Tuesday, 24 March 2015

Jonathan’s Lies, Damned Lies and Statistics

By Ibraheem Dooba

“A well-wrapped statistic,” said Darrell Huff, six decades ago, “is better than Hitler’s “big lie”; it misleads, yet it cannot be pinned on you.” So, to save the ordinary citizen from the mischievous use of statistics, Huff deemed it fit to teach regular folks how to do wicked statistics, hence his book, How to Lie with Statistics.

Here’s how Huff defended his project: “This book is a sort of primer in ways to use statistics to deceive. It may seem altogether too much like a manual for swindlers. [But] the crooks already know these tricks; honest men must learn them in self-defence.”
Unlike Huff, I wouldn’t (and can’t) cover all the angles people employ to deceive with statistics. Like him, however, it is a concern seeing how people abuse statistics and deploy them as a weapon to tell lies. So I propose to share three tips of how to spot a bad statistic.

What is the baseline?
Seeing that it hasn’t fulfilled any shining promise, our government distributes and publishes statistics and makes nothing of the baselines. But you can use that to catch them. For example, if I present a report that says HIV kills six million people per year in Lagos, you can ask one question to quickly discover I couldn’t be telling the truth: what is the total number of people who die yearly from all causes (from accidents, cancer, of natural causes, etc.) in Lagos? If the answer is less than six million people per year, you’d know that statistic is a lie. It’s the baseline technique that Obasanjo’s minister of agriculture, Adamu Bello, used to expose the lies of Jonathan’s minister of agriculture, Dr. Akinwumi Adesina.

Before then, the man and his ministry overwhelmed us with statistics, “We ended 40 years of fertiliser corruption in 90 days,” they informed us. Also, ‘the growth that the Jonathan administration has engendered in the agriculture sector is unprecedented’. It happened that this was a complete fabrication. 

Sharing data from the National Bureau of Statistics and the Central Bank of Nigeria, Adamu Bello argued that instead of growth, there has been a decline in the sector since Obasanjo left office. Statistics show that between 2003 and 2007, when Obasanjo left office, the growth in the agriculture sector averaged 6.85 percent per annum. The growth started declining immediately after he left office. And in 2013, under Jonathan, growth in the sector was 4.5 percent; the government failed its own target of 8.0 percent growth. 

Also, the ministry of agriculture was simply being mischievous when it said that it eliminated N26 billion annual corruption in fertiliser distribution. Using the same baseline technique, we can ask what the total budget for fertiliser per year was in those years. That was what Adamu Bello did. The answer may surprise you: N3 billion!

How can you steal 26 from 3? Bello said the total subsidy between 1999 and 2007 was under N25billion. “This can be verified from the Budget Office of the Federal Ministry of Finance,” he said.

Another baseline question, if N870 billion fertiliser subsidy was embezzled in 40 years, what was the total budget for agriculture for those years? 

Adamu Bello doubted if the entire budget for the ministry of agriculture was up to that amount. “To claim that there was subsidy of N870 billion spent on fertilisers, since the use of fertilisers was initially encouraged by the government about 40 years ago, is most unfair, as I doubt if the entire agricultural budgets for the whole period was up to that sum,” he said.

Dr. Adesina knew he could no longer win with statistics, so he changed tactics. He challenged Adamu Bello to tell the citizens his achievements in seven years as minister of agriculture.

“Bello never told Nigerians the truth about his actual performance in office. Let him give details commodity-by-commodity, and let him state where and how this impact was felt,” Adesina said through his assistant on media, Oyeleye Olukayode.
Under a better president, Adesina could be a good agriculture minister. However, what we get from the ministry now is too much noise and a good deal of initiatives and acronyms: phone for farmers, e-wallet, Growth Enhancement Scheme (GES), Youth Employment in Agriculture Programme (YEAP), Fund for Agricultural Finance in Nigeria (FAFIN), NIRSAL, and so forth. As a result, we’re confused and overwhelmed.

What is the method?
Imagine Daily Trust has four columnists – Mohammed Haruna, Adamu Adamu, Mahmud Jega and Ibraheem Dooba. Respectively, they make N3 million, N2 million, N1 million and N5 million per month. What is their average monthly pay? If we add all the money together and divide by 4, we get N2.75 million. From here we can draw a conclusion that a Daily Trust columnist belongs to the middle class.

What if, however, Dangote starts writing for Daily Trust? Now, the paper has five columnists, but Dangote also brings his income of, say, N100 million per month. If we add all the five monthly pays together and divide by the five writers, we get N22.2 million per month. Now, we can say a Daily Trust columnist is a rich man. Is this conclusion valid? Yes. But the method that led to the conclusion is flawed. When we have numbers that are too high or low (outliers) in a data set, statisticians advise us to take the median not the mean.

Now, if we do the calculation again using the median, we’ll get an entirely different result. To do that, we arrange the numbers in ascending order: N1m, N2m, N3m, N5m and N100m. Now we take the middle number, which is N3 million. Which is quite close to what we figured earlier before Dangote crashed the party – and a far cry from N22m per month.
So when a politician gives you an average that sounds suspicious, you can ask which average he chose. 

The Bureau of Statistics demonstrated recently the importance of appropriate method selection in statistical analysis. Sokoto State was the poorest state in Nigeria, it said. Niger State was the least poor.

In response, the governor of Sokoto State, Magartarda Wamako, stepped forward, breathing fire: “An agency that failed to conduct a proper survey on what we have been doing as government to mop up the rate of unemployment in the state should not attempt to misinform Nigerians and the rest of the world that Sokoto is the poorest state in Nigeria with 81.2 per cent rate. Such a rating is unthinkable, utterly incorrect and misinforming,” the Punch newspaper reported him as saying on January 14.
It took NBS almost three weeks to deny the reports that it rated Sokoto State as the poorest state in Nigeria. On January 25, ThisDay published a statement from Dr. Yemi Kale, the statistician general, addressed to the chambers of Dr. Alex Iziyon (SAN), the lawyer to the Sokoto State Government.
I was confused further when last month, at a retreat organised by the gubernatorial campaign council of Niger State APC, officials from the Governors’ Forum told us during a presentation that Niger is one of the poorest states in the country. The conclusions of the bureau and that of the Governors’ Forum, while measuring the same thing, are at the opposite ends of the spectrum.
When I asked the question, the Governors’ Forum said they are using a method completely different from the one the statistics bureau used. Instead of measuring income alone, they also measured other indicators of well-being such as health, education status, etc.
So if it sounds fishy, ask for the method.

Is it a lie?
When some conclusions are drawn from statistics, you immediately know that they are lies. A recent example of this is the recalculation of our GDP. President Jonathan and his mutual admiration society usually imply that the president grew our GDP to 500 billion dollars. No. He did not.
Recalculating is different from creating. Take for example, if the value of the products in my village’s weekly market had not been calculated for many years (even after its expansion) until now, and it’s found that the current value is far higher than what was calculated several years before, are we going to credit the incumbent chairman for this achievement even though there were many chairmen who contributed to the growth of the market before him?

Credit the incumbent, Jonathanians say: “Buhari’s shameful past is dwarfed by the achievements of Goodluck Jonathan. Under Jonathan, Nigeria has emerged as by far the largest economy in Africa with a GDP of $503 billion; nearly double the previous estimates,” Femi Aribisala wrote.

When Jideofor Adibe, a Daily Trust columnist, implied the same thing, I immediately threw away 70 percent of the respect I had for him; and I’m now sussing him out on the remaining 30 percent.
Actually, the Economist argued that the Nigerian economy is growing not because of Jonathan but in spite of him: “The single bright spot of his rule has been Nigeria’s economy, one of the world’s fastest-growing. Yet that is largely despite the government rather than because of it, and falling oil prices will temper the boom. The prosperity has not been broadly shared: under Mr Jonathan poverty has increased. Nigerians typically die eight years younger than their poorer neighbours in nearby Ghana.” 

So next time you feel you’re being conned with numbers, ask yourself the three foregoing questions.
Let’s end the way we started, with a wise counsel from Darrell Huff: “But without writers who use the words with honesty and understanding and readers who know what they mean, the result can only be semantic nonsense.”

Ibraheem Dooba, Ph.D, is a data scientist.

Source: premiumtimesng

No comments: