Measuring Mental Well-Being in Singapore

Why I did this project

I was volunteering with The Majurity Trust, a Singapore-based philanthropy organization. Their work is really fascinating - they collect donations from large donors and disburse these donations as "grants" to NGOs; almost like a venture capitalist.

I was part of the Strategy & Insights team, which aimed to build a strategy around the organization's philanthropy. At the time, Singapore was starting to see rising prevalence of Single Family Offices, SFOs, which are essentially offices set up to manage the wealth of rich families. These SFOs wanted to donate more money, but being "investment" bodies, they wanted to quantify the results of their donations. To illustrate, they wanted it to be very clear that their $100,000 investment yielded $1mn in "returns".

One field in which SFOs were willing to donate but wanted to see results was in the promotion of well-being. They wanted to understand to what degree their investment in well-being was actually improving outcomes. For this, we needed to quantify improvemnets in well-being.

Methods

How do you quantify well-being? I learnt that there are pre-existing indices already in place - for example, the Short Warwick-Edinburgh Mental Well-Being Scale (SWEMWBS) is a measure of well-being that has been validated in several contexts as a good measure of well-being. The SWEMWBS essentially uses a questionnaire to assess individuals' self-reported state of well-being.

But evidently this comes with methodological challenges. I'll point out two.

1. The SWEMWBS scale only has 7 questions - and some of them are really subjective. One of the questions, for instance, is "I’ve been dealing with problems well". What does it mean to "deal with a problem well"? For example, does it mean being analytical in dealing with problems? Does it mean ensuring you can think about the problems while still maintaining your composure and sanity? There are clearly a range of interpretations.
2. Second, does SWEMWBS actually measure what you want to measure - well-being? Do these seven questions provide a (relatively) good picture of people's state of well-being?

But these problems are not insurmountable.

Confirmatory Factor Analysis (CFA) is a statistical "stress test" used to prove that a specific set of survey questions actually measures the single concept it claims to—in this case, ensuring the seven items of the scale accurately represent the "latent" or hidden trait of mental well-being.

Once the scale's structure is confirmed, Measurement Invariance acts as a fairness check to determine if the survey "works" the same way across different groups, such as men versus women or different education levels. It moves through increasingly strict levels to ensure that differences in scores reflect real differences in well-being rather than just different groups interpreting the questions or response scales in unique ways. While the researchers found the scale was highly consistent for most groups, they noted that for race and marital status, people might not be using the "starting point" of the scale in exactly the same way, suggesting those specific group comparisons should be handled with extra care.

My work

The data collection and analysis was led by Kwan Jin Yao, a research fellow with The Majurity Trust. My job was to work on the Introduction and Literature Review sections.

Through my work, I learnt how to go about assembling a high quality literature review. The backbone of a good literature review is an annotated bibliography. And in turn, a good annotated bibliography requires a good system to find high quality sources.

I also learnt how to incorporate responsible use of AI in a literature review. My AI workflow for this project was:

1. Using AI to refine my search queries
2. Using AI to skim papers before creating a "keep/discard pile" of valuable studies.

Using AI came with teething problems. At the time, my AI model of choice - Gemini - was quite hallucinatory. Thus, I had to double-check its results before utilizing it.

You can read our paper, published in the Asia-Pacific Journal of Social Work and Development, here.