Note: This post is a special year-end issue in my blog series, reflecting on recent policy discussions from a health economics conference in Nepal. “The policy logic behind taxing High-Sugar, Salt, and Fatty (HFSS) foods is compelling and urgent: curb consumption, expand fiscal space, and fund health to fight non-communicable diseases (NCDs).” The public health framing was clear, and the intent, unquestionably serious. Yet, as discussions unfolded, I was struck by a gap. The debate…

Imagine two researchers working with the same dataset to answer the same high-stakes policy question: Does a new vocational training program helps unemployed adults earn more? Both are well-trained. Both follow standard practices. Both make what they believe are reasonable, defensible choices in their models. Yet when they present their results, the headline numbers don’t match. One reports a strong positive effect, the other finds only a modest improvement, and a third version of the…

In my previous post, we discussed the signal — the functional form. We saw how getting the shape of a relationship wrong (for example, assuming a straight line when the true effect is curved) can lead to biased policy estimates. Today, we turn to the other half of the story: the noise or the stochastic component of a model.  In this post, we will unpack what this “noise” really means and why the specific distributional…

It has been a while since my last post—since March, to be exact. Life threw me an unexpected challenge: a detached retina. The fix? Surgery where surgeons injected silicone oil into my eye to press the retina back into place. But the oil does not work magic on its own. For it to work, I had to maintain a face-down position for several hours each day. That way, gravity could help the oil push the…

When a new programme or intervention is implemented, we all want to know: “Did it work? ” A seemingly straightforward way to find it out is to compare outcomes before and after. If GPA scores go up, health improves, or incomes rise, many mistakenly assume this is clear proof of impact. But is it? In my last blog, we discussed why correlation doesn’t imply causation. This time, we will focus on a popular but often…

Did you know that countries with higher rates of internet penetration tend to have longer lifespans? Or that nations with larger militaries often have healthier populations? At first glance, these patterns might seem logical. After all, internet access could improve healthcare access, and a strong military might indicate a stable government. But does this mean one causes the other? Or are we mistaking correlation for causation? Statements like “increased internet access leads to longer lifespans”…

Imagine planning a household survey in a rural village where access to a reliable and comprehensive sampling frame is limited or logistically challenging. OpenStreetMap may be useful in this context but the dataset is often riddled with inconsistencies. In my previous blog, we explored a method to extract households using OSM tags to guide sampling – an effective strategy in well-mapped areas. But what happens when these tags are inconsistent or missing altogether? This time,…

Have you ever found yourself staring at a blank map, wondering how to select the right households for your research? Field researchers often face this challenge, especially when a reliable sampling frame is scarce. One common workaround is a crude method where enumerators would spin a bottle to determine a walking direction and then select households at regular intervals along the way. While this method may seem practical, it is far from precise. Selection bias…