When should I quantify? – Quantification – Part 4

Every 48 seconds someone dies of malaria. Every one of those deaths is a human being with passions and loves and feelings. When we talk about quantifying our impact on the world it is important not to forget what those numbers mean. They mean people. Every single year of happy life we can give, is a joyful thing to the person living it. This is ultimately why we want to have as much impact as possible. Because more people living happy joyful lives is better than fewer.


Image source

The last 3 posts have been about quantifying our impact on the world and analysing our effect with numerical methods. But general theories don’t help people with malaria. Their specific applications do. This post is a checklist to run through to decide if you should use quantitative methods on your goal. I will run through the example of malaria treatment, but the checklist is general.

Step 1: What is the goal?

Before you begin any project it is important to be very clear about what you are trying to do and why. This way you don’t end up like the fool in every story about a genie, who gets exactly what he wishes for but not what he wants.

If you want to get rid of malaria remember that you want this because malaria is a terrible disease that kills people. If you ever find yourself with some course of action that could eradicate malaria but at the cost of many many lives it is important that you remember your real goal. The ends often justify the means, but you can only be sure of that if you remember what your true ends are.

Step 2: Can the goal be quantified?

Ask yourself if the goal is of a type that can probably be measured well. A good way to think about this is to imagine two different states of the world 5 years from now. One where you have been very successful, and one where you haven’t. How would you prove to a questioning child which one was which? More often than not you’ll point to some thing in the world that there is more of.

Imagine two worlds. One where I spend the whole 5 years working in a charity shop and eventually raise £100,000 to buy malaria nets in the UK and ship them over to Uganda. One where I become a research chemist and discover a really cheap chemical that can be sprayed from the air and which kills only malaria carrying mosquitos and I cure the disease for good. In the first example I’ve bought maybe 20,000 nets and saved about 100 lives, in the second example I’ve saved 655,000 lives in the first year alone.

The second example is much better, but why? It must be because there aren’t as many people dead. This focus on “how will I know if I have done well” can tell us which of three types of goal we have:
* Inherently numerical goals, e.g. “get as much money as possible”.
* Goals that can be put in numerical terms only at some loss of detail e.g. “make the lives of people with malaria better, which I will approximate as meaning more QALYs
* Goals that cannot be put in numerical terms at all e.g. “produce the best work of art ever”

If your goal is in the first category then clearly quantitative methods should be considered. If your goal is in the last category then stop here, there is clearly no hope for quantitative methods.

The example I am thinking of falls in the middle category. If this is the case then quantitative methods should be considered, but with caution. They may be very effective, but they may lead us to bad outcomes unless we are careful.

Step 3: What interventions are there?

This step requires the most creativity. What possible interventions are there. List all the way you could possibly help the problem. Focus not on finding good ideas but on finding ideas you wouldn’t have thought of normally. Don’t worry that 99% of your ideas are terrible at this stage.

My first few ideas were: Working for a large chemical firm making insecticides. Working in a bank and using the money to buy nets. Becoming a research scientist finding better treatments. A public education campaign going from village to village teaching people to make their own mosquito nets. Boosting the health services of poor countries by campaigning against corruption.

Step 4: How good do you expect the non-quantifiable interventions to be?

Campaigning against corruption is a plan that would be very hard to analyse quantitatively. How do you measure corruption? On the other hand the effect of buying mosquito nets is very well understood. In general there will be some interventions that you can deal with quantitatively and some where it just doesn’t make sense.

The question now is, which do you think will be the most effective? Specifically do you think that the non-quantifiable interventions are reliably going to be better than the best that the quantifiable methods have to offer? If so you should stop here and go for something from this camp. But if not remember that you are much more likely to find the best possible intervention in a category with quantitative methods than otherwise.

Step 5: Research, decide, act.
If you are still reading at this point then you have a quantitative goal and you think there is a general category of interventions that will be best and which can be analysed quantitatively. All you have to do now is to find that best intervention and do it.


You might also be interested in: