One can treat data to yield the desired results, goes an old saying. But when the data in question concerns something sensitive, like religion (religious minorities, to be precise) the stakes are higher than usual.

Last month, on August 29, economist Surjit S. Bhalla wrote his Indian Express column on the recently released 2011 census data on religion—“Census, Christians, Conversions.” Bhalla began by arguing that the Muslim population share was rising because of a higher fertility rate, which is an indication of higher relative poverty of Muslims; he even made a case for affirmative action for Muslims. He then moved on to Christians. It was his inference here—“Christianity practises proselytisation in modern times”—that became controversial.

Bhalla’s argument is that the share of Christians in India’s total population has remained constant at 2.3% in the last three censuses. Christians have “the highest per capita consumption, the highest level of female education and the lowest fertility”, which, he argues, should have led to a reduction in the community’s population growth rate. Since it didn’t, conversions to Christianity explain the constant share, according to Bhalla.

Bhalla also put a number to conversions. Sikhs and Christians have more or less similar average income as well as fertility rates, but the annual growth rate of Sikh population in last two decades is 1.2% as against 1.9% for Christians: this increase in “Christian population of 3.7 million in 2011 is very likely due to conversions”.

One among the people irked by Bhalla’s arguments on conversion was Tony Joseph, former editor of Businessworld. Joseph wrote a scathing piece in Express on September 1, charging Bhalla to have “tortured his data to make it say what he wants to hear”.

Joseph alleged that Bhalla chose the years arbitrarily to fit his theory: “If you take 1971 as the starting point, instead of 1991, as Bhalla has done, you will find that the Christian population has declined from 2.6 to less than 2.3 per cent, a decrease of over 0.3 percentage points in 40 years.”

Joseph then comes to his main argument, the difference in sex ratio — the number of females to males in a population — between the Sikhs and Christians: In communities with a strong preference for sons, he argues, the sex ratio becomes highly skewed against women. “Among Christians, there are 1,023 females for every 1,000 males, while among Sikhs, there are only 903 females for every 1,000 males. It is a well-known and academically accepted fact that highly skewed sex ratios like this have a significant impact on population growth,” he writes. And so on…

Bhalla made his riposte on September 5. He was irritated with the support Joseph’s arguments had received on social media. “Writers seem to care more about the support they get on Twitter than the veracity, or logic, of their arguments,” he said, adding that Joseph had forsaken truth for momentary fame. This isn’t true, for it seems Joseph is new to Twitter and has considerably less fan following than Bhalla makes it appear.

Bhalla challenged Joseph to show even “one study that shows that, all other things being equal, a skewed sex ratio significantly affects population growth.” He added: the effect of fertility levels and adult sex ratios was analysed for 15 major states of India, for the period 1981-2011. This revealed, as expected, that fertility levels are strongly related to population growth—and that there is no relationship between sex ratios and population after controlling for fertility.”

Crucially, neither Bhalla nor Joseph mention the “academic studies” from which they derived their conclusion on whether sex ratio affects population growth, other things remaining same. For the same reason, some of the economists I called up refused to be quoted saying they need to take a look at the papers Bhalla and Joseph are referring to.

Laveesh Bhandari, economist and founder of Indicus analytics, on being explained the arguments, sided with Bhalla saying “fertility rate is surely one of the factors in population growth,” but added a caveat, “there are so many other important factors too — like migration,” which Bhalla doesn’t consider significant. It seems both Bhalla and Joseph have been making extraordinary claims without enough evidence. In a way this debate also shows the limitations of arguing complex issues in 700-800 words over op-ed pages.

On September 7, Rohini Prabha Pande, who researches gender and population, also took on Bhalla. Pande argued that it’s not fertility rate per se that determines population growth: “Even with a slightly lower fertility rate, the Christian population may have grown larger than that of Sikhs because of a larger population base to which this fertility rate is applied, rather than lakhs of conversions. If the age distribution among Christians favours the reproductive age groups, then even with lower fertility rates the population could grow.”

Pande went on to show that for the last two 10-year census periods — slightly technical stuff, sorry — “the (compounded) rate of growth of the Christian population across the two periods is slowing faster than the rate of growth of the Sikh population.”

So Bhalla is right when he says “Numbers don’t lie—people do”. Maybe people don’t lie consciously, but they certainly allow themselves to be trapped by biases which they fail to defend convincingly.