fallacy" in the argument!

Educated women marry, on average, 9 months later, doesn't imply that if you educate one, she will be 9 months older when she marries.

Also, it definitely does not mean one less baby. Your argument is based on assumption that people have following policy on babies: produce as many as you can until the woman crosses age=X mark, then stop. Which of course, is not likely to be true. What is more plausible policy is: produce as many as you can until number of babies=N. Which simply means that last baby will be born 9 months later than otherwise, nothing to check the population. Only exception being that someone keeps on having babies until menopause (that would be 20+ babies, I guess), in which case 9 month later marriage means one less baby. But this scenario is not typical, is it?

Sorry to pick on hair's skin!

Dilip D'Souza said...

Thanks Anoop.

Ashish, consider it this way. Suppose you want to reduce the growth of a population. One way is to get people to have fewer babies.

But another way is to get women to have babies later. This works for two reasons: one, the later a woman has a baby, the less babies she is likely to have. Two, if succeeding generations also have babies later, that will steadily slow the growth of the population.

So what do these figures show? That if a woman goes to primary school, on average she gets married (and thus starts having babies) 9 months later than if she didn't get any education at all. That's 9 months less, of the time she is fertile, that she can bear a child.

I can't see the logical fallacy.

Tanuj said...

to me, what pops out is that the mean ages are less than the legal permissible age for women to get married in India! it is obvious that no one cares (or knows?) about the law.

as for logic, etc., i tend to agree with ashishg. dilip, your logic assumes that women continuously reproduce until their bodies give up (say 20+ babies per woman), which is obviously not the case, even in india. if the average number of babies born per woman is 3 - 4 (my estimate), 9 months (taken away from 25+ possible years of bearing babies) is hardly likely to impact that average.

sure, education has many benefits, but this is probably not one of them.

Niket said...

Can't agree more Dilip... though the real impact may be because some education (even minimal) would result in the woman getting a greater say in making decision about child birth.

It would be interesting to compare the current stats. How much have we improved in the last 20 years.

Melody said...

Hey there! Remembered your url & decided to pop right in! Great place you have here... especially impressed to know you're a writer **Mel bows down**

Interesting post too (though I'm not sure about the 9 months thingie)... Will be interesting to see what 2005 figures say.

Anyway, really popped in only to say Hi, so will shove off now...

Dilip D'Souza said...

Tanuj: 9 months (taken away from 25+ possible years of bearing babies) is hardly likely to impact that average.

Here's something of an analogy. Let's say you run your washing machine every day, washing 25 pieces of clothing each time. One day, you decide to pick out one shirt every day from the load and wash it by hand.

What's one shirt less from 25, you think. But consider that in a year, you will have washed 365 shirts by hand. The equivalent of over 14 loads.

That's the way to look at this. I'm not saying the primary-educated woman will necessarily have one less baby. I'm saying, a primary education shaves 9 months off a woman's likely child-bearing years. Aggregate over millions of women, and you're going to have less babies.

Or put it this way: if we got figures for the average number of babies born to illiterate women, and figures for the average number of babies born to women with a primary education, I will bet there's a difference. This 9 months is a pointer to that difference.

Melody, I hope that explains for you as well. Good to see you! Soon again, I hope.

Quizman said...

I agree with the conclusion, but as Ashish G has indicated, there is a logical flaw. This is not a controlled experiment, where all the other factors are supressed. Hence, you cannot rationally conclude that primary education alone had a causative effect.

It could, for example, be the already existing progressive attitudes of the parents (for educating the girl child) and husband (for marrying one), which could be causative factors.

That said, there is no doubt that primary education is beneficial for a variety of reasons.

Anonymous said...

Only thing I can get out of this is that the quality of sex must be terribly high in India. If couples get hitched at 16 & 17, given the low divorce rate in your country, they have some seriously prime sextime in their marital life - atleast 20 years of mindblowingly good sex, 10 years average sex & last few years of erectile dysfunction, handjobs, porno induced semen outflow.
In USA for example, couples marry in their 30s, have only 5+ years of average sex before moving on to porno, even that assuming no divorces & prenups.

Regarding gestation - you have free condoms supplied by GoI, you have family planning clinics, in time junta'll learn how to use them efficiently. So count your blessings and don't despair - go have that sex right now.

Purushottam said...

A few points:

1. from The Department of Women and Child Development, Ministry of Human Resource Development, Government of India ...
http://nrcw.nic.in/index2.asp?sublinkid=280

A small table at the beginning reports the effective age of marriage for women in 1981 to be 18.3 years, which is different from what Dilip found in the Journal (16.5 for rural and 17.6 for urban) ... not sure if/why there is a discrepancy or a case of depends on who-reports-the-data.

2. The mean in 1991 and 1997 are reported as 19.5 years (surprisingly exactly same!). The Child Marriage Restraint Act, 1976, which raised the age of marriage for girls from 15 to 18 years, is also mentioned as a possible reason and may have certainly influenced those numbers.

3. Dilip, in the nos. you mentioned,
are there any more breakups, other than rural, urban, illiterate and primary educated?

Purushottam said...

also check ...
http://www.censusindia.net/results/C_Series/data_highlights_c6.pdf

1. Page 2 reports the Mean Age of Marriage for 1991-2001 and has a little more detailed breakup of catergories. The number reported in the link I posted in my previous comment does not seem to match for the year 1991 reported in this report. (19.5 yrs in previous and 17.7 in this one)

2. From page 3, atleast 3% of all women are below the age of 10 yrs at marriage (3.6%: illiterate and 2%: literate)
And what would the category 'Literate but below primary' mean?
... can write name? sign? ... any idea?

tejal said...

i have read previous posts on your blog Dilip and have found the articles as well as the comments to be quite sensitive and sensible.
but i couldnt help the feeling that this post (more the comments than the post) lack that sensitivity. Everyone has spoken about womens education as a tool to end/slow down population growth and equated women to baby making machines. I think the more important issues here are the fact that there are so many girls married off before they turn 18 and this poses a very big threat to their health, reproductive and otherwise. the "logical fallacy" here is that it is assumed that women are inherently capable of producing babies for the 20 years between 15 and 35. I think the most terrifying and "striking" part of the statistic is the fact that there is this enormous attrocity (physical and mental) being meted out against women(literate or illiterate) and not the effect that it will have on how many babies an educated girl will have or not have. I am sorry if I am being too blunt.

Quizman said...

Tejal,

To be fair to Dilip and the folks who wrote the comments, I'm sure it goes without saying that all of them would agree with your statement on the statistic. Yeah, okay. So what? Nothing new in the 'atrocity' being pointed out.

If I may speak for Dilip, I think he was trying to find a viable solution to that problem. And he thinks that the data shows that primary education is a key factor. And others are pointing out that there could be other ways in which that data could be read.

That is where the challenge is! Merely accepting that atrocity exists, is a good step, but not good enough.

Dilip D'Souza said...

This is not a controlled experiment, where all the other factors are supressed. Hence, you cannot rationally conclude that primary education alone had a causative effect.

But it is not meant to be a controlled experiment. Census figures never are. This is just a way of looking at data, and this data suggests that a primary education has an effect on when a woman gets married. That impression gets strengthened with the rest of the data in that table: the more educated women get, the later they get married on average. The other suggestion from the data -- women who live in urban areas get married later on average than women in rural areas -- would this be as contentious as the other suggestions seems to have been?

These are the other categories in the table:
Middle School: (r 17.8 u 18.1). Matriculates (r 19.3 u 19.8). Graduates (r 21.5 u 21.9). (Puru, hope this answers your point #3).

More in next comment.

Dilip D'Souza said...

Puru, I'm surprised by that NRCW table. All I can say is, it looks wrong. you pointed out one apparent flaw, the identical 1991 and 1997 figures (19.5). Another one that strikes me is that the male figure for 1981 is 23.3, which is higher than the male figure for 2001 (22.6) from the censusindia.net document you pointed us to.

The latter document seems to be more in line with the table I'm looking at: I can see the 1981 numbers fitting into the 1991+ numbers it cites.

As for literate but below primary, that's a good one! What on earth is that supposed to mean?

Of course this censusindia document also bears out the impression that the more education a woman gets, the later she gets married.

Tejal, believe me I see your point. This is why I ended my post saying "Apologies for putting this so bluntly and baldly". I don't mean at all to speak of women as baby machines. My only point is that education has a bearing on when women get married, and therefore on the choices they can make.

Finally, thanks for all the discussion, and I hope it hasn't ended.

I'd like to say that I actually put up this post for one more reason. On the new blog I'm part of, How the Other Half Lives, there have been a few comments suggesting that we should "use statistics to make arguments", offer "solid, statistical reporting", and the like. So I made this post regarding statistics to make a small point: even solid statistics can be looked at differently.

Tanuj said...

dilip,

theoretically speaking, i agree with your hypothesis. if you take away 2-3% of the reproductive age of 60% of all indian women, something might happen to fertility rates.

but let's look at some real numbers: from various sources you will find that the average fertility rate for women in india is anywhere between three and four babies. (several results via google)

An example:
"TFR or the estimated number of children an average woman will bear, as reported by NFHS1 was 3.4 (3.7 for rural areas and 2.7 children for urban areas). Women with at least high school education are estimated to have only 2.2 children per woman which is close to the number of 2.1 as said to be desired for stabilising the population. Illiterate women are estimated to have 4 children."

source (for whatever it is worth): http://www.indianngos.com/issue/population/resources/articles3.htm

mathematically speaking, this could mean (for illiterate women), either giving birth to 4 kids on the trot (1 every 9 months) and calling it a day, or 1 baby every eight years (or multiple scenarios between these). i imagine that in either scenario shaving 9 months off the reproductive age of a woman is unlikely to make any realistic difference to average number of babies born.

Dilip D'Souza said...

Tanuj, I don't get it: you yourself are quoting me statistics that show the TFR goes down with education. (4 for illiterate, 3.4 average, 2.2 educated). That's precisely the point I'm making with this post.

With education, women get married later on average, and therefore have less babies. If you accept the "have less babies", I'm unable to understand why you won't accept the "get married later". Or I'm unable to understand why you won't accept that "get married later" means "have less babies" (as a broad rule).

The 9 months is just incidental -- it really was just the coincidence with gestation that got me thinking about this.

Anyway, I hope you see what I meant about statistics.

Vivek Kumar said...

The title was rather unfortunate because it made one digress from the point. But I get the point you are trying to make - more education for women, lower fertility rates.

I don't think Dilip was trying to make an air-tight statistical argument. Social issues can hardly ever be argued based entirely on statistics. Statistical conclusions have many underlying assumptions about the data set that do not hold true for real life data.

Id it is said...

Numbers don't lie! However, I observe that marriage seems to be a necessary precursor to inception. That I find very reassuring. We in the USA have 13 and 14 year olds, obviously not married and not considering it either, are mothers already. The High School where I teach has a free child care unit to look after the numerous babies of 13 to 16 year olds. This service is provided so that the girls do not drop out of school! However, free childcare has resulted, not in academic advancement for the young mothers, but in steady increase of the number of babies being serviced.
The above is another side to the picture that you etched so vividly with numbers.
Great post.

Tanuj said...

d,

sure, educated women have a lower fertility rate. the question really is what's causing that lower fertility rate.

there are two possible reasons:
1. logistical: time spent on education takes away a few months out of a potential 25 year reproductive age period (yes, by marrying later - that was hardly the issue). hence, fewer kids per person.
2. behavioral change: education impacts the way women think, behave, react. among other things, they also choose not to have more than x babies.

to me, your post appears to advocate reason 1. the point of my prev comment was that reason 1 is most likely insignificant.

Neela said...

Dilip: Nice argument. As one of the commenters, I asked for a mix of stats and anecdotes. And so am happy to see this.

Only one thing worries me, as it did quizman - the old correlation does not imply causality argument. You have shown the first condition of causation: i..e there is a statistically significant relationship b/w the numbers. What would you suggest for condition 2 and 3 i..e (2) the causal variable must precede the dependent variable and (3) there must be no other factors or potential third variable?

Quizman, your point is valid. However, one could appreciate that conducting the "perfect experiment" in this case might not be possible because of ethical reasons (though as my professor argued: why not? when medical experiments are being carried out every day, why not stuff like this?).

Thanks

Neela

PS: Dilip, of course solid statistics can be looked at differently - no one denies that. The point is that in case of stats, one can, with enough motivation and expertise logically argue out a conclusion because ultimately numbers have to add up. whereas with anecdotes it becomes a "your story against mine" which leads to emotional arguments along the lines of "are you doubting my word/me as a person/my honesty/integrity etc ?" whcih can easily deflect the argument at hand.

Dilip D'Souza said...

And now here...

Vivek, I chose that title very deliberately, to emphasize the 9-month connection that got me thinking about these figures.

Id, I don't know that marriage is a necessary precursor. But I think there is a broad link between getting married later and having fewer babies.

Tanuj, it is not logistical at all. After all, why would the time you spent in school between the ages of 5 and 10 (say) shift the age you get married at by 9 months? The way I look at it, It is probably the awareness and behaviour changes that even a primary education brings that results in that 9 month shift.

That said, Neela, I am not advocating a causal relationship. I'm saying, if you look at the figures, it's reasonable to conclude that if you have more education, you get married later. That's all this post says, though I know I was rather more longwinded when I wrote it!

Nope, I don't believe simply using facts results in logical conclusions that all sides agree on. That's just the point: people look at stats differently and draw different conclusions. (Did Shaw speak of lies, damned lies and statistics?). As for anecdotes resulting in emotional arguments and accusations of dishonesty, in my experience those are possible features of any argument. Not just those laced with anecdotes.

Quizman said...

Neela,

You wrote Quizman, your point is valid. However, one could appreciate that conducting the "perfect experiment" in this case might not be possible because of ethical reasons.

True, it is the same argument that my former prof made about the lack of good statistical data on epidurals!

But, regardless of whether there is a causal factor, one could argue that the sheer act of educating the girl child could be an indicator of a (relatively) progressive attitude in the household. Whether the girl child herself causes it or it is caused by other factors may, in the long run not be material to the fact.

Educate them, I say!

You can begin by contributing to the group I've run for! :-)

Anirudh said...

The amount of talk about how to educate more people (which is important) seems to be far greater than the more important problem of how to educate people. As this post shows, all we're doing is keeping a girl in school for 9 months longer, not "educating" her.

Dilip D'Souza said...

As this post shows, all we're doing is keeping a girl in school for 9 months longer...

Wrong, Anirudh! This post says something quite different: you give a girl a primary education, and on average she gets married 9 months later than if she was illiterate.

Ashish Gupta said...

I see your point Dilip, which I guess is lost in the discussion here. Eudcation does not mean one less baby (my argument still holds) but it increases interval between two generations by 9 months.

So even if women have as many babies as they had earlier (N), then rate of growth will increase from be N over 17 years (assuming thats mean marriage age, I forgot your numbers) to N over 17.75 years. In long run, that is definitely beneficial.

Ashish Gupta said...

sorry..rate of growth will *decrease*, i meant above

Sunil said...

Good post .
Could you clarify 2 points for me ?
You say "Or put it this way: if we got figures for the average number of babies born to illiterate women, and figures for the average number of babies born to women with a primary education, I will bet there's a difference. This 9 months is a pointer to that difference.'
Can you point to figures that can support your hunch ?
You also say , Two, if succeeding generations also have babies later, that will steadily slow the growth of the population.
You are initially comparing literate vs illeterate women, where does the generation factor come in ? Are second generation literate women likely to marry later than their first generation literate mothers.

Anirudh said...

D-Mama,

I wasn't talking about the point you were trying to make. All I'm saying is that we're all very concerned about literacy, marks etc. Nobody cares about education.

Dilip D'Souza said...

Sunil, Can you point to figures that can support your hunch?

From Tanuj's comment above, this:
"TFR or the estimated number of children an average woman will bear, as reported by NFHS1 was 3.4 (3.7 for rural areas and 2.7 children for urban areas). Women with at least high school education are estimated to have only 2.2 children per woman which is close to the number of 2.1 as said to be desired for stabilising the population. Illiterate women are estimated to have 4 children."

Are second generation literate women likely to marry later than their first generation literate mothers.

I have no idea.

Look at this way. Woman W is born in 1950. At 20 years old, in 1970, she has a baby daughter X. When X gets to 20, in 1990, she has a baby daughter Y. In turn, Y has baby daughter Z when she reaches 20, in 2010.

In 60 years, you have these four women.

Contrast to this other case. Woman W1 is born in 1950. At 30 years old, in 1980, she has a baby daughter X1. When X1 gets to 30, in 2010, she has a baby daughter Y1.

In 60 years, you have these three women.

That's what I mean about having babies later. In other words, these are two ways to slow a population's growth: people have fewer babies, or people have babies later.

Anonymous said...

What about the guys these women married? did they have anything to do with the babies ?.

After all chances are if "she" has a primary education, her parents would marry her to someone with a similar or higher level of education. Does he bring nothing to this mathematical/ logical equation everyone is trying to prove or disprove ?.