Subscribe to the monthly socialservice.sg newsletter and check out the socialservice.sg podcast!

Poll proportions alone (or percentages) – for instance, that 81 per cent of Singaporeans are afraid of infection and that 35 per cent would still attend events even with mild symptoms (ST, Feb. 17) – may provide some useful descriptive information, yet news agencies could explore more effective ways to present the information, include cross tabulations to explore the relationship between different variables, and consequently use the data to better inform policy decisions.

In this particular ST survey of 401 Singaporean households on the coronavirus – ostensibly representative of the country’s demographics – the key findings include:

• Women were more afraid than men of getting infected;
• Women were more likely to have stockpiled medical supplies and groceries as well as to have worn masks outdoor despite being suck, even though men were more likely to feel more prepared; and
• Younger Singaporeans were more likely to attend an important event despite being sick and less likely to express fear over the outbreak.

Besides the fact that it is difficult to ascertain which of the findings are more important, visual representations – such as the use of tables or graphs – probably provide clearer illustrations. For example (assuming that respondents chose from dichotomous yes / no options), the tables could look like these:

Are you afraid of getting infected?	All	Men	Women
Yes	81.0%	75.5%	85.9%
No	19.0%	24.5%	14.1%

Would you attend an important event if you had symptoms of being sick?	All	15 to 29 years old	30 to 49 years old	50 years old and older
Yes	~35%	62.5%	35.2%	27.1%
No	~65%	37.5%	64.8%	72.9%

Would you advise others on public transport to wear a mask or seek medical help if the looked sickly?	All	Married with children under 13	Married with no pre-teen children	Singles
Yes	31.2%	41.6%	22.6%	32.9%
No	68.8%	58.4%	77.4%	67.1%

Even with these tables, the observations are still singly organised based on the respondent’s sex, age, and marital status. However, there is value in jointly considering these variables – sex and age or age and marital status – and in understanding the extent to which respondents are guided by their perceived fear of the coronavirus and by their knowledge of it.

How much do respondents know in the first place? Are younger Singaporeans more willing to head out because they think older individuals are more likely to be infected? And what accounts for the difference between those married with no pre-teen children and the singles?

	Married with children under 13	Married with no pre-teen children	Singles
15 to 29 years old	…	…	…
30 to 49 years old	…	…	…
50 years old and older	…	…	…

A cross tabulation like this – which relates the two variables of age and marital status together – would provide some insights into the differences, as aforementioned, between those married with no pre-teen children and the singles. Policy decisions are improved when we know who to target and even how to do so, and good data (presentation) can help with that.