Home

Monday, May 27, 2019

The Gender STEM Paradox, and Its Implications for Future Technology

A few months ago, one of my friends Lukas wrote a nice post about the STEM gender paradox. The paradox is this: in countries where women are relatively empowered, women are less likely to choose math and science subjects and are consequently less likely to enter math and science professions. For example, Finland and Norway score the highest in the Global Gender Gap Index (which measures things like female economic participation and political empowerment), yet have one of the lowest proportions of female STEM graduates in the world (~20%). Conversely, in Algeria – a country with a notoriously bad track record for women’s rights – over 40% of STEM graduates are female. Figure 1, from a paper by Stoet and Geary (2018), illustrates this more comprehensively:

Figure 1: The STEM Gender Equality Paradox 


Source: Stoet and Geary (2018)

Why is this?


One touted explanation for this paradox is that men have a natural aptitude for science, and this absolute advantage leads to their disproportionate representation in more gender-equal (and developed) countries. In less developed countries, this natural advantage may be masked by information asymmetries in University admissions and greater inequalities in education access from the outset. Nonetheless, as countries develop, these imperfections slowly recede and a natural gender bias re-emerges.

This idea, sadly, has some precedent. In January 2005 Lawrence Summers – Economics Nobel laureate and then-President of Harvard University – received heavy backlash after claiming that lack of female representation in STEM could be caused by ‘natural differences in high-end aptitude’ between the genders, prompting a media storm and his eventual resignation two months later.

Needless to say, there is very little evidence to support Summers’ view. In looking at numeracy test scores across 67 countries and regions, the Stoet and Geary paper found that girls performed just as well or better than boys in most countries, and in almost all countries girls would have been capable of college-level STEM classes had they enrolled in them.

A better explanation, put forward by the authors themselves, suggests that men have a comparative but not absolute advantage in STEM subjects. In all countries except Romania and Lebanon, the researchers find that boys’ best subject was science and girls’ best subject was reading. That is, girls performed just as well or better in science – they just happened to perform even better in reading.

Moreover – and here’s the kicker – the researchers find that the more gender-equal the country, the larger this gap between boys and girls in having science as their best subject. This completes the explanation: as countries get more gender-equal, boys’ comparative advantage in STEM increases, widening the STEM gender-gap at University.

Stoet and Geary also put forward another explanation that invokes life quality pressures. In less developed and more gender-unequal countries, women face greater economic and life risks. As a result, those fortunate enough to obtain a tertiary education are more likely to choose STEM subjects, under the (generally well-founded) belief that these are more likely to lead to a secure and lucrative job.

The authors then argue that these financial considerations become less burdensome as countries develop, giving people greater freedom to pursue their true interests. One of their findings supports this: even when girls’ ability in science in school equalled or excelled that of boys, they nevertheless tended to register a lower interest in science subjects. This explanation has a strangely uplifting implication: it’s not that gender equality discourages girls from pursuing science; it’s that it allows them not to if they’re not interested.

Outstanding Issues


I find both these explanations unsatisfying. At its heart, the life pressures explanation rests on the idea that men are naturally more interested in science than women – an idea with very little intuitive appeal. Indeed, by invoking the idea of an ‘interest gap’, all we seem to do is pass the buck on to another, no-less tractable problem.

Likewise, the comparative advantage explanation relies on the idea that greater gender-equality leads to more pronounced relative advantages for women in non-STEM fields. Again, there is no obviously intuitive explanation for why this might be the case.

More fundamentally, though, both explanations seem to overlook a large elephant in the room. In emphasising a benign sorting process based on interest or comparative skill, both explanations conveniently side-step the numerous structural barriers that prospective female STEM students face in countries we otherwise consider relatively gender-equal. These barriers are extremely well-documented: in recounting her experience as a woman in STEM, Ellen Pao describes the accumulation of micro-aggressions as like “a thousand tiny papercuts”. Given the pervasiveness of these barriers, any explanation which stays silent on them seems patently incomplete.

Why Does This Matter?


If progress towards economic growth and gender-equality comes at the expense of equal STEM gender-ratios, this has a number of unsettling implications. First, gender STEM ratios today are likely to heavily influence the gender composition of tech companies tomorrow. This matters because – as recent research shows – algorithms routinely inherit the biases of their creators. I like the example Lukas gives: many ‘smart’ heating systems take the metabolic heartrate of a middle-aged man as the default, making the entire algorithm geared towards maximising the experience of this demographic. If more women were included in the design process, this unfair gender-bias may have been noted earlier and rectified.

We can push the envelope further. As Yuval Harari and Jamie Susskind recently argue, the commodification of personal data is likely to progressively shift power towards tech companies away from traditional sources of authority, like nation states and religious institutions. The future leaders of these companies are likely being educated today – in STEM lecture halls marked by a conspicuous absence of females. If we care about the continuation of patriarchal power structures, we need to take the STEM paradox seriously.

Word count: 954 words

References

For a great discussion of the implications of tech on the future of politics and power, I would recommend:
  • Harari (2015), ‘Homo Deus’
  • Susskind (2018), ‘Future Politics’
Posted by at No comments:

Wednesday, May 22, 2019

The Story of Kane & Abel, or Why Early Childhood Education is So Important

Kane & Abel


In the early hours of Christmas morning in 1984, Kane Kawooya is born in Nsambya Hospital, a private facility in central Kampala. His Mother, one of the few female Deputy Ministers in President Obote’s government, enjoys a trouble-free pregnancy and birth – aided by a healthy diet, good antenatal care and generous maternity leave.

At the same time, another boy named Abel Masane is born in a bush hospital near Gulu, Northern Uganda. His birth marks the end of a long and arduous pregnancy haunted by chronic stress: Abel was conceived during the height of the Ugandan Bush War.

Kane is raised by a loving family. Though both his parents are full-time professionals, they spend most evenings and weekends interacting with Kane and taking him to play with other children. During the week, Kane is looked after by house staff who feed him, speak with him and introduce him to the many toys that litter his home. At age 4, Kane is enrolled in a nursery school – one of the first in Uganda – where he quickly develops basic reading, writing and numeracy skills.

Abel is also raised by a loving family. However, under the looming threat of poverty, his parents spend most of their time working in agriculture and petty trade. Due to the limited time they spend with each of their five children, Abel’s parents have adopted an authoritarian parenting style that tends to issue orders, not explanations.

Abel’s Mother did not go to school, and knows little about nutrition, sanitation or parenting best-practices. He does not attend nursery as there are no government-run pre-primary schools in Gulu at the time. However, by age 6, his parents have managed to scrape enough money together to send him to primary school, just as Kane is putting on his first school uniform in Kampala…

Before their first day, the two boys look and sound very different. Kane towers 4 inches above Abel – a tell-tale sign of Abel’s stunted growth. Kane’s language is more sophisticated and he is better able to cooperate and empathise with other children. Even their brains are different: due to worse nutrition and less mental stimulation, Abel is likely to have a lower density of grey matter in his prefrontal cortex and a smaller anterior cingulate, an area important for emotional and cognitive tasks.

Early Childhood Development


These boys are, of course, fictional - named after the biblical tale of Cain and Abel (or Jeffrey Archer's novel). But the research that underpins their lives is very real. Poor children hear up to 30 million less words than their richer counterparts by age 3. Parents from low-income backgrounds are more likely to adopt authoritarian and less engaging parenting styles. When mothers are exposed to stress during pregnancy, their children are more likely to experience emotional or cognitive problems, which can affect their willingness to cooperate even in much later life. Stunted children exhibit slower brain development and interact less with their environment. As Pritchett and Beatty show, these initial differences are likely to compound once primary school starts, as disadvantaged kids get left further and further behind the curriculum.

Findings like these strike at our emotions in a way unlike many others. The fact that some kids are so far behind before they even step foot in a classroom violates our innate sense of justice and puts to rest many hopes we harbour about social mobility.  

How Can We Help Kids like Abel?


If we can tackle childhood inequalities early, this may provide benefits which accumulate through time, providing much more bang-for-buck than interventions designed to help later on. Preliminary research is broadly consistent with this: a study by Gertler et al. (2014) suggests that early childhood interventions can increase labour market returns by 25%, compared to 7-10% for school-age interventions.

One example of an effective intervention is parental education. In Jamaica, Gertler et al.’s study finds that teaching parents of growth-stunted children parenting techniques that are conducive to cognitive and socioemotional development increased the adult earnings of these kids by 25% 20 years later, enough for them to catch up to the earnings of their non-stunted peers. In a similar study in Nepal, Levere et al. (2016) find that parental education interventions are most effective when deployed simultaneously with a cash transfer, as this alleviates financial constraints that can prevent good practices from being adopted.

Another example is to expand access to pre-primary education (PPE). Barriers to PPE will vary by context: in Zanzibar, the biggest barrier is infrastructure – there are simply not enough pre-primary schools. The graph below is revealing: while Zanzibar has made some progress towards universal pre-primary enrolment, private schoolchildren are still far more likely to have access than their government counterparts.

Source: Zanzibar EMIS Data

PPE can be delivered in a number of ways. Zanzibar has three types of delivery modality: standalone pre-primary schools, integrated pre-primary and primary schools and ‘Tutus’ – community­-run centres that rely predominantly on radio-based instruction. These differ in cost and likely effectiveness – standalone pre-primary schools are the most expensive but likely to provide the best-quality pre-primary education.

To the best of our knowledge, very little has been done to distinguish the cost-effectiveness of these different delivery modalities. In our opinion, this represents a fruitful avenue of research to the enterprising student.

Word count: 996 words

References


[1] This is known as the Matthew Effect or, colloquially, ‘the rich get richer’.

Posted by at 3 comments:

Sunday, May 5, 2019

How to Write an Essay

People occasionally ask me if I have any advice to offer them on how to write a good essay. Most of the time, I reply with fairly long-winded and jumbled voice notes as I wrack my brain for all the disparate pieces of advice I’ve taken onboard from friends, teachers and articles over the years. I’ve decided that instead of doing this, it would be more useful to consolidate this advice into one document and then upload it online so that anyone else can access it too.

This post draws from a number of sources; but by far the biggest inspiration is a wonderful guide written by the late Bob Hargrave, a Professor of Philosophy at Balliol College. A link to this is provided at the end – I implore you to read it!

Disclaimer: The type of essay I had in mind when writing this guide was a typical assignment or exam essay you would be set in a social science or humanities undergraduate or Masters course. This guide is based on my experiences as a student of Philosophy, Politics and Economics – as such, I cannot guarantee that it carries over to History, Geography, English Literature etc. Likewise, parts of this guide do not extend to creative writing pieces, blogs, or more journalistic forms of writing.

Further disclaimer: I am not an examiner or teacher. My only real authority on this topic stems from the great advice I’ve received over the years, and from the fact that I have tended to punch above my weight in essay-based exams. You follow my advice at your own risk. With these caveats in mind, I present to you the 10 commandments of good essay-writing.

Commandment 1: Answer the damn question


This one comes first as it is by far the most important. It also represents the most common pitfall that poor essays fall into. Open any Examiners Report on any essay-based subject at any University and I can almost guarantee you will come across a statement like this:

“Quite a few students attempted to squeeze prepared answers into the frameworks provided by the questions, rather than thinking about the question and attempting to answer it directly.” - Theory of Politics Examiners Report, University of Oxford, 2015.

Do not attempt to shoe-horn in related essays which do not directly answer the question posed. Remember: essays are not are not a test of how much you know about a subject. They are a test of how well you can explain and argue what you do know.

A good way to test whether you have answered the question is to play the “Guess the Question” game: if you were to cover up the title, would someone reading your essay be able to guess what it was? From Professor Hargrave’s guide, here are some unfortunate answers to this game from Oxford finals examinations:
  • Write down all you can remember about Locke on Personal Identity, in no particular order. Abandon any attempt at accuracy or rigour. Above all, do not draw any conclusions.
  • Think up a really stupid view on Personal Identity, which no philosopher has ever held, and blame it on Hume. Remember to end your essay in the middle of a sentence.
  • Compose a brief piece on Personal Identity parodying the style of Readers' Digest, and beginning with the phrase "Ever since the dawn of time...".
  • Copy out from memory your tutorial essay on Personal Identity.
  • Confuse Hume with Berkeley. If you have never heard of Berkeley, just confuse Hume.
To avoid being immortalised in future essay writing guides such as this, write this in two-foot letters on your wall: answer the question exactly, fully, and only the question in front of you.

Commandment 2: Argue a case


All essays make a case – in fact the word ‘essay’ itself originates from the Latin word ‘exagium’, which means to weigh up or ascertain. A case is essentially a structured argument (see Commandment 4) which aims to convince the reader that the authors' view on the question at hand is correct. An essay is not a literature review, nor is it an invitation to knowledge-dump everything you know about the subject.

In making a case, you should avoid sitting on the fence (e.g. “this evidence points this way, this evidence points that way, so in the end I don’t really know”). This does not mean you have to take ridiculously one-sided or polemic positions. Be nuanced by all means: but if you think a ‘middle way’ provides the best answer to the question, then defend that position with fervour.

Of course, to argue a case, you need to know your view on the matter. Which assumes you actually have a view on the matter. Which leads to…

Commandment 3: Defend a position you actually believe


This may sound obvious, but it wasn’t for me when I first started writing essays. Rather than thinking critically about the question, I would tend to assemble the most convenient argument based on what I’d read. For example, if I’d read more about compatibilism than determinism then I would defend compatibilism in my essay on free will based on my (flawed) assumption that I had more evidence in favour of the former, so writing the essay would be easier.
If you’re not sure about your opinion on a particular topic – that’s okay! Think about it, read about it, speak with classmates about it until your view becomes clearer. In an exam, I would not recommend answering questions upon which you don’t have a settled view. The pressure of an exam is not the time to be deciding what your genuine opinion is; this should have been done long ago during your revision.
Defending a position you actually believe in will make your prose sound more authoritative and convincing (see Commandment 10). It will also (hopefully) make the whole ordeal of essay-writing less pain-staking.

Commandment 4: Plan your structure


There is a simple way to think about structure. Imagine you are in a court trying to convince an intelligent, scholarly, but not-necessarily-well-informed-on-the-question-at-hand jury that your view on the question is correct. The structure you would use to make this argument should be the same as the structure you use to write your essay.

What might this look like? It should be logically coherent – in the sense that if the jury accepts every point you make, it shouldn’t be a great leap for them to accept your conclusion. It should also be evidence-based, where evidence can take the form of real-world statistics or anecdotal examples. Above all, it should be clear. You are far better off concentrating on simple points that are under your command rather than sophisticated points which are not under your command. Apply this advice to many more walks of life.

Commandment 5: Deliver a strong thesis statement


A thesis statement is a one sentence summary of your argument which exactly answers the question posed. This is useful for many reasons: i) it leaves the reader with no doubt in their mind where you stand on the topic ii) it forces you to articulate clearly what you actually think iii) it provides a good focal point for the rest of your essay to refer back to.

Your thesis statement should be delivered in your introduction, preferably as early as possible. Which leads on to the next point…

Commandment 6: Nail your introduction


One of the most powerful biases humans exhibit is the ‘halo effect’ – the idea that if you make a good first impression, then everything you do subsequently is more likely to be perceived in a favourable light (and vice-versa). Examiners and readers are no different: if you start off erudite you can get away with murder later in your essay; if you start off muddled then it will be an uphill battle to change their opinion about you.

Good introductions have the same essential ingredients: a crisp and punchy opening sentence, a thesis statement delivered clearly and unambiguously, an indication as to how you will go about defending this statement, and absolutely no flannel, waffle, or anything else unrelated to the question at hand. For examples of good, bad, and ugly opening paragraphs, refer to the next section.

It is also important to make sure that your introduction and conclusion speak to each other. A tutor once told me that he could predict, with a high-degree of accuracy, how good an essay would be just by reading the introduction (to see where the author thought they were going) and conclusion (to see where the author thought they had ended up). If the two paragraphs were sensible, clear and aligned, then this was a strong indication that the author had produced a good essay.

First (and last) impressions matter. If you’re sitting an exam in which you’re asked to write three essays, I would recommend concocting a ‘shit sandwich’: putting your strongest essay in the envelope first, the weakest in the middle and the final one at the end. This is a small point and probably won’t make much difference – but you might as well game the system as best as you can.

Commandment 7: Keep it short and focused


If your tutor has imposed a word limit, stick to it. This is not to encourage laziness: it is to sharpen your writing and force you to think carefully about structure.

Make every paragraph count. Each paragraph should be linked to and drive forward your thesis statement like a tentacle of an octopus. Each tentacle should make its own independent contribution: consequently, if any tentacle is dismembered, this should make your overall argument weaker. 

A good way to test this is to cover up each paragraph and then ask yourself if your essay is any way weaker without it. If the answer is no, delete that paragraph.

Commandment 8: Cite, but don’t name-drop


Citing is important: in an assignment it shields you from plagiarism and in an exam it signals to the examiner that you are well-informed on the topic. But don’t name-drop. What I mean by this is don’t cite for the sake of citing e.g. just to mention an author you want to show off that you’ve read. No one cares that you suffered through the entirety of Hobbes’ Leviathan unless it directly relates to your argument.

In terms of technicalities, it is fine to just put the authors surname and the year in parentheses at the end of your sentence (Harvard referencing); in an exam, you can drop the year altogether. In general, avoid quoting text verbatim unless you literally cannot put it better in your own words (which will be rare!)

To get an idea of an essay which is citation rather than point-led, this is a genuine excerpt from my A-Level History coursework:

The historian Creveld acknowledges this point by noting “there was not a single cartridge fired that had not been manufactured before the outbreak of hostilities”, lending credence to Blackbourn’s interpretation. Furthermore, Taylor’s conviction in the importance of railways furthers this argument, as he states “the basis for their [Prussia’s] success was their railway timetable”. Indeed, Prussia’s ability to utilise railways when mobilizing was only made possible due to the economic boom, which the historian Henderson acknowledges in saying “existing industries – coal and iron – were expanded to supply fuel and rails for the railways”, which further supports the importance of economic strength in securing victory. However, the efficacy of the railways is questioned by Sheehan, as he argues that whilst “railroads changed the scale as well as the tempo of war”, unfortunately “with this new scale came profound problems of command, communication and control”. The historian Voth lends further weight to this counter-assertion, as he stipulates that “Moltke’s plans to supply his armies in Austria by rail nearly ended in catastrophe”. Moreover, the importance of economic growth in developing Prussia’s munitions industry is further refuted by Taylor, who testifies that “Prussia was inferior in the application of industry in war”. This viewpoint is corroborated by Sheehan…

Yikes.

Commandment 9: Don’t obsess over definitions


I occasionally see essays that feel the need to define every term in the question, which reads like something out of Clockwork Orange. As a general rule, you only need to define a term if it is inherently ambiguous (e.g. ‘neoliberalism’) and if your answer to the question hinges crucially on what you take that word to mean.

For instance, my answer to the question “Is realist theory obsolete?” clearly depends on my interpretation of the word “obsolete”. I could plausibly take it to mean one of three things:
  1. It is no longer written about/discussed
  2. It is no longer useful to scholars
  3. It is no longer a tenable explanation of how the world works
1) is patently false; you would have a hard time defending 2); but 3) is a very reasonable view that many IR scholars hold. Clearly, which definition I choose makes a massive difference to my argument.

Commandment 10: Write authoritatively


Imagine you are writing for the Economist or the Financial Times. You wouldn’t read phrases like “in my opinion” or “it seems that” in these magazines and for good reason – they depower the author and make them seem hesitant and non-committal. So don’t write phrases like this in your essay – be authoritative![1]  

It is also useful to bear in mind George Orwell’s timeless ‘six elementary rules’ of the English Language:
  1. Never use a metaphor, simile, or other figure of speech which you are used to seeing in print
  2. Never use a long word where a short word will do
  3. If it is possible to cut a word out, always cut it out
  4. Never use the passive where you can use the active
  5. Never use a foreign phrase, a scientific word, or a jargon word if you can think of an everyday English equivalent
  6. Break any of these rules sooner than say anything outright barbarous
If you want an example of good academic prose, look no further than the writings of Robert Nozick. I am definitely more of a champagne-Rawlsian, but the way in which Nozick captures the mind through prose is second-to-none.

The Good, the Bad and the Ugly of Essay Introductions


I also nicked this idea from Professor Hargrave, but thought it would be fun to re-do using my crap first year undergraduate Philosophy essays as fodder. Without further ado, here are the good, the bad and the ugly of essay introductions, using genuine essays I wrote during my undergraduate degree. 


The Ugly


 Are the Mind and Body Two Separate Substances?

The question of mind-body dualism has been a prominent area of philosophical discussion throughout the course of human history, and still retains the same relevance and controversy today as it has done in the past. In order to establish a thorough conclusion to the question of this essay, it is necessary to examine the main viewpoints that have been presented on this topic, and compare their validity against each other.

Commentary
Oh dear. Platitudinous first sentence, and the second for that matter. In fact, this introduction isn’t really saying anything. For instance, putting myself in the shoes of a reader, I have no idea what the question is and no idea what line of argument the writer is taking.

If you think about it, this introduction could be about any essay. Replace ‘mind-body dualism’ and ‘philosophical’ with ‘Darwinian evolution’ and ‘biological’ and this could literally be copy and pasted into another essay, which is never a good sign.

In short: this is a waste of paper – but in fairness, this was my first ever essay at Oxford.

What Constitutes Personal Identity?

The idea of defining our personal identity and capturing the real me invokes a greater sense of importance than many other areas of general philosophy, and has therefore always been a prominent area of discussion. In this essay, I am going to argue that we must tear up all our naturally held assumption about ourselves; we cannot be certain that our identity dies when our body dies; we cannot be certain that our identity implies an owner (or a self), and we cannot even assume we retain the same identity throughout our bodily lives.

Commentary
This introduction is slightly better, but still majorly deficient. Again, we have a platitudinous first sentence. While there is a semblance of a thesis statement in here, it is very woolly and ill-defined. If I had to play the ‘Guess the Title’ game, I would probably say something along the lines of “write down a list of vague, disconnected thoughts loosely related to the topic of personal identity”. This was my second ever essay – baby steps!

The Bad


 Do We Have Free Will?

At first glance, the issue of free will can often be dismissed as trivial, and the preceding arguments embarrassingly one-sided. Surely we can be certain that free will exists? However, as I will demonstrate at the beginning of this essay, the doctrine of hard determinism puts forward a strong argument to suggest that the causal nature of our behaviour renders us unable to influence both the present and the future, and thus makes the notion of freedom highly questionable. I will then show that any attempts to bypass or transcend determinism are either impossible or insufficient in proving we have free will, and thus the only way to show we are really free is to somehow reconcile determinism with freedom. After dismissing a few of the more common arguments for compatibilism, I will finally settle on an argument proving that freedom and determinism can indeed coexist, and hence show that human beings do have free will. 

Commentary
Things are getting better slowly. There are some good things about this introduction: first, we have a thesis statement, highlighted in bold. Anyone reading this will know exactly that this is an essay defending freedom of the will. There is also an (albeit slightly unclear) indication of how I’ll go about doing this: by reconciling determinism with freedom.

Still, things could be improved. The first sentence doesn’t really make sense – I sort of know what I must have meant at the time but it is very woolly and leaves a bad first impression. The second sentence is not much better: the hanging question is insipid and disempowering.

A final point: when confronting questions that have occupied human minds for millennia, a good rule of thumb is to add a healthy dose of nuance and not come down too heavily on one side. As much as I would have liked, it is unlikely that my 19-year-old self was able to crack the problem once and for all.

Should principles of justice be chosen behind a ‘veil of ignorance’?

The veil of ignorance is a methodological thought experiment first introduced by John Rawls in his Theory of Justice. In this essay, I will ultimately argue that we should choose principles of justice from behind a veil of ignorance, since, as I shall explain in a subsequent paragraph, two fundamental requirements of justice are best ensured by assuming ignorance. I will then explore three major objections to Rawls’s device, examining each one in turn before explaining why I think they are all fundamentally misguided.

Commentary
Again, we have an identifiable thesis statement, and this time the opening sentence goes some way in clarifying and introducing the question. However, this introduction is too short, even for an exam essay. The reader is left with no idea of how I will defend my thesis statement which may cause confusion later down the line. At the very least, I should have stated what these objections are and given a short indication of how I will go about countering them.

The Good


 ‘In the final analysis, a state’s foreign policy choices will be determined by whichever domestic interest groups are the strongest.’ Do you agree?

The role of domestic interest groups in foreign policy choices embodies the puzzling tangle between domestic and international politics that the IR theorist must contend with. In approaching this puzzle, I will adopt Puttnam’s two-tiered strategic framework to assess the role of constitutive actors and exogenous forces in shaping foreign policy direction. Using this framework, I will argue that foreign policy choices are not determined by whichever domestic interest groups are the strongest, since these interest groups may be cleverly circumvented by astute Level I negotiators. Moreover, the fact that Level I negotiators often pursue policies that are overtly at odds with the interests of their domestic constituents shows that it is they, and not domestic interest groups, who are the most important actors in the conduct of foreign policy. This must be caveated by acknowledging the importance of exogenous variables such as ideology, since these can influence the behaviour of all agents within Puttnam’s underlying framework.

Commentary
This essay has a clear thesis statement and a good explanation of how I will go about defending it. Referencing Puttnam’s model sets out a logical framework from the outset which will make the whole essay clearer and more structured. Not all essays warrant a framework, but in topics as broad as ‘foreign policy choices’ introducing one can make unwieldy questions a lot more manageable. The added nuance at the end is also a nice touch – if done well, self-critiquing a framework/paradigm can be a powerful way to signal both intellectual rigour and command of material.

What explains the continued salience of ethnicity in so many African political systems?

Garrick Bailey defines an ethnic group as “a named social category or people based on perceptions of shared social experience”. The emphasis on perception implies that Bailey interprets ethnicity as a social construct, as opposed to a purely descriptive account of different physiological characteristics. For the purpose of this essay, I will view the concept of ethnicity under similar guises. Thus, since ethnicity is a construct, to explain its continued salience we must examine the motivations people had for creating and sustaining these identities. To do this, I will distinguish between the motivations of two types of agents: elites and commoners (non-elites). Embodied in this distinction is an important message: ethnic divides are not simply ‘top-down’ or ‘bottom-up’ impositions; they are embraced and perpetuated by elites and commoners alike. Thus, this essay focuses on two questions:
  1. What motivation did elites have to appeal to ethnicity?
  2. What motivation did commoners have to accept and advance these ethnic divides?
Using this framework, I will argue that the continued salience of ethnicity is attributable to elites and commoners using ethnicity to pursue two aims simultaneously: elites used ethnicity as a means to gain or consolidate their power; commoners used ethnicity as a means to retain a sense of personal identity and to enhance their claim to state resources. This argument will be made in reference to two case studies in particular: Rwanda and Kenya.

Commentary
This, I think, is about as good as my introductions ever got. There is a strong thesis statement and a clear indication of how I will go about defending it. The definition given at the start seems both reasonable and necessary: ethnicity is a murky and controversial topic so defining it from the outset helps avoid any ambiguities later. Finally, the reference to Rwanda and Kenya at the end signals that the essay will be evidence-based and helpfully narrows the scope.

References


The two guides written by Professor Hargrave can be accessed here:

https://www.st-andrews.ac.uk/~mnat/~ball0888/oxfordopen/Rodin.htm
https://www.st-andrews.ac.uk/~mnat/~ball0888/oxfordopen/paragraphs.htm

This guide is also available in PDF format, available here:

https://www.dropbox.com/s/2bc6hx035hi3moq/How%20to%20Write%20an%20Essay.pdf?dl=0

[1] There is a good argument to be made that rewarding authoritative language in essays unfairly biases men over women. I am certainly sympathetic to this argument: but unless something changes, my advice will remain to make a conscious effort to write authoritatively.
Posted by at 3 comments:

Sunday, April 28, 2019

Colonialism, Race and Education Inequality

In the words of Engerman and Sokolof, colonialism is back – not in practice, thankfully – but in the minds of scholars as an important ‘deep determinant’ of contemporary development patterns. This post focuses on just one of the many long-term effects of colonialism: its transformation of the construct of race and the effect that has had on education inequality and institutional development.

A Brief History of Race in Africa


Colonialism did not ‘invent’ ethnicity. Many ethnic groups – such as the Ashanti in Ghana and the Baganda in Uganda – have roots that long pre-date European occupation. Back then, ethnicity was an extremely fluid construct: socio-cultural boundaries were marked by fuzziness and flexibility, and many Africans existed within a reality of multiple, overlapping and alternative collective identities. This was largely driven by the scourges of war, famine and disease, which would routinely destroy and dislocate entire communities.
By delineating Africans based on birth and defining their relationship with colonial state in outwardly ethnic terms, colonialism made the concept of race both inflexible and politicised. This was done for several reasons. First, inventing the notion of ‘tribes’ proved to be a costless and efficient way of locating, demarcating and counting the population of a colony for purposes of surveillance and control. Tribal divisions also made it easier for administrators to monitor their local ‘Big Men’: indigenous rulers who served as intermediaries between the colonial state and the wider populace. These relationships were defined by patrimonial logic: colonial elites would keep local Big Men happy through the dispensing of resources who would in turn pacify ethnic constituents by similar means.

The notion of tribalism was also supported by genuine ‘scientific’ beliefs at the time that Africans were naturally tribal and that some tribes were inherently superior to others. For instance, in Rwanda, Belgian occupiers viewed the Tutsi as ‘less African’ than the Hutu and privileged their status accordingly.

Figure 1: The ‘Scientific’ Demarcation of Race in Rwanda

Source: The New Times

The net results of this was a power structure that categorised and prioritised Africans in avowedly ethnic terms. Education was a popular vehicle to deploy this favouritism: for example, in Burundi and Rwanda, Western-style education was reserved almost exclusively for the Tutsi; in Uganda the Baganda were educated at the expense of the Bonyoro, while in Nigeria Christian missionaries set up schools in the Igbo South whilst leaving the Hausa North virtually untouched (though this was for religious as well as ethnic reasons). Largely by design, colonial administrators fomented a new indigenous ‘super-class’ which outwardly reflected the racial biases manifest in colonial rule.

The Stickiness of Colonial Institutions


At independence, inaugural elites – themselves largely beneficiaries of these power structures – found it expedient to maintain rather than dismantle them. By emphasising ethnic divides, they could consolidate support within their ethnic ‘in-group’ while disempowering the rest. Ethnicity proved such a useful ‘mobilising’ tool because it could coexist with other types of consciousness, papering over traditional cleavages like age, religion and class.

As before, education proved an effective means to exhibit ethnic favouritism. Not only were investments in education ‘visible’ to the populace, they also safeguarded the dominance of a group between generations: by deciding who to educate today, elites could effectively choose who would rule tomorrow.

In many countries, the endurance of this logic led to a remarkable stickiness of racial hierarchies. Educational attainment amongst the Brahmin in India, the Igbo/Yoruba in Nigeria and the Baganda in Uganda continues to outstrip those of their ethnic neighbours. In each case, contemporary dominance can largely be traced to the whims of colonial favouritism.

Even in countries where ethnic groups exhibit greater mobility, education policy remains shaped by ethnic considerations. As Franck and Rainer (2009) show, the education enrolment of a particular ethnic group increases when they have 'their man’ in power.

This discussion points to an important consequence of colonialism: by creating communities who were greatly advantaged in terms of human capital and legal rights, colonial administrators left behind societies that were grossly unequal at the outset. This incipient inequality greatly inhibited the development of inclusive political institutions, as the narrow group of elites who inherited power were heavily incentivised to maintain institutions that worked to their advantage (Engerman and Sokolof, 2005).

Final Words


A final word on colonialism. Imagine a spectrum on which theories of colonialism can be placed. At each end are two ridiculously polemical positions: on the left sits the view that colonialism has made no difference to contemporary African development; on the right sits the view that colonialism explains all idiosyncrasies of modern African states.


Clearly, any position too close to the left of this spectrum is untenable – this post has shown just one (of many) long-term consequences of colonialism on economic and institutional outcomes in Africa. However, we should also be wary of arguments that sit too close to the right. Deferring to colonialism as the ultimate cause of everything does a poor job in explaining contemporary differences between Kenya and Tanzania: two countries with ostensibly similar colonial experiences and initial conditions, yet with vastly divergent postcolonial economic, political and social histories. It also robs Africans of agency. Colonialism did not set in motion an unerring chain of events irresponsive to the will of incumbent elites – such a view uniquely misconstrues African leaders as passive objects of manipulation rather than active shapers of history.

Why am I saying this? It is not to exonerate colonialism; rather, it is to undermine the fatalistic idea that colonialism has doomed Africa to underdevelopment. This is not just alarmingly defeatist – it is also patently untrue, as demonstrated by development successes in Botswana, Mauritius and more recently in Ethiopia and Rwanda. Making sense of these necessitates a balanced perspective: we must remain sensitive to the long shadows cast by history while at the same time avoiding overly-deterministic modes of thinking which under-value prospects of future development.

Word count: 984 words

References

  • Engerman and Sokolof (2005), ‘Colonialism, Inequality and Long-Run Paths of Development’
  • Franck and Rainer (2009), ‘Does the Leader’s Ethnicity Matter? Ethnic Favouritism, Education and Health in Sub-Saharan Africa’
Posted by at No comments:

Monday, April 22, 2019

Making Schools Safe

Open any op-ed, journal article or Econ Twitter thread about education delivery in developing countries and there is a very good chance it will centre on one of two issues: access (how to get kids in school?) or quality (how to make sure kids are learning in school?). Of course, there is good reason for this – if kids aren’t in school (or if they aren’t being taught if they are) then it is unlikely they will be learning key competencies such as reading, writing and numeracy – with devastating consequences for economic growth and individual flourishing. However, in making these issues cardinal, we run the risk of blind-sighting ourselves to issues that are at once simpler yet more fundamental.

Accidents Waiting to Happen


Consider the issue of school safety. If the government mandates a policy of compulsory schooling – which compels children to meet at the same place each day, five days a week – then we had better make damn sure that these places are safe for children to congregate. Safety is multi-faceted. Perhaps the most obvious component is the physical safety afforded by school infrastructure – are schoolchildren at risk of falling debris, open sewage or fire?

It is easy to take these questions for granted – I certainly did before I stepped foot in a school in Zanzibar. But consider these photos I took last week when visiting a pre-primary school – a school designed for children of four and five years of age. These children – barely out of nappies – play largely unsupervised in areas riddled with decaying ceilings, open gutters and precipitous drops unguarded by railing. The fact that no learning is taking place in this school seems secondary to the fact that there is an accident seemingly waiting to happen every single day.

Figure 1: A Pre-Primary School in Zanzibar


Of course, accidents do happen, and the results can be devastating. In 2004, a primary school in the Indian state of Tamil Nadu caught fire due to an illegal thatched roof. 91 primary schoolchildren died that day – a tragedy of a scale that's difficult to comprehend. It is hard to think of a commensurate failure in education delivery.

The Threat Posed by Teachers


Unsafe buildings are not the only threat posed to children: a more frequent and intentional threat is posed by aggressive, lecherous and untrained members of staff. A 2011 study by UNICEF in Zanzibar revealed that over 70% of students reported having experienced some form of physical or humiliating punishment at the hands of educators – a fact which has devastating impacts on children’s incentives to attend school. Despite large decreases in recent years, approximately 69 countries still allow corporal punishment in schools, including parts of the United States, some Australian states and large swathes of Africa and Asia (Gershoff, 2017).

More shockingly, sexual abuse perpetrated by teachers remains ubiquitous, particularly in countries towards the bottom of the income distribution. In Liberia, 1 in 4 primary school age children admitted to having had sex with a teacher, and 1 in 3 admitted to having sex with a member of staff (Steiner et al., 2018). These grim statistics are reflected in Kerala, India, with 21% of schoolchildren reporting having been sexually abused in a school environment over the past year, with surprisingly far more boys reporting abuse (29.5%) than girls (6.2%) (Kumar et al., 2017).

Issues like these strike at our emotions in a way unlike issues related to enrolment or pedagogy. Despite this, they comprise a much smaller part of the overall conversation. If you wanted to find out what works in raising attendance or test scores in schools, there are literally hundreds of studies that have tested all manner of interventions in all manner of contexts. However, if you wanted to assess the prevalence of sexual violence or unsafe infrastructure in schools – and what can be done to improve it – you would struggle to find anything like the body of evidence behind access and quality. As Justin Sandefur notes, there is a strange paradox at play here: what we choose to focus our research on seems disproportionate to what we care about as parents, siblings and people. I'm not a parent - but if I were I imagine I'd care more about whether my child was vulnerable to sexual assault than I would about whether my child learnt anything today in school. 

In a scintillating open letter to a new Minister for Education, Paul Skidmore recommends immediately conducting an independent, internationally-recognised literacy and numeracy assessment to see whether children are learning anything of value in school. This is undoubtedly a good idea – but does not dig deep enough. As well as asking whether children are learning anything in schools, a more fundamental question would be: are schools safe places for children? Such a question would necessitate a school and teacher audit to go alongside any measure of basic competency.

Likewise, academic researchers should recalibrate their focus and make a more concerted effort to identify ways in which unsafe practices in schools can be reduced or eliminated. Recent work by the Behavioural Insights Team which assesses different strategies for changing teachers’ attitudes towards corporal punishment in the Nyarugusu Refugee Camp in Tanzania makes for a promising start (Behavioural Insights Team, 2017).

The benefits of this would be manifold. Crucially, making schools safer will have positive knock-on effects on the two topics currently in the limelight: access and quality. It doesn't take a PhD in Economics to figure out why in environments where they can be hurt and abused, children may be less engaged in class and less willing to show up altogether. More fundamentally though, eliminating unsafe practices can ensure that child's rights are better protected - a benefit which is difficult to overstate.

Word count: 967 words

References

  • Behavioural Insights Team (2017), ‘Changing Teachers’ Attitudes Towards Corporal Punishment’
  • Gershoff (2017), ‘School Corporal Punishment in Global Perspective: Prevalence, Outcomes, and Efforts at Intervention’
  • Kumar et al. (2017), ‘Prevalence of Child Abuse in School Environment in Kerala, India: An ICAST-CI Based Survey’
  • Steiner et al. (2018), ‘Sexual Violence of Liberian School Age Students: An Investigation of Perpetration, Gender, and Forms of Abuse’
Posted by at 26 comments:

Friday, March 8, 2019

Uncovering Causality II: Instrumental Variables

While randomised controlled trials may be the hottest tool in the econometricians toolbox, the most contentious, capricious but potentially brilliant tool is the instrumental variable (IV). A bad instrument will never see the light of day; but a good instrument can elevate the statistician from humble programmer to creative genius, catapulting the researcher to academic stardom.

Instrumental Variables


As is often the case, instrumental variables are best explained by means of example. Suppose I hope to estimate the causal effect of class sizes on student test scores. If I collect data on both, I may get a scatterplot that looks something like this:


What is the problem with extrapolating a causal relationship from this result? As before, there may be ‘confounding’ variables that influence both X and Y, obscuring the true relationship between the two. For example, low class sizes may be correlated with school wealth (e.g. richer schools can afford lower student-classroom ratios) which also affects test scores through other ways (e.g. better textbooks and non-classroom facilities).

One way to avoid this problem is to identify a variable/instrument (Z) which i) affects class size (X) and ii) is not correlated with anything else that affects test scores (Y). In econometrics jargon, i) is known as the relevance requirement and ii) is known as the exogeneity requirement. If both criteria are met, we can effectively ‘isolate’ the random variation in X and identify a true causal effect.




For example, in the context of class sizes and test scores, a candidate instrument could be the occurrence of earthquakes. In places like California, schools sometimes shut down due to infrastructure damage caused by periodic earthquakes. While the schools are being repaired, affected students are temporarily transferred to undamaged nearby schools, causing a sudden increase in class sizes.

We could also argue that earthquakes do not affect test scores (Y) in any other way apart from via their effect on class sizes (X). They do not change the school syllabus, they do not change the timing of the exams and they do not change the underlying ability of students. Assuming this argument is correct, we are left with a nice natural experiment: we have a random, isolated increase in class sizes (X) that allows us to identify its independent contribution to test scores (Y).

The Exogeneity Requirement: An Achilles Heel


Unfortunately, plausible instruments are difficult to identify. The relevance requirement is fairly easy to meet: it is not difficult to find variables that are correlated with the treatment (e.g. class sizes). However, the exogeneity requirement represents a perennial stumbling block. In my previous example, the exogeneity requirement is almost certainly not met. We can theorise many potential effects of earthquakes on test scores besides its impact on class sizes: for instance, earthquakes can have localised income effects as well as psychological effects on affected students, both of which effect test results. Alas, academic stardom will have to wait.

To make matters worse, the standards for an acceptable IV have generally become a lot more stringent. This is partly due to the emergence of new causal identification techniques that have raised the bar when it comes to credible identification strategies. The graph below gives a nice illustration of these new econometric techniques: some of which I may discuss in future posts.


Source: The Economist

As a result, some of the most iconic IVs appear unable to bear the weight of contemporary scrutiny. For example, perhaps the most famous IV in all of Economics is the use of settler (colonial) mortality rates to isolate the causal effect of institutions on growth (Acemoglu and Robinson, 2001). The authors theorised that in places where settler mortality rates were high (e.g. Africa, South America), settlers were more likely set up ‘extractive’ institutions in which a small group of individuals exploited the rest of the population. Conversely, in places where settler mortality rates were low (e.g. North America), settlers were more likely to establish trade relationships and ‘inclusive’ institutions, which included many people in the process of governing.

The authors go on to argue that these initial conditions greatly influenced the quality of economic and political arrangements today, due to the ‘sticky’ nature of institutions. Hence, the relevance requirement is met: Z (settler mortality rates) affect X (modern institutions). Then comes the leap of faith: the authors claim that historic settler mortality rates (Z) have not affected modern day growth (Y) except via their effect on the institution’s they brought about (X).

Needless to say, the validity of this instrument has received a hammering in recent years. For instance, settler mortality rates (Z) seem correlated with historic disease environment which is itself correlated with modern day growth (Y). These criticisms appear unanswerable: I remember a senior faculty member at Oxford once telling me that the IV would probably not get published today. Still, the above example serves as a testament to human creativity and to the rapid advancement of scientific standards.

All Hope is Not Lost


With this being said, instrumental variables are not completely defunct. One of the best IVs I have seen concerns estimating the effect of incarceration length (X) on subsequent employment outcomes (Y) (Kling, 2006). The immediate difficulty in estimating this effect is that people who get longer prison sentences are likely to be different from those who get shorter sentences, in ways that matter for future labour earnings.

To get around this, we can use the fact that judges are randomly assigned to cases, and that some judges appear systematically harsher at sentencing than others. This appears to meet the conditions of a valid instrument: the particular judge (Z) a defendant gets certainly effects his/her likely incarceration length (X), but is unlikely to be correlated with any other factor which effects his/her future earnings (Y).

This, I think, represents a plausible identification strategy with many avenues for future research (see this blog post). Though standards have risen, instrumental variables are certainly not obsolete. But remember: IV regression is only as good as the instrument itself!

Word count: 1,000 words

References

Posted by at 1 comment:
Subscribe to: Posts (Atom)