Effective Strategies for Recruiting Data Staff

Recruiting Papua New Guinea NSO June 2023

Principles when recruiting data staff It is (usually) easier to teach a technically competent recruit subject matter and context than a subject matter expert technical skills There are exceptions! Watch out for existing staff with an interest in developing technical skills. But recruitment should focus on technical competency. Recruitment process should focus on skills and aptitude to do the actual job. Interview questions should be behavioural and based in their actual skills tell us about a time you did You must include a practical test of some sort.

Some example interview questions Tell us about one of the more difficult data analysis tasks you had to perform. What was the problem, what sort of data you had, and how did you go about it? Tell us about a complex data management problem where you had to combine two datasets What s your experience with statistical modelling and inference? Tell us about a time you ve had to draw conclusions about a broad population using data just from a sample. Have you ever worked with a database? What s a time you ve had to extract data from a database to use with a real life problem. In this job you ll have to make a lot of use of Excel. What s something you ve done before with Excel that was a challenge. What did you learn? How did you go about solving the problem?

Example practical exercise The accompanying file "ruritania-population-data.csv" contains information on the population by age and sex and the deaths in the past 12 months of the fictional small country of Ruritania. Some of the population and death counts are missing. In your response, you should provide: a) a very brief report (1.5 pages maximum) with the results requested above and brief explanation, aimed at a not-very- technical audience, of how you derived the results and dealt with missing data Using this data, your job is to: b) a final dataset with the same grain as the original data (i.e. one row for each combination of age and sex) but with any additional variables you have created to help produce the outputs (for example, calculated death rates, or imputed values of population and death counts) 1) calculate the total population of Ruritania 2) For each of males and females in Ruritania, calculate the a) crude death rate; c) your workings in sufficient detail to allow peer review. For example, this might be your computer code if done with a language like R, or an Excel workbook with enough commentary and descriptions that it can be seen exactly what you did. b) under 5 mortality rate; and c) life expectancy at birth d) average age at death for someone aged 30 3) draw a population pyramid with ages in 5 year brackets This exercise is aimed at testing your ability to do some fundamental demographic tasks, rather than consume unreasonable amounts of your time in advanced or cutting edge techniques. For example, there is no need to greatly polish the outputs, provide confidence intervals, or estimate growth rates or other numbers not asked for in the tasking above. You can use whichever tools you prefer and have available.

Data and model answer