Science of Reading

As a parent of young children, I’ve been trying to figure out how kids actually learn to read. You hear all these conflicting terms: "phonics," "whole language," "sight words." Schools have always talked about "balanced literacy" and encouraging a "love of reading."

But lately, I’ve been hearing a lot about the "Science of Reading" (SoR). It’s everywhere, from news articles to parent forums. It turns out this isn't a new fad, but a huge body of research. Diving into it has been a real eye-opener for me.

What is the "Science of Reading" (SoR)?

First off, it’s not a curriculum or a single program. "Science of Reading" is a term for a massive, interdisciplinary body of research from the last 50+ years. It pulls from cognitive psychology, neuroscience, and education, and it basically explains how the human brain learns to read.

The biggest takeaway? Reading is not a natural process.

We're wired to learn to speak, but our brains are not wired to read. To become readers, we have to build new pathways in our brain that connect the sounds of our language to the letters on a page.

So what was wrong with the old way? The "Reading Wars."

For decades, many schools used an approach called "Balanced Literacy" or "Whole Language."

A key part of this was a strategy called "three-cueing" (or MSV: Meaning, Structure, Visual). If a kid got stuck on a word, they were taught to:

1. Look at the picture (Visual cue)

2. Think about what word would make sense in the sentence (Meaning cue)

3. Look at the first letter and guess (Structure/Visual cue)

The "Sold a Story" podcast [1] really exposed the problem: scientific research shows that skilled readers do the exact opposite. They don't guess. They decode the word instantly by sounding it out. Teaching kids to "cue" is actually teaching them the habits of poor readers.

What does SoR say we should do? The Five Pillars.

So if guessing is bad, what's good? The research points to five key components that need to be taught explicitly and systematically.

(1) Phonemic Awareness This is all about sounds. It's the ability to hear and manipulate the individual sounds in spoken words. Before kids can read "cat," they need to be able to hear that it's made of three sounds: /k/ /a/ /t/.

(2) Phonics This is the part everyone's talking about. It’s the explicit connection between letters (graphemes) and sounds (phonemes). It’s the "code." This isn't just a 20-minute drill; it's the core of early reading.

(3) Fluency This is the bridge from decoding to understanding. It’s the ability to read a text accurately, at a good pace, and with expression. You can't understand a paragraph if you're struggling with every single word.

(4) Vocabulary This one's simple: kids need to know what words mean. You can sound out "gargantuan" perfectly, but if you don't know what it means, you're not comprehending. This is built by talking, and especially by being read to!

(5) Comprehension This is the ultimate goal. It's thinking about and understanding what you've read. The SoR model shows that comprehension is the product of the other skills: you can't comprehend the text if you can't decode the words or don't know what they mean.

Conclusion?

It’s not about "killing the love of reading" with boring drills. It's about giving kids the actual tools they need to be able to read in the first place.

My takeaway is that explicit phonics instruction isn't optional. It's the foundation.

New cigarette taxes in Malaysia – will they work?

So, big news for tobacco control last week! The Malaysian government announced a tax hike on cigarettes – the first in 10 years.

This is a policy reform I'd been working on (with a local think tank), so it's cool to see progress.

The existing excise tax is RM0.40 per stick (RM8 per pack). Starting November 2025, the tax rate be RM0.42 per stick. (Besides excise tax, there's also sales tax and import tax.)

Why tax tobacco? Reduce smoking

Well, many Malaysians (around 1 in 5) still smoke (not good!). So reduce smoking, one thing we can do is keep cigarette prices high. And the way to do this? Tax cigarettes.

Additionally – to stop cigarettes becoming affordable – taxes needs to keep up with inflation and rising income.

That's the idea. But will it work?

What about illicit cigarettes?

Some ask "Would people just switch to cheaper alternatives? Or illicit untaxed cigarettes?"

A reasonable question, especially in Malaysia where illicit cigarettes supposedly make up over half of smoking volume.

(A caveat on illicit trade data: only data generated by the tobacco industry is regularly available and updated. Obviously a concern; however independent estimates also do suggest illicit cigarettes are a big problem in Malaysia.)

So will the new tax work? Three considerations

Well, it's hard to say for sure. But here are three considerations that give us some idea:

(1) An obvious place to start: Malaysian historical data.

Before 2015, Malaysia's cigarette excise tax had steadily increased. What actually happened as a result?

• Firstly, consumption continued to decline (though not necessarily entirely due to higher taxes) (see Figure 1 ).

• Next – despite this declining consumption – tobacco tax revenues increased (see Figure 2). Between 2016-2019, when taxes were stagnant, tax revenues decreased.

• And lastly, illicit cigarette volume did increase – especially between 2008-2009 and between 2015-2016 (see the chart here ).

Let's park these observations for now.

(2) Second consideration: the research on illicit trade in Malaysia.

There are studies / models examining how tobacco tax (or prices) affects smoking consumption and tax revenues – and how illicit trade interacts. What are some key insights?

• Firstly, higher taxes do reduce smoking. This is somewhat obvious, and confirmed by various studies (like this review of Southeast Asia evidence ).

• Next, higher taxes also generate higher tax revenue – even though people will smoke less. This is due to price inelasticity.

• And lastly, the main driver of illicit cigarettes is not taxes – instead it is corruption. This study shows many countries have tobacco taxes/prices higher than Malaysia, but do not have high levels of illicit trade. In other words, higher taxes won't lead to more illicit trade – as long as government improves enforcement. (Good news: this is exactly what the Malaysian government has been doing in the past couple years – strengthening enforcement)

Let's keep this in mind, and look at a third consideration.

(3) The new tax hike is not that large.

The actual policy decision is fairly modest. An additional 2 sen/stick would probably lead to a roughly ~2-3% price increase.

As such, whatever effect the tax will have – on smoking reduction, switching to illicit trade, and tax revenues – will probably be modest as well.

Conclusion? A prediction.

With all these points in mind, I think a reasonable prediction would be:

(a) Illicit trade will continue to decline. This is due to increasingly strong enforcement, and because the new tax is too modest to incentive switching to (cheaper) illicit cigarettes.

(b) Smoking prevalence will continue to decline steadily – but not necessarily total smoking. The downward trend of smoking prevalence (the percentage of people smoking) will continue, due to various factors (including that fewer youths are taking up smoking). However, total smoking (the absolute number of smokers) may not reduce much, as the modest tax hike will have some – but limited – cessation impact.

(c) Tax revenue will increase slightly. If we assume little change to smoking habits, that means a similar number of cigarettes will be purchased, but with an additional tax paid. This translates to around ~RM100 million/year (we'll save the math for another post) in extra revenue – not a huge some, but not negligible either.

What next?

So will the tax work? Well, at least it'll work at the margins. And additionally, it could be a good first step – toward regular increments in the coming years.

Would be interesting to see the data, as it becomes available.

Should stalking victims do "risk assessments"?

I worked in an NGO dealing with gender-based violence in Malaysia. Advising victims to do "risk assessments" was common practice. This made sense, but I was curious to see what the research said on this.

In short yes, stalking victims should do risk assessments (phew!).

What is a risk assessment?

Essentially "risk assessments" involve examining the details of a stalking case (e.g. is the stalker a former partner, does the stalker have a weapon, etc.) to determine how dangerous the situation is. This in turn helps victims (and authorities) decide what precautions to take.

Why are they helpful?

Basically, risk assessments work because there are factors (that have been shown to be) associated with increased danger.

For example, if a stalker is monitoring the victim's daily activities, physical violence and homicide are more likely (according to this article, citing other articles). And so the victim (and authorities) may want to be on high alert.

This article has a useful summary of factors that increase the risk of violence in stalking cases.

Do protection orders (PO) work for stalking victims?

I worked on a campaign to enact stalking laws in Malaysia. A key thing we advocated for was creating protection orders (or restraining orders).

This seem a logical "good practice", but what does the research say?

What I found

Well, the first thing I found: there doesn't seem to be much evidence available. This is understandable – as you can imagine, it's hard to conduct experiments to test protection orders.

I did find this meta-analysis, that suggests POs but aren't enough to keep victims safe – in cases of domestic violence that involve stalking.

This isn't to say that POs aren't useful. In practice, it means that POs must be combined with other interventions.

Key points about the meta-analysis

The meta-analysis examines whether protection orders (PO) are effective in reducing domestic violence. Stalking comes in, as some domestic violence cases involve stalking.

Overall, the analysis finds there is "emerging evidence" that POs reduce "subsequent violence to some extent". However, this is for domestic violence cases that don't involve stalking and a few other factors.

So what about domestic violence cases that do involve stalking? Doesn't look promising. The analysis reports:

"All five studies [included in the meta-analysis] investigating stalking reported stalking to be a significant predictive factor for ongoing PO violations and increased physical, psychological, and sexual violence to victims, as well as poorer mental health and a lower sense of safety."

Why? Potentially because stalking is harder to prove and is taken less urgently by authorities.

The analysis concludes:

"effectiveness of POs [for stalking victims] may only be achieved ... when POs are combined with community-based interventions that offer further education, support, and resources to victims together with greater surveillance of offenders."

A bit on the meta-analysis's methodology:

• The analyses include 25 studies, covering 31,586 participants.

• All studies were US-based, except 1 study (Sweden).

• Most victims in the studies were female, while most offenders were male.

• Of the 25 studies: "19 were case series, 3 used a case control study design, 2 were retrospective cohort studies, and 1 was a prospective cohort study."

How do countries challenge human rights norms? (case of China)

Enjoyed this paper by Alexander Dukalskis. The paper (1) argues that, through the UN Human Rights Council, China aims to change international human rights norms; and (2) proposes four tactics China uses to make these changes.

(The author says this point 2 is his main contribution, as it's been studied less compared to point 1.)

The paper draws evidence from China's voting behaviour at the UN Human Rights Council.

On what norms China challenges:

"the PRC [a] keeps its own record out of the spotlight, [b] is hesitant to vote for resolutions that target one country, [c] advances norms associated with its conceptions of development and multilateralism, and [d] works against norms associated with liberal democracy."

On how China tries to challenge the norms:

"the “how” is characterized by four modes of action: [a] mobilizing like-minded states, [b] real or implied coercion, [c] tactical deception, and [d] repression of critical voices."

A clear, well evidenced, and well argued article.

A few quick reflections:

• It's implied (I think, though not explicitly said) that the current norms are desirable, and conversely China's attempts to challenge norms are undesirable. I'd generally agree, but worth reflecting on some bits. For instance, more emphasis on a "right to development" could be useful, insofar as development can (of course not necessarily) facilitate the realisation of most other (including civil and political) rights. Of course, we can think about whether we can conceptualise a "right" to development (vs seeing development as a means to achieving other rights), but the practical question is: whether focusing more on "development" in human rights discourse is useful or not. The other instances of China's norm challenging seem more clearly (and blatantly) undesirable.

• The fact that China (and probably other states) put much effort in defending it's record and challenging norms suggests that states do feel that norms affect them in some material ways (and perhaps in some less material ways e.g. "honor" / "face"). This could strengthen the case for the usefulness of defending (desirable) international norms. Also, would be interesting to explore what ultimately states like China care about.

• If this analysis were to be done for 2025-2035 or so, how will the findings change? Would states like China care less about (unenforceable) intersectional norms, or perhaps take the opportunity to cement even more changes?

Are historical "Truth Commissions" useful? (Belgium's colonial legacy)

Just read Tine Destrooper's paper critiquing the Belgium Truth Commission.

The author argues, in this particular case, the Commission has many limitations – but still could have "expressive" value (e.g. further socialise struggles for justice, and lead to future change).

The background:

• In 2020 the Belgian Parliament established a "Special Parliamentary Commission" (referred to as a "Truth Commission") to enquire into Belgium’s overseas colonial legacy.

• This Commission is notable -- Belgium is the "first consolidated democracy to establish a truth commission to investigate the historical and ongoing injustices related to overseas colonialism".

I found the paper a helpful commentary on the Commission:

• The author analyses the Commission's establishment, as well as (the largely negative) reaction of various stakeholders -- and offers her own critique.

• The Commissions can play an important "expressive" role -- i.e sending a message to larger society (e.g. cementing values, norms). This is separate from any "accountability" role the Commission may (or may not) have.

• While the author is quite critical of the Commission (e.g. over-ambitious, impossiblly short timelines, insufficient representation of experts), the author also offers reasons to (potentially) be hopeful:

"While acknowledging the pertinence of this concern, I have nevertheless argued that even under the current circumstances, the work of the commission could have a more far-reaching impact than can be gauged at present, if we consider its expressivist potential. It could generate momentum as well as a dynamic of rhetorical entrapment or socialization that could provide breeding grounds for further struggles for justice and thick accountability that may then have more potential to disrupt the status quo, lead to more transformative justice efforts, and challenge epistemic injustice."

Some quick thoughts:

• It seems to me that any exercise like this (a Truth Commission) is bound to be imperfect, due to various constraints (many were alluded to in the paper). Some of these constraints may be technical (e.g. policymaker capacities, time limits, subject matter complexity).

• Other constraints relate to political acceptability (e.g. the scope/ambit of the Commission, commitments following Commission findings). The author makes the point that having government backing is key, as it legitimises a Commission. And so if a society wants to initiate a Commission exercise, the boundaries of the government (at that particular time) are an implicit limit to the Commission's ambition.

• These constraints are not permanent, but there's rarely a "perfect time". And so, I appreciate the author's effort to find a (potential) silver lining.

Should reform campaigners focus on raising awareness? (interesting paper on child criminal responsibility in Australia)

Apparently in (parts of) Australia, children as young as 10 can face criminal trial as adults. That seems instinctively super low. And indeed, around 14 is the norm.

Just read Suzanne Rock and co's article, on the efforts to raise the "minimum age of criminal responsibility" (MACR) in Western Australia.

Wanted to highlight two cool things about this paper:

First cool thing: paper's findings are very "real-world" relevant.

While we chase generalisability (and "seek universal truths"!), policy research must also equally value specificity ("in this specific country or context, what works?").

The paper's findings aren't necessarily novel, nor does it develop a theoretical concept. But it does however, fill a specific knowledge gap, to advance a specific policy challenge in Western Australia.

The policy challenge: If 14 (not 10) is the international norm for the MACR, why haven't Australian policymakers followed suit?

The paper's main argument (paraphrased):

1. Currently the public doesn't know about the MACR. Therefore politicians don't care either.

2. But, if made aware, the public will care A LOT -- and will support raising the MACR. This would in turn make politicians care.

3. Therefore, if campaigners want to increase the MACR from 10 (to say 14), they need to first raise awareness.

This may seem obvious, and indeed prior surveys have shown similar findings. But the paper provides qualitative evidence, that illustrates just how strongly the public may care about raising the MACR -- if they are made aware of the issue (e.g. some respondents used strong words like, the current 10 is "ludicrous").

Second cool thing: paper an excellent "learning" tool.

Besides presenting useful findings, the paper also seems a useful "example paper" for introductory learning/teaching of research, because:

1. The paper lays out it's methodology in detail -- including the sampling, interviews, and analysis. (So it's like a guide for new researchers.)

2. The methodology and subject matter isn't too complex. (So it's accessible and doesn't require much prior background.)

3. There are many pros and cons of the paper's methodology and approach, which the paper does address. (So it enables discussion of research methods.)

Indeed, the authors surely designed the paper with learning/teaching in mind. The paper was written as part of a criminology class -- where students learnt research methods and applied their skills through carrying out the research.

In conclusion

While the paper may not end up with the most academic citations (it might yet!); it could more than compensate through impactful real-world policy and classroom references.