Getting the ABCs right | The Big Shift

The big shift in literacy in India is clear enough: rates of literacy have grown steadily over time, and gaps that once existed- between men and women, rural and urban, etc. – are converging, as my colleague Abhishek Waghmare found. Irrespective of how we measure it, the story is going to be the same.

Yet, how we measure it still matters.

Some measures matter because they affect what the trajectory of what we’re trying to measure could look like. That’s certainly the case when you study wage growth in India, and take a call on whether to chart the trend with or without the impact of inflation. In this case, the trajectory is unlikely to change, but we still need to better understand what precisely we’re measuring, and to improve measurement practices.

That’s what Abhishek decided to do in his work for us on how literacy in India is measured. First the basics: the globally accepted United Nations definition of literacy is being able to read and write a simple sentence in any language with understanding, and Indian surveys also use this definition. How do surveyors capture this? Typically by asking the head of the household about the educational status of all members of the household.

Except, this has some obvious drawbacks. The head of a household may not have an accurate sense of the literacy levels of all household members. Not just that – people who declare themselves to be literate may also be knowingly or unknowingly overstating their actual abilities.

Abhishek identified a novel experiment. Until 2005, India’s National Family Health Survey canvassed the literacy question by only asking adult men and women about their education levels. Since 2005, the NFHS has switched to administering an on-the-spot literacy test to those respondents who report having only studied up to the primary level.

What he found was quite remarkable. In the year that the NFHS used both methods, it found that actual literacy was six percentage points lower than respondents had reported.

The old NFHS method of only asking respondents had produced numbers that were quite similar to what India’s National Sample Surveys found by asking the head of the household. Over the years, the gap between the NFHS’s new methods and the NSS’s household head method has grown, and at the last count is at over ten percentage points.

As I said in the beginning, none of this will change the trajectory of the growth of literacy in India, which I can picture will reach 100% within my lifetime. That remains the big shift here. But for our statistics to show real integrity to lived realities, how we count matters.