Hacker News Reader

Comment by ngc248 | Hacker News Reader

BS headlines everywhere.

Straight from the study: https://cgdev.org/publication/three-new-estimates-indias-all...

"Estimating COVID-deaths with statistical confidence may prove elusive. But all estimates suggest that the death toll from the pandemic is likely to be an order of magnitude greater than the official count of 400,000;"

But no one is adding this disclaimer, everyone is just parroting the 4 million number.

Another thing to ponder, if this amount of people had died, do foreigners really think Indian people would have sat silently? There may be underreporting or missing stats, but I doubt if it is a 10x difference.

View on HN

> an order of magnitude greater than the official count of 400,000

Wikipedia:

> An order of magnitude is an approximation of the logarithm of a value relative to some contextually understood reference value, usually ten...

An order of magnitude seems to usually mean something at least multiplied by 10. 400,000 times 10 is 4 million.

An order-of-magnitude means "about 10."

In contexts, it's a way of expressing uncertainty. If I say "About a hundred, but I might be off by an order-of-magnitude," that means "probably between ten and a thousand, perhaps even as low as 5 or as high as 2000"

In the same context, if I say "I'm might be low by an order-of-magnitude," that means "between 100 and 2000"

Bringing that back to a number is nonsense.

Conversations aren't the same as written articles. In a study or article you expect a proper and more precise use of language, hence the 4 million.

The entire point of language like "order-of-magnitude" in an article is to convey uncertainty, and that you're sure at the level of the log of the number.

Error bars are important.

A different poster pointed out other language in the article which DID estimate 4 million deaths more precisely, but that level of imprecision is exactly what scientists and engineers mean by "order-of-magnitude."

It's common for press to take a result with huge error bars, and to condense it to a number.

> if this amount of people had died, do foreigners really think Indian people would have sat silently

Yes? I mean, the UK has had 129k deaths and the government's approval rating remains leading the opinion polls of voting attention.

The statistical deaths of strangers just don't move the needle in politics.

In India, you have the opposition party and then lots of regional parties which are in power in state governments. These parties would never sit silent if the numbers were like this, since they could have used it politically, or people from all parties should be in on the game, which I can't believe.

They don't sit silently at all, but Modi's phenomenal approval ratings barelly took a hit for things which would've been a political suicide to any politician in the West.

Modi's approval rating has taken quite a hit recently though. Your point still stands. Even after the hit in ratings those would be the envy of many.

> do foreigners really think Indian people would have sat silently?

Most of my Indian relatives and family (who all live in India) find this 10x under-reporting very plausible.

An order of magnitude above 400000 is 4 million.

3 million and 5 million are also an order-of-magnitude above 400,000, as are quite a few other numbers.

OP has conveniently left out a few lines just before their quoted bit:

> We report excess mortality estimates from three different data sources from the pandemic’s start through June 2021. First, extrapolation of state-level civil registration from seven states suggests 3.4 million excess deaths. Second, applying international estimates of age-specific infection fatality rates (IFR) to Indian seroprevalence data implies a higher toll of around 4 million. Third, our analysis of the Consumer Pyramid Household Survey, a longitudinal panel of over 800,000 individuals across all states, yields an estimate of 4.9 million excess deaths.

If their (or your) complaint is with these actual methods of estimation, then pick on those, but the assumption that the study authors just want to Jedi mind-trick you into 10x'ing the official Covid deaths number is flawed.

Reply on news.ycombinator.com