After another hard election FactCheck.org – the original independent fact-checking project in the USA – wrote: “We saw more aggressive fact-checking by journalists in this election than ever before. Unfortunately, as a post-election Annenberg Public Policy Center poll confirms, millions of voters were bamboozled anyway.”

That was 2008.

At the last moment that blog post swerved away from what might have been the conclusion: “are we discouraged that our efforts didn’t prevent this? Not at all. If we hadn’t tried, it might have been worse.”

It should have been discouraging, but it shouldn’t have been surprising. And it’s not surprising that they could have written it again in 2012 or 2016.

The idea that fact-checking can work by correcting the public’s inaccurate beliefs on a mass scale alone doesn’t stack up. Nobody should be surprised when, despite fact checkers publishing lots of fact checks, people still believe inaccurate things and politicians still spin and distort. Fact-checking can work but not if this is all we do.

It doesn’t work like that because current fact-checkers don’t have the reach to get the message through to all those it needs to reach. There is no realistic prospect of anyone funding the communications effort it would take to get that reach. Fact checkers are outspent by campaigns 100 to 1 or more at election times.

The most effective campaigns

And it doesn’t work like that because of how our brains work. Imagine walking late into a party where people have been arguing about politics, announcing that you have a PhD in what they’re talking about, and correcting everyone there. Would you expect them to be grateful?

The most effective campaigns of persuasion engage people’s emotions and not just their heads. Organisations that take sides are better placed to run campaigns of persuasion than fact-checkers.

This doesn’t mean that this approach to fact-checking is not worth doing. ‘Publish and pray’ is not a completely justified criticism. Simply publishing fact checks has two clear benefits, seen both in experience and in research.

The first benefit is to the audience it does reach, who it can inform and most of the time successfully correct when needed. That audience is often larger and broader than you might imagine, as search helps fact checkers reach millions of people who wouldn’t be regular readers, and now as the increasingly widespread work with Facebook puts fact-checking in front of some of its users.

When the fact-checker is known and trusted as fair, they can have influential audiences too of other journalists and policymakers using their work, and in that way they can help to inform debate, even when their work is not seen directly by everyone it affects.

Fact-checkers need to do more

The second benefit of publishing fact checks is to the wider public. It’s what we call the ‘they know we check’ effect. We see individuals’ and institutions’ surprise at being asked to justify the facts behind what they say, and we see governments, media outlets, and pressure groups making changes to avoid being vulnerable to factual criticism.

The first generation of fact-checking organisations tried fact checking with only this model of pure journalism. Many are still doing so. In our view the benefits are real but they are not nearly enough.

First generation fact-checking has a place as part of a wider set of tactics in a fact-checking organisation, or as part of the offer of major media outlets that can include fact-checking accountability in mainstream interviews or bring a large audience to the fact checks. This is where it has always belonged, at the heart of good journalism. Independent fact checkers can take some credit for inspiring a renewal of interest in fact checking in the mainstream media.

But, in our view, if we want a better answer than “if we hadn’t tried, it might have been worse” then independent fact-checking organisations need to do more than just publishing fact checks.

Some fact checkers have already moved on. Fact checkers who came after FactCheck.org got to learn from their open evaluation of their experience and come up with more effective models for tackling misinformation and disinformation. We call this second generation fact checking.

Second generation fact-checkers – like Full Fact, Africa Check, and Chequeado – do more than publishing fact checks. Our focus is on what we change not just who we reach. Our work is about power and accountability.

‘Publish and act’, not just publishing

What that looks like varies from place to place. It needs creativity and a sensitive understanding of the political and media environment to see what’s possible. But much of the logic is the same even when the details in different countries can be very different.

First, we move from just publishing to ‘publish and act’. We seek corrections on the record, pressure people not to make the same mistake again, complain where possible to a standards body. In other words, we use whatever forms of moral, public, or where appropriate regulatory pressure are available to stop the spread of specific bits of misinformation.

Secondly, we recognise that our fact-checking provides a unique evidence base that gives us important insight into where misleading claims come from in public life and how they are spread.

Fact checkers do the legwork of understanding the problems of misinformation and disinformation in a way that nobody else does. We turn this into decision-ready information to inform our own work, others in civil society, and platforms, regulators, policymakers, and others with the powers to act to tackle distorted information.

Thirdly, we work for system change. Using the evidence from our fact-checks we identify patterns and common causes, points where we can intervene to significantly reduce particular kinds or sources of information. The pattern might be who’s publishing something, where it’s published, a particular subject that there’s a lot of false information about, or something else.  The interventions can range from educating children or adults to advocating for policy changes.

Investing more in research

Finally, culture. We are trying to create institutions in different societies that can help anchor public debate to reality and to challenge the casual acceptance of deceptive and misleading behaviour. This is a long-term task: it involves earning good trusted reputations and not just getting attention. It needs funders to think long-term as well as fact-checkers.

It’s not enough to have an effective approach if it’s badly done. The simple description on paper hides the complexity of doing this in practice, the experiments and the failures and the learning that have got us to where we are. We still have a lot to learn, so now we’re investing more systematically than ever before in research to ground that learning in the best available evidence about all the different aspects of our work and the environments we work in. We know we will never stop misinformation and disinformation completely, but we can aim to reduce the harm it causes.

There is an old slogan among activists: educate, agitate, organise. Fact checkers have concentrated on education for a long time. In the future it will be our job to agitate our engaged regular readers and to organise to challenge those powerful people and organisations that provide and promote misinformation without accountability.

At Full Fact, we’ve been working on this broader second-generation approach to fact-checking since we began, inspired by that honest post-election blog from FactCheck.org. We made our first corrections requests in 2010; started working on misinformation policy in the UK in 2011; started working with the authorities to improve official statistics in the UK in 2012. We’ve had successes and we’ve made mistakes, and we still have a long way to go.

We cannot risk being left behind

At Chequeado, we developed the first live collaborative fact-checking model in 2013 that now many others organisations have replicated; created the first Education Programme to teach “critical thinking” for young people; created #LatamChequea, a regional network of fact-checkers; and convinced the President’s office to give them in advance all the sources of the President’s speech of “State of the Union” in 2018 and 2019.

At Africa Check, we have reached out to political parties and campaign groups to change the way they communicate in South Africa; we are bringing together the health ministry, health practitioners, and community-based groups in Nigeria to tackle health misinformation; and we are developing fact-checking curricula for schools across the continent.

Together we’ve shown that this approach of going beyond just publishing fact-checks can work across the world, even though it doesn’t work exactly the same way in different places.

Fact-checkers can’t afford to stay stuck on the first generation model, because already we can see that even second generation approaches are no longer enough. The third generation of fact checking will have to address all these issues – but also be able to function at internet scale, be massively collaborative, and work across international borders. We’ve already seen some of the early work that’s getting us closer to that third generation. The work of fact-checkers is too important to risk getting left behind.

We’d love to talk to other fact-checkers about how we can all have more impact.

The post BLOG: Fact checking doesn’t work (the way you think it does) appeared first on Africa Check.

Many people will consult doctors for relief from their tight chests and shortness of breath.

Enter inhalers: the solution to the problem of how to administer medication directly to the lungs. They have become important in managing “obstructive” lung diseases, like asthma and chronic obstructive pulmonary disease (COPD).

Because so many people need inhalers we should all be worried by news reports published by the Star newspaper in South Africa claiming that studies have shown that “inhalers can damage lungs”.

Original source unclear

Despite the alarming headline, the short article aims to raise awareness of COPD, and how it is frequently misdiagnosed. It is a republished version of an original story that was published in the UK’s Mail Online.

The connection between lung damage and inhaler use is never clearly explained in the Star’s version of the report. Instead it refers to the effect of inhaled corticosteroids (a class of medication used to suppress inflammation and the immune system) on the lungs, and concern that these drugs leave users more susceptible to infections.

We contacted the Star and the original Mail Online journalist to ask which study they were reporting on. We did not get a reply to our emails.

Without these answers, we are left with two important questions. Are inhalers themselves causing lung damage? Do inhaled corticosteroids cause lung damage?

Inhalers have been tested for safety

Inhalers are a drug delivery mechanism and are used to administer a variety of medications. A blanket statement like “inhalers can damage lungs” is like saying “eye drops can damage eyes” or “creams can damage skin”.

Inhalers come in two variants: metered-dose inhalers and dry-powder inhalers.

Metered dose inhalers consist of a metal canister, which contains both the drug and the substance that carries the drug — the propellant. The standard propellant used in modern inhalers has been shown to be so safe that dosages 200 times and upwards of the expected human exposure have been found to be non-toxic. This is in part because most of the inhaled propellant is exhaled again almost immediately.

Dry powder inhalers, on the other hand, are typically made from plastics. They contain a mixture of the active medication (in powder form) and a stabilising agent (typically lactose). The stabilising agent must be both safe to inhale and prevent clumping of the medication, so that the medication is delivered evenly to the lungs. This substance must also be non-irritating and be able to be broken down by the body or swept up and out by the cells lining the airways.

How do we know lactose is safe? As with the propellant in metered dose inhalers, lactose was first tested. One of the earliest reviews was an extensive series of animal studies on rats, dogs, monkeys and macaques. Researchers found no evidence that lactose had side effects on the lungs.

Other studies have investigated if dry powder inhalers are more dangerous than metered dose inhalers. A Brazilian study showed that there was no difference in side effects or efficacy of drug delivery. A Swedish study found that lactose had no negative effect on the airways.

Are steroids harmful to the lungs?

If research shows that inhalers are safe, that leaves one final avenue of investigation.

Corticosteroids — a type of steroid — are part of the typical treatment in the management of asthma. They have been demonstrated to improve quality of life, reduce risk of death from asthma and decrease hospitalisation rates. This is largely thanks to their ability to control inflammation.

“There may be some side effects to inhaled steroids, but significantly less so than taking these drugs orally,” Professor Richard van Zyl-Smit, head of the University of Cape Town’s Lung Institute, told Africa Check.

Taking a prolonged course of oral corticosteroids is associated with a host of side-effects such as osteoporosis, hypertension, weight gain, mood changes and cataracts.

Unlike in asthma, the narrowing of the airways in chronic obstructive pulmonary disease (COPD) is typically thought to be permanent, which means steroids have a less clear role to play in treatment. But what does the data suggest?

In 2012, the Cochrane group published a meta-review analysis of 55 studies. They found that using inhaled corticosteroids for more than six months did not significantly stall the slow but steady decline of lung function in COPD. Despite not changing the long-term outcomes of the disease, there was a minor decrease in symptoms and a small associated improvement in quality of life. On the other hand, the same review identified another more worrying trend: higher doses of inhaled corticosteroid were associated with a modest increase in hospitalisations for pneumonia.

This is the finding that brings us closest to “lung damage”. It’s important to distinguish that it’s not that the drug itself that damages the lung. It’s likely that the drug’s ability to suppress inflammation and hinder the disease-fighting function of the immune-cells may increase the risk of lung infections.

In 2014 the Cochrane society published another review, this time specifically analysing the connection between inhaled corticosteroids in COPD and pneumonia. Depending on the type of inhaled steroid used, pneumonia admissions to hospital were increased by 8 to 12 additional admissions per 1,000 people treated per year, but the difference in deaths between study groups was not significant.

Could the same be true for asthma? A group of researchers from the UK reported a two-fold increase in the risk of pneumonia or lower respiratory tract infection with the highest dose of inhaled corticosteroids.

The benefits outweigh any harm

In both asthma and chronic obstructive pulmonary disease, high-dose inhaled corticosteroids have been shown to slightly increase the incidence of hospitalisations for infections of the lung, which makes sense because of the immune-impairing effect of corticosteroids. But these increases are small in absolute terms.

In asthma the balance of positive effects on quality of life and general outcomes greatly favours the use of corticosteroids. In the case of chronic obstructive pulmonary disease its best to determine their use on a case-by-case basis with the aid of a doctor.

Despite this side effect, to claim that all inhalers — or specifically corticosteroids — can damage lungs is not only unsubstantiated but harmful.

“There isn’t really an argument over whether they harm the lung,” said Prof van Zyl-Smit. “Rather they save lives from asthma and there are many articles showing the benefits of inhalers”.

Petrie Jansen van Vuuren is a medical doctor based in Gauteng, with a postgraduate background in human physiology research at the University of Pretoria.

The post Analysis: Do inhalers ‘damage lungs’? appeared first on Africa Check.

He had contracted poliomyelitis as a child in 1963, a few years before the polio vaccine became available in Nigeria, leaving him unable to walk. In 2002, when his son Umar was born, an unfounded rumour about polio vaccination spread in the region and led authorities to ban the vaccine. Soon afterwards, baby Umar contracted polio too.

I was then living in Nigeria as bureau chief of the AFP news agency. Over the next few years, I followed reports of how false claims about polio vaccines (we didn’t use the term “fake news” then and should not now) led to a surge in the number of polio cases in Nigeria and surrounding countries. Thousands of victims – of polio, but also of misinformation.

I would later learn how a rumour about the tetanus vaccine started in 1994 at a meeting in Tanzania would spread to countries as far apart as Senegal and Kenya. Tetanus is a bacterial disease that causes muscles to tighten painfully. When it affects the jaw, people battle to breathe and can die. The false rumour about the vaccine means Kenya is one of just 16 countries that have not yet eliminated maternal and neonatal tetanus. And it is estimated that on average one Kenyan child dies from tetanus every day.

Almost two decades since the polio vaccine rumour spread, it might seem that the problem of misinformation has only worsened. Then rumours spread slowly, peer-to-peer on the human grapevine and on to the media. Today they fly via social media. This month alone has seen false rumours shared on Facebook stirring social and religious tensions in Nigeria, scamming people to pay for non-existent jobs in Kenya, announcing the death of a still healthy religious leader in Senegal and trying to stir resentment in South Africa.

The 2019 elections in Nigeria, Senegal and South Africa have all been information battlegrounds. Tactics have ranged from politicians distorting statistics to smears of political opponents, and from drives to suppress the vote to efforts to undermine faith in election outcomes. The revelation of the role of the firm Cambridge Analytica in past votes in Kenya and Nigeria has led to much speculation about the impact of disinformation on the recent polls.

But since 2002, some things have changed for the better. For one, misinformation is now widely recognised as a major problem worldwide and so is the focus of greater attention. What should we, and others concerned about these problems in countries from the DR Congo to Zimbabwe do about them?

Stepping down today seven years after launching Africa Check I take away four lessons, and a question.

1. Say what is right, not just what is wrong

In 2018, Africa Check organised a debate in Senegal on access to reliable information about public health. This information ranged from the prevalence of key diseases to what is spent on public hospitals. Even policymakers are often in the dark, yet health ministry officials failed to show up. And the picture is the same on the other side of the continent, in Kenya, where our team has struggled in recent weeks to get data on maternal mortality.  

“It is not that data does not exist,” I wrote in Africa Check’s first blog post, in 2012. “It does… But nevertheless, accessing that information is still harder than necessary and the quality of the data is not what always it should be. So, it will, I am sure, be a challenge.” It was and so it remains as we heard when Africa Check brought a dozen fact-checking initiatives together in Johannesburg in November 2017.

This is a particular challenge because fact-checking, in my view, is not only about calling out what is wrong but also showing what is right. This is why, earlier in 2019, Africa Check launched a new beta version of Info Finder, a tool aimed at making it easier for the public and policy-makers to find reliable information on a wide range of topics. It is why we will be working with the Open Data Institute on a project to open up data in Africa over the next three years. Providing information citizens and policymakers can trust is vital.

2. Look for the kernel of truth in a false claim

Yet being ‘right’ is often not enough to win arguments. I thought about this recently in South Africa when I asked an elderly couple how things were going. I heard a litany of woes, starting with: “Electricity has got so much worse in the past 20 years.” Apartheid ended in 1994, but we all knew what they meant. And with the country suffering frequent power outages, many would agree there are problems.

Do these mean electricity supply generally has “got worse” since 1994? It depends on where you start from – as an individual.

In 1994, as this Africa Check factsheet found, almost 100% of white households had access to uninterrupted power supply but only 37% of black households had the same. Overall, 54% of households had access to electricity. The comparable number in 2017 was 87%.

So whether electricity has “got worse” or “got better” depends on which end of the telescope you look at it from. Statistics often provide a correct national picture, but people live their lives locally. And we need to acknowledge that, or readers will dismiss our fact-checks. As this 2010 study showed, if we acknowledge our readers’ lived experience, or the kernel of truth in a flawed argument, our fact-check findings are more likely to be accepted.

3. Journalistic fact-checking alone is not enough

Independent fact-checking organisations were first established in the early 2000s in the United States as specialist units of traditional media. Deploying a journalistic approach to tackling misinformation, fact-checkers investigate claims to establish their accuracy and publish their findings. More than 30 years after my first reporting job, I still see myself as a journalist and there are sound reasons for limiting the approach to this.

But, in my view, if the goal is to reduce the harm misinformation causes in complex countries, such as those Africa Check works in, a more holistic approach is needed.

To achieve its overall goal of improving debate, Africa Check has five key targets. The first four are to:

• Identify and reduce the circulation of harmful misinformation
• Ensure accurate information is more widely available
• Ensure users retain a more accurate understanding of topics it covers
• Foster fact-checking skills and awareness among the wider public

In the first three months of 2019, our work reached an estimated 86.5 million people via traditional journalistic fact-checking, according to media monitors Newsclip. But as well as identifying misinformation, we also work to reduce the supply at source by reaching out to public figures and institutions to seek corrections and discuss an evidence-based approach to communication.

We also work with traditional and social media. Since 2015, Africa Check has trained over 2,000 newsroom journalists. As a result, as one respondent to a 2016 academic study said: “My writers now know they have to verify facts.” And to improve the fact-checking skills of the wider public, we are reaching out to schools, and recently brought in a new head of education and training, Carina van Wyk, with a great background in training work, to help develop a curriculum for teaching news literacy to young people.

4. Harness the power of other communities

Africa Check works in four countries with a team of just 30 people. The need we seek to meet is continental. For this reason, Africa Check has – since we were launched – sought to meet a fifth target of building a network of non-partisan fact-checkers across the continent and working with wider communities.

To build our Info Finder and make it a truly transformative tool, we seek to work with panels of independent experts, tapping into specialist knowledge we could not have. To develop fact-checking beyond the four countries we work in, we provide training and mentoring to newsrooms, are this year running our sixth annual African Fact-Checking Awards, and since 2017 bring together emerging fact-checking organisations including Pesa Check in Kenya, Dubawa in Nigeria, Congo Check in DR Congo and ZimFact in Zimbabwe.  

The growth of the fact-checking community in Africa is remarkable (and will be evident at Global Fact 6 in Cape Town next month), but not enough. As the Ebola crisis of 2014 showed, to reach those most vulnerable to misinformation in Africa, messages may need to be delivered in local languages by peers, not in English or French via a website. We therefore need to engage with those who can reach out at the community level. We’ll be doing that, later this year, with an exciting new project in Nigeria.

So, does fact-checking work?

When most people think about fact-checking, they think of journalistic fact-checking alone. They question whether exposure to a fact-check will persuade a die-hard adherent to one world view to change their minds. (The answer, according to this November 2016 study, is perhaps more than you think.)

My view however is that, to be effective – at least in the types of complex society in which Africa Check operates – fact-checking has to be more broad-based, and encompass a wider range of activities. It’s not just publishing reports, but liaising with traditional and social media, reaching out to public figures and institutions to both correct misleading claims and change the way they communicate, and working with educators and others to put news literacy into schools programmes and build awareness in society. That is what fact-checking needs to be judged on.

On this basis, what is the evidence? The honest answer is it is too early to tell, but my belief is that our impact accumulates over time.

In February 2019, a major South African political party approached Africa Check to fact-check their manifesto before publication – anxious to avoid the reputational damage of adverse fact-checks. (We did not oblige, to avoid partisan advantage for one party.) In Senegal, an economist writing the manifesto for one of the major political parties told a researcher our work means that “those writing the manifestos of the political parties, more and more pay attention to the accuracy of our figures”. Every month we see politicians and institutions correcting false and misleading claims. And we know that, as this study found, teaching the young to spot misinformation works.

Certainly, there are no simple solutions to such complex problems, but my view is a multiple-pronged approach such as ours appears to be the best hope we have of providing an answer. Future generations of people like Umar Tudun-Wada, the young man affected by the 2002 polio vaccine scare in Nigeria, deserve that we try.

The post COMMENT: Fighting falsehood, reducing harm? Lessons from Africa’s complex countries appeared first on Africa Check.

The first, expressed most recently in President Cyril Ramaphosa’s speeches, paints a picture of a country at war with its women, with the highest rape numbers in the world. The second, shown in police statistics, tells the story of an epidemic of sexual violence brought steadily under control.

Both these versions of the present are alive with the ghosts of politics past.   

Seventies estimate that only 5% of rapes reported

We can date the start of the first story to 1977, when a member of parliament estimated¹ that just 5% of rapes were reported. He called for a judicial commission of inquiry into all aspects of rape.

A dazzling array of numbers appeared, all supposedly revealing the true extent of rape. By 1981 media reports were describing South Africa as having the highest rate of rape in the world. Reflecting on these increases another member of parliament speculated² that rape “was a new form of terrorism.”

The numbers had an impact. In 1982 the minister of justice asked the South African Law Commission to investigate the law applicable to rape.

But where had these estimates of under-reporting of rape come from?

US estimate becomes local South African fact

No national studies had been done to quantify the number of unreported rapes in South Africa. But research and interviews³ from the period suggest that the source of the figure was US researcher Menachim Amir’s highly influential 1971 study Forcible Rape.

A section of the book cited prior studies of under-reporting, with the figure of 5% derived from research probably undertaken in the US before 1960. Nonetheless, the claim that only one in 20 rapes was reported in South Africa became a solid, local fact.

By 1990, with political violence escalating, this foreign estimate was informing the sort of speculation offered by Jacklyn Cock in her preface to Lloyd Vogelman’s book The Sexual Face of Violence: “If war is defined by the scale and intensity of political violence, then perhaps there is a hidden war being waged against women in our society… sources estimate that over one thousand women are raped daily in South Africa. This would suggest that almost two million women have been raped here since 1984.”

The figures were also recalculated to argue that a rape occurred every 83 seconds, or every 23 seconds (a statistic that continues to be cited). These claims were made more salient by the transition to democracy and the need to ensure that rape was placed high on any new government’s political agenda.

South Africa’s status as the “rape capital of the world” was cemented by a 1995 Human Rights Watch and a 1996 Interpol report. The police fanned the flames further by estimating the number of reported rapes to be only one in 36.

Rape, it was said, had reached “epidemic” proportions. At a time when the country was undergoing both a political transition and an emerging HIV epidemic, the argument resonated. Widespread sexual violence not only explained women’s disproportionate rates of infection but also supported demands that the government make post-exposure prophylaxis available to prevent HIV-infection in rape survivors.

More reliable figures emerge

The number of rapes recorded by the police in the 1990s increased significantly. But this upward trend predates the decade.

According to one researcher, reports of rape increased by 132% between 1955 and 1990⁴ – an increase that would likely have been greater had the Bantustans’ figures been taken into account.

From 1976 – when the first so-called “self-governing territory” was established – to 1993, the police excluded figures for the Bantustans from their crime reports. The reincorporation of the territories into South Africa in 1994 is part of the explanation for some of the recorded increases in rape in the mid-nineties.

Research institutions were also trying to understand the extent of violence in the country, with the 1998 South African Demographic and Health Survey finally providing a more reliable indicator of the prevalence of rape in South Africa. The researchers found 7% of their sample had been persuaded or forced to have sex against their will in the course of their lifetimes. Of the women who reported being physically forced to have sex, 15% said they had reported the incident to the police. (This is often reported as one in nine, but is more accurately calculated as one in seven.)

In 2000, using data from the 1999 victims of crime survey, Statistics South Africa calculated that 55,000 women (or 143 per 100,000 of the population aged 16 or more) had been raped in 1997. Of this number 68 per 100,000 had reported to the police. The Medical Research Council, however, recalculated Statistics South Africa’s prevalence rate as 390 per 100,000 women over the age of 16.

This difference was likely because Statistics South Africa used population figures for both men and women – even though the legal definition of rape at the time only applied to women. This meant prevalence rates for rape could logically only be calculated in relation to the female population. This use of total population figures may also explain why the SAPS prevalence estimate for rape in 1997 – 212 per 100,000 women – is also higher than Statistics South Africa’s estimate.

Data indicates a reduction in rape

The figures appeared at a time when it was imperative to show the police and government had crime under control.

In July 2000 a moratorium was declared on the release of the crime statistics to allow the police to revise their crime information systems. In January 2004 the police were set the goal of reducing contact crimes such as rape by 7% to 10% a year.

Police data for 2008/09 to 2017/18 suggests that the police have achieved their goal of reducing rape. Its incidence fell by 16.5% over that 10-year period while its prevalence decreased from 94 people per 100,000, to 70.5 people per 100,000 of the population.    

But these figures warrant a closer look.

The last decade: 2008-2018

The Sexual Offences and Related Matters Amendment Act, introduced at the end of 2007, widened the definition of rape to include an array of penetrative offences. It also recognised that that men and women could be both the perpetrators and victims of this crime.

The crime of “sexual assault” was introduced, largely referring to unwanted touching of the genital area, buttocks and breasts, as well as non-consensual kissing.  Other crimes were codified, such as grooming, making and/or distributing child pornography and all aspects of child prostitution.

The changes should, in theory, have led to an increase in reported sexual offences from 2008/09 onward. New crimes now existed and some existing crimes had had their definitions expanded. This did not happen.

While rape has decreased, the incidence of sexual assault increased by 6%, with its prevalence showing little change (12.9 per 100,000 to 12 per 100,000).

The drop in reported contact sexual offences is remarkable – a full 90%. The massive decrease in this category of crime explains most of the 27.6% decline overall in reported sexual offences.   

The notion of a war on women in South Africa persists, despite the decline in reports. Following the #TotalShutdown marches, Ramaphosa called for an end to this war in his Women’s Day speech on 9 August 2018.

Another old theme reappeared in his speech at the presidential summit on gender-based violence in November 2018. This was that South Africa had the highest rate of rape in the world.

New questions about statistics needed

This brief history suggests that the ways we think about rape statistics were fixed 20 to 40 years ago, trapping us in the two arguments that there is either a war against women, or the police are winning the war against crime. But what do the rape statistics actually say?

Why, despite some fluctuations, do the figures appear to have gone down? How did the police – or anyone else for that matter – achieve this drop? Does the decline reflect a reduction in the rate of reporting, an actual decrease in the prevalence of rape, or some combination of the two?

Fresh questions about rape statistics are long overdue if we want to meaningfully address the serious problem of rape in South Africa.

Footnotes

1. Hansard 6 June 1977, column 9331
2. Hansard 16 September 1981, column 3952
3. Cole, J. (1987). Die Vrou as Slagoffer van Verkragting. Unpublished Master of Arts dissertation, University of Pretoria; and ‘Anne Mayne’ in Russell, D. (1989). Lives of Courage: Women for a New South Africa. New York: Basic Books.
4. Glanz, L. (1993). Crime in South Africa: incidence, trends, perceptions’ in Lorraine Glanz (ed) Managing Crime in the New South Africa. Pretoria: Human Sciences Research Council.

Lisa Vetten is a Mellon Doctoral Fellow based at the Wits City Institute at the University of the Witwatersrand. This article is based on some of her PHD research, which provides a critical history of rape in South Africa since the 1970s.

The post ANALYSIS: War on women or crime under control? The two stories of South Africa’s rape stats appeared first on Africa Check.

The project – led by our colleagues at the UK fact-checkers Full Fact – is an ambitious plan to use artificial intelligence to transform the international fight against misinformation – in a collaboration with Africa Check, Latin America’s leading fact-checking organisation Chequeado and the Open Data Institute, co-founded by the inventor of the web Sir Tim Berners-Lee and artificial intelligence expert Sir Nigel Shadbolt.

And today the project has been named as one of just 20 projects to receive transformative funding from Google.org and support from Google’s AI experts.  

What we’ll be doing

Full Fact, pioneers in this field, have been working on automated fact-checking since 2013. Since then, with Chequeado and Africa Check, they have built technologies used across three continents and are working to roll them out to 50 organisations worldwide.

Over the next five years, we’ll come together to use machine learning to dramatically improve and scale global fact-checking efforts, working with international experts to define how artificial intelligence could transform this work, developing new tools and deploying and evaluating them.

We’ll help media outlets, civil society, platforms and public policy-makers worldwide understand how AI can help people decide what information to trust and bring the benefit of automated fact-checking tools to everyone.

In five years, we hope our project will let individual citizens and internet users place trust with confidence, help internet companies make fair and informed judgements at scale, and enable policy-makers to better understand how they can respond to misinformation while robustly protecting free speech.

A global effort

In Europe, across Africa and Latin America we have witnessed the very real harms that can be done to society by misinformation. But it is not limited to the countries we work in directly. Since January 2017, 33 people have been lynched by mobs in separate incidents fuelled by child kidnapping rumours that ran rampant on WhatsApp in India. Meanwhile, Minnesota battled its largest measles outbreak in nearly 30 years, with 79 cases. Researchers say the outbreak was fuelled by anti-vaccine messages.

The harm caused by misinformation and disinformation affects millions of people’s lives, health, safety and ability to participate in society, while governments around the world say they threaten to undermine democracy.

Technological responses to these challenges are risky and potentially harmful. They require a deep understanding of public debate, commitment to protecting free speech, close attention to the responsible limits of AI in this field, and recognition of how much these can vary between countries, languages and social and political contexts. Technological solutions to global problems have too often been developed in and for the Global North which is why we are thrilled to be working from the outset on these solutions with Full Fact, Chequeado and the ODI, helping ensure they meet the needs of society in Africa and Latin America too. We believe an effective response requires collaboration across all sectors of society.

But we can’t solve this problem completely in a few short years. By establishing a global network of experts in misinformation and disinformation, we’ll continue to innovate and apply emerging technologies to keep up with this ever-changing problem in years to come.

The Google AI Impact Challenge gives Google.org grant funding to organisations that are using AI to address social and environmental challenges. Full Fact, Africa Check, Chequeado and the ODI will share $2 million over three years, as well as receiving coaching from Google’s AI experts and credits and consulting from Google Cloud.

Read more about the automated factchecking tools we’re building. Read the press release here. 

The post Africa Check part of international group that wins ‘Google AI Impact Challenge’ appeared first on Africa Check.

Africa Check was established in 2012 at the Wits Journalism Department in South Africa. We also have fact-checking units in Kenya, Nigeria and Senegal doing the same work.

South Africans will vote on 8 May 2019. Africa Check has spent the last four months fact-checking the top three political parties election manifestos. It is not easy or quick work.

Our first step is always to contact the political parties and ask for the evidence they’ve based their claims on. Once we have this we compare it to the most recent and reliable research on the topic.

We then speak to experts and academics for context or analysis. We reach a verdict and then send it back to the political parties. We tell them what we found, we tell them why and we ask if they would like to comment. This can take three hours, three weeks or three months.

Political parties do get it right

Our aim is not to find errors and out people for getting their numbers wrong. We often find that political parties have their facts right – and we say so.

This year the incumbent African National Congress (ANC) claimed in its election manifesto that government had “invested more than R2 trillion in infrastructure projects over the past 10 years to build more schools, clinics, roads and the freight logistics network”. They were right. Treasury figures put the exact figure at R2.3 trillion.

The Democratic Alliance (DA) had its facts straight too when it claimed that “youth unemployment [in South Africa] is amongst the highest in the world”. While country comparisons of youth unemployment data have limitations, the available data does put South Africa among the worst performers. In 2018 the International Labour Organization estimated that South Africa had the highest unemployment rate for 15- to 24-year-olds: a staggering 53%.

The Economic Freedom Fighters (EFF) were also in the green when they claimed that “close to 40% of South Africans who need jobs are unemployed”. Almost 9.8 million people – or 37.7% of the labour force – were unemployed according to the expanded definition of unemployment.

Don’t muddy the water with false information

We give credit where credit is due but we’re quick to point out when political parties get it wrong.

Land and its redistribution is a key election issue this year. It is important that political parties do not muddy the waters or mislead voters with incorrect information.

The EFF was incorrect when it claimed that “the post-1994 governments have cumulatively bought less than 7% of the targeted 30% of land meant for redistribution”. The government has bought 6.7% of South Africa’s total land. But this is also a third of the land initially targeted for redistribution.

The DA’s claim that “the national government’s failure rate of land reform projects is currently standing at 92%” is baseless. There has been no national audit or study of the success or failure of land reform projects in South Africa. This statistic has been recycled and reprinted for years without anyone checking if it is based on rigorous research.

It was incorrect for the ANC to claim that “since 1994, over 4.7 million free houses have been built”. Government housing statistics show this isn’t true. Just under 3.3 million housing units were built between 1994 and December 2018. The figure of “over 4.7 million” refers to “housing opportunities”, which include serviced sites without an actual dwelling.

Facts matter – especially during election season

Why does Africa Check fact-check election manifestos? There are a couple of reasons.

At election time voters have to make a decision about their future. This decision may be based on personal preference or ideology but it’s also based on the information political parties share. Facts matter and they matter even more during election season.

The second reason is more abstract but equally important. It’s about policy.

The three political parties before us today are competing to govern South Africa. Should they win they will take their facts and misperceptions into government.

South Africa faces serious problems on a number of fronts. To solve them, any government will need an accurate understanding of these problems. A wrong understanding of how much land has been redistributed, how successful land reform projects have been or how many houses have been built could have serious policy implications.

Misconceptions could send us down the wrong road and waste valuable resources. We need our government and its officials to be well informed. The wrong information won’t help us make the right decisions.

Taking active steps to combat false information

Misinformation, propaganda and fake news have become a global issue. It has never been easier for false information to travel so quickly and widely. We must take active steps to combat false information. We each have a role to play.

Political parties and politicians need to commit to getting their facts right. This means investing in research. It also means issuing corrections when they get it wrong. Political parties hold enormous power and should use it with extreme care.

Journalists need to fact-check politicians – not just report what they say. The media should not simply act as a loudhailer, amplifying misinformation. They need to do the hard work of checking and contextualising information.

The public also has an important role to play. It is an integral cog in social media and the online space. We all need to take care to not share and retweet information that may be incorrect.

The challenges are immense. We likely won’t overcome them this election or the next. But by working together – and holding each other accountable – we can tackle misinformation and improve accuracy in public debate.

This is an edited version of a speech delivered by Kate Wilkinson at the Johannesburg Stock Exchange on 25 April 2019.

The post Voters’ choice and future governments: fact-checking in election season appeared first on Africa Check.

This was according to a 2017 World Health Organization report, host Trevor Ombija said during the Burden of Depression current affairs show. It was broadcast on Citizen TV, the country’s most popular channel.  

“That same report ranks Kenya as number six in Africa in terms of depression levels,” he added.

Suicide was also “a leading cause of death among youth”, the audience was told. This was linked to depression.

But does available evidence bear this out? As we found out after weeks of digging, data on suicide in Kenya is significantly unreliable. And that’s a major obstacle to finding solutions.

Over a thousand suicides – or 302?

The WHO report mentioned in the show is titled Depression and other Common Mental Disorders: global health estimates.

It estimates that 4.4% of Kenya’s population – 1,952,981 people – suffered depression in 2015.

Kenya does rank sixth in Africa for the number of people with depression, but places eighth for the share of population suffering the disorder. The report doesn’t give an estimate of suicides.

The UN health agency’s most recent suicide data on Kenya is from August 2018. This estimates the suicide rate as 3.2 cases for every 100,000 people in 2016. Using Kenya’s official population of 45.4 million people that year, this works out to 1,453 suicides.

But according to national statistics there were 302 suicides in 2016 (up from 221 in 2015). That’s less than a quarter of the WHO estimate.

What’s the source of WHO data?

How did the WHO arrive at its estimate? For months, Africa Check asked the health agency for its source data but received little feedback. The best we got from the WHO’s Kenya office was that most of the data on suicide in Kenya was prepared at their headquarters in Geneva.

A review of the agency’s research methods showed that where there was no data it drew from analysis by the Institute of Health Metrics and Evaluation, an independent research centre at the University of Washington.

This “made use of available death registration data as well as other sources of information on deaths, covariate regression modelling, and patterns of causes of death for similar countries”.

Lack of data on the causes of death

In its 2017 report, the WHO said only 60 of its member states had vital registration data good enough for the direct estimate of suicide rates. Kenya was not among them, due to a lack of data on the causes of death.

Joyce Nanjala, who works in the WHO’s Kenya office, said the lack of a “national survey” on suicide and “very limited surveillance data” made it difficult to track suicide trends.

“It would be helpful for the country to conduct systematic review of local epidemiological studies to have a better understanding of the local suicide trends,” she told Africa Check.

Police only record attempted suicides

Police data puts Kenya’s cases of suicide in 2017 at a provisional 421, according to the 2018 economic survey by the Kenya National Bureau of Statistics.

But this is unlikely to be the number of actual suicides in Kenya in 2017. The police only record suicide attempts. Once a person dies, there isn’t anyone to charge. This is according to a source in the office of the inspector general. (They are not authorised to speak on the record.)

Kenya’s penal code states that “any person who attempts to kill himself is guilty of a misdemeanour”.  

Further complicating the picture, less than half (41.2% in 2017) the deaths in Kenya are officially recorded, according to the statistics agency.

‘Intention of death is not recorded’

Most of Kenya’s data on suicides comes from “lay reporting” and there is little coded data, Samuel Cheburet, the head of the civil registration and vital statistics unit at Kenya’s health ministry, told Africa Check.  

“Most of the medical certification of deaths doesn’t say it is suicide, because the intention is not recorded.”

So even if a person kills themself “in a traffic accident, or in a homicide, the intention is not recorded”, Cheburet said.

Experts agree on lack of data

All the experts we spoke to highlighted the lack of accurate data.

Inaccurate coding of cause of death data meant most cases of suicide were not captured, Prof David Ndetei, a mental health expert at the department of psychiatry of the University of Nairobi, told Africa Check.

“There are those who throw themselves in front of a speeding vehicle… it will not be recorded as suicide.”

Stigma, misclassification and criminalisation

There isn’t a “specific place to get reliable suicide data on Kenya because it is not routinely collected in our health systems”, said Mary Bitta, a researcher at KEMRI-Wellcome Trust, a health research institute in Kenya and doctoral candidate at the University of Oxford.

Bitta has studied suicide in coastal Kenya. She gave a number of reasons for the poor state of data:

• No suicide report verification using coroner certificates
• The criminalisation of suicide in Kenya
• Low reporting due to culture-based stigma
• Misclassification of deaths by other causes such as accidents

What’s the result?

Does Kenya’s lack of accurate suicide data have consequences?

Misleading data could cause “unnecessary alarm”, Nairobi University’s Ndetei told us. He is the founder of the Africa Mental Health Foundation and has studied suicidal behaviour in Kenya’s youth.

And, he said, underestimates could mean not enough was done to prevent suicide.

Further reading:

The post ANALYSIS: Poor data on suicide leaves Kenya groping in the dark for answers appeared first on Africa Check.

We’ve debunked the claim that South Africa has the worst maths and science education in the world three times. It’s still wrong! We’re also three fact-checks deep into the often repeated (and incorrect) claim that Zimbabwe has the highest literacy rate in Africa.

Growth of South Africa’s economy

Since the beginning of 2019 Africa Check has twice rated as misleading a claim by the African National Congress, South Africa’s ruling party, that the country’s gross domestic product (GDP) has tripled since 1994.

I first fact-checked the claim when President Cyril Ramaphosa said it in the party’s traditional January 8 statement: “Despite economic challenges over the last few years, our economy has tripled in size over the last 25 years.”

GDP is a measure of the size of a country’s economy. It’s the value of all goods and services produced in a certain period, usually a year.

Data from the World Bank shows that South Africa’s current or “nominal” GDP grew 2.5 times from US$139.7 billion in 1994 to $349.4 billion in 2017.

But this growth is in current or nominal GDP, which includes the effects of price inflation over the years. Comparing nominal GDP from different years gives you “completely distorted figures”, says Prof Jannie Rossouw, head of the school of economic and business sciences at Wits University.  

It’s important to separate real economic growth from inflation, he says.

World Bank data based on constant or real GDP (where the effect of inflation is removed) shows that the country’s economy grew 1.9 times from $225.6 billion in 1994 to $426.7 billion in 2017.

Claim repeated – twice

Yet the misleading claim was repeated two more times.

Later in January, we fact-checked a tweet by the ANC’s head of risk management, Bo Mbindwane, which said the party had grown GDP “three fold” since 1994.

The South African government’s Twitter feed posted the same claim in February.

“How many more times could we publish the same fact-check?” I thought.

We’re not trying to catch people out

So imagine my delight when during a parliamentary question and answer session on 7 March, Ramaphosa changed his tune on the country’s GDP growth.

“Do you honestly believe that the economic policy of the last 25 years has brought real change to the economy and the people of this country?” an opposition member of parliament asked him.

Ramaphosa said there had been progress. “The GDP of our country has almost doubled.”

The president was spot on. South Africa’s real GDP has almost doubled. It’s now 1.9 times the size it was in 1994.

I can’t be sure that our fact-checks helped persuade the president to re-examine the claim. But it’s a great reminder of why we sometimes fact check the same claim over and over again.

We’re not trying to catch people out. We’re hoping that, eventually, they get it right.  

The post Checking the claim again and again – until they get it right appeared first on Africa Check.

On 1 June, seven years after Africa’s first fact-checking organisation was set up, current deputy director Noko Makgato will take up the role of executive director, replacing the founder, Peter Cunliffe-Jones.

“Founding executive director Peter Cunliffe-Jones has laid a solid foundation for this important organisation to build on. Having worked with the new executive director, Noko Makgato, for some years, I know we are in solid hands to consolidate and grow further. Noko has laid out excellent plans for the coming years and I am confident we will bring them to fruition,” said board chairperson Anton Harber.

At the same time, the organisation unveiled a new, eight-member governing board of the Africa Check Trust, which oversees Africa Check’s growing operations across the continent.

Chaired by University of the Witwatersrand Journalism professor Anton Harber, the board includes Cami Mbulawa, an SA-based executive with investment advisers RisCura as Treasurer, Dapo Olorunyomi, the publisher of the Premium Times and former chief of staff at the Economic and Financial Crimes Commission in Nigeria, Karien Bezuidenhout, a director of the Shuttleworth Foundation, Mandla Mchunu, the former Chief Electoral Officer of South Africa and chairman of the AfriCore group, Pamella Sittoni, the Executive Editor of the Daily Nation in Kenya, Paula Fray, the founder of FrayIntermedia in South Africa, and Tidiane Sy, the head of the EJICOM journalism school in Senegal.

Set up in 2012 in South Africa to reduce the spread of misinformation and promote evidence-based understanding, Africa Check today works in South Africa, Nigeria, Kenya and Senegal, checking the accuracy of hundreds of key claims made each year on leading topics.

As well as publishing its reports on its own site, it distributes its fact-checks via partnerships with numerous media houses and is working directly with Facebook to reduce the spread of misinformation on the largest social media platform.

Since 2014, Africa Check has run an annual awards ceremony to support the growth of fact-checking by the media. And since 2015, when it launched its training and research commercial unit TRi Facts, it has trained almost 2,000 journalists and others around Africa in fact-checking, to build a culture of fact-checking in the media. In 2017, it set up the African fact-checking network to support emerging independent fact-checking operations such as Congo Check in DRC, Dubawa in Nigeria, PesaCheck in Kenya and ZimFact in Zimbabwe.

“When the public and policy-makers make decisions based on misinformation, it can bring real harm,” said out-going executive director Peter Cunliffe-Jones. “In 2012, we were the only independent fact-checking organisation operating on the continent and working in just one country. Today, we have teams in four countries and are working with a growing number of partners around Africa. In June, we will welcome colleagues from fact-checking organisations around the world to Cape Town – the first time the Global Fact meeting has been held in Africa. The gathering pace of change is thanks to both a hard-working team and the recognition by so many partners and allies of the real importance and need for the work we do.”

Africa Check is the African representative of the advisory board of the International Fact-Checking Network (IFCN), the standards body of the fact-checking movement, based at the Poynter Institute in the United States.

The post [PRESS RELEASE] New Executive Director and board at Africa Check appeared first on Africa Check.

In an interview with SAfm’s Stephen Grootes the next day, the minister blamed this on inadequate funding and “the issue of foreign nationals”.

Grootes asked if there were statistics on the share of foreign nationals using public healthcare. Motsoaledi said the health department did not officially collect this data as it was a “very sensitive matter”.

But he did say he had “a lot of anecdotal data”, such as the number of “not South African” babies born at three public hospitals – two in Gauteng province and one in Limpopo.

The minister said he “did not want to raise any debate or storm” but was “just mentioning the facts which nurses are dealing with every day”.

Expectant mother turned away from Gauteng hospitals

Motsoaledi’s first example was from Charlotte Maxeke Academic Hospital in Johannesburg. He said “two years ago” it was accused of being one of three hospitals that turned away a pregnant “Burundian woman”, who later gave birth at an inner city train station.

The incident was “in June that year”, the minister said. But in May the hospital “helped deliver 700 babies – and 400 of them were not South African at all”.

Lungiswa Mvumvu, a spokesperson for Charlotte Maxeke, told Africa Check that in May 2016, the hospital recorded “838 births, 200 of which were to non-South African citizens”.

But the incident the minister was referring to took place in June 2017 not June 2016,  And the woman involved came from the Democratic Republic of the Congo, not Burundi.

In May 2017 the hospital said it listed 771 births, 381 of which were to mothers from outside South Africa.      

Many foreign mothers ‘don’t go for treatment’

Motsoaledi then told Grootes that “on Christmas, 60% of the babies” delivered at the Steve Biko Academic Hospital in Pretoria were “not South African”.

The minister meant Christmas Day in 2017, national health department spokesperson Popo Maja told Africa Check.

Maja said Steve Biko was a referral hospital, which took in patients suffering complications. For women giving birth, these complications could be a result of them not getting healthcare during their pregnancy.

“Most people sent to Steve Biko are usually in critical condition. Many are foreign nationals who do not go for prenatal treatment as they are undocumented and so they often have birthing complications.”

Steve Biko hospital spokesperson Lovey Mogapi told Africa Check that six babies were born in the hospital on 25 December 2017. Only one was to a foreign mother.

‘Foreign’ babies born in border hospitals

Motsoaledi also said that when the head – or MEC – of Limpopo province’s health department visited Messina (also known as Musina) Hospital recently, “out of 10 babies that were delivered, seven were not South Africans”.

We have not been able to confirm the date of the MEC’s visit with her office. But “on average per week, the hospital delivers about 21 babies”, department spokesperson Neil Shikwambana told Africa Check.

“Out of those, 11 are foreign nationals and 10 are South Africans.” Shikwambana could not tell us how this average was calculated.

At face value, that means more than half the babies born at the hospital are “not South African”. This is lower than the 70% claimed by Motsoaledi.

We asked Lorena Nunez Carrasco, an associate professor in the department of sociology at Wits University, if she could explain.

“It makes sense,” she said. “If you need to give birth and there is a better service right there, a few kilometres away, but it belongs to another country, then you would go.”

Messina Hospital in northern Limpopo is just 15 kilometres from the Zimbabwe border. It’s also within reach of people from Mozambique to the east, and Botswana to the west.

Do foreign mothers ‘flood’ South Africa to give birth?

Carrasco said there was “no rationality behind this idea that women are flooding the country to give birth”. Her areas of study include how social and cultural factors affect health.

There might be “a few cases”, but these were only in border cities. And, she said, the people who did migrate were often the physically strongest and often have family or some other connection to South Africa.

“You don’t have many frail or sick people migrating… migrants don’t make irrational decisions. To travel at eight or nine months pregnant is risky and costly.”

What of Motsoaledi’s examples from the two Gauteng hospitals? These were “not extraordinary” and in line with global averages, Carrasco said.

“One has to first take into account the size of the migrant population in South Africa. It varies from three to five percent of the total population.”

Gauteng was home to over a quarter (25.4%) of South Africa’s population, so most migrants were likely to live there, Carrasco said.

Identifying ‘foreign’ babies

But how do officials know a baby is “not South African”?

Patients were identified as foreign nationals if they were unable to produce identification or proof of residence, or could not speak English or other local languages, the national health department’s Maja said.  

But there were weaknesses in this approach, Jo Hunter-Adams, a researcher at the University of Cape Town’s Health Economics Unit, told Africa Check.

“Hospital staff would not necessarily recognise all South African languages and those who are ill may not be able to communicate.”

What impact do migrants have on the health system?

“At a national level, very little,” Jo Veary, associate professor with the African Centre for Migration & Society, told Africa Check. Among her research works is how health development in South Africa is affected by migration.

South Africa did have progressive laws, Veary said. But in practice migrants from elsewhere in Africa were not attracted to South Africa for the quality of its healthcare system.  

“Individuals know that the system is not functioning at the level it should be, and people know that the treatment received as foreign nationals is poor.”

Hunter-Adams agreed that many migrants found it difficult to get public healthcare in South Africa.  

She said there was a “national discourse” that encouraged “medical xenophobia”.

“The narrative that foreign nationals are a ‘burden’ to the South African health system seems to be an example of scapegoating, where migrants are blamed for broader systemic problems with the health system.”

The post ANALYSIS: Are South Africa’s public hospitals ‘overburdened’ by foreign patients? appeared first on Africa Check.