These software and how we use them have developed dramatically since the 1980s. In the same period, nobody in Statistics Norway has given more trainings in statistical software at home and abroad than Kristian Lønø.
Results?
While totally necessary for most statistics production, the ability to program is rarely the end goal of development cooperation projects in statistics. So how does he see the results of his work?
“It is not so easy to see”, says Kristian. “We come in, work for a while and then leave. We do not track how the skills are used in the long run. Turnover in the NSOs has also always been a challenge. Many people leave the statistics offices after having built experience. But some stay - almost as if they have a calling.
One of my favourite examples is Frank Kakungu in Zambia. In 1994 he attended one of my trainings. Soon after he came to Statistics Norway to work further on his project with us. 23 years later I happened to meet him in a hallway at ZamStats. By then, in 2017, he had become manager of the IT-department.”
An evolving landscape
When Kristian Lønø boarded a plane to Zambia together with a colleague in 1992, he had little idea that this trip would mark the beginning of a three-decade-long journey of also shaping statistical capacity across continents. In their bags they had extra chalk for the blackboard, extra lightbulbs for the overhead projector and an open mind towards what may and may not be meeting them in the statistics office.
Since then, Kristian has gone abroad 60 times to conduct training or work with colleagues. He has been to 18 different countries in Africa, Asia and Eastern Europe. He has taught SAS, SPSS, Stata, R, CSPro and Python, assisted in translating between languages and contributed to IT mapping and strategies. The arrival of internet has made the trips and the workshops much easier to organise, but flexibility is still an important element for success. “The need does not always match what is stated in the terms of reference,” he says. “You just have to dive in and make the best out of it.”
Hierarchical cultures and fear of mistakes can stifle initiative. “If you’re afraid to fail, the safest option is to do nothing,” he explains. “But things do change. Encouraging participants to present solutions in class-something unthinkable years ago-has helped build confidence and foster discussion. And another clear improvement over time is digital literacy. People are now more comfortable online and quicker to pick up new tools.”
Time is however still a hurdle. “Capacity building takes more than a one-week course,” Kristian emphasizes. “We want our colleagues to be able to do their own programming without support, but we have often ended up fixing problems for them because there is not enough time. It is always a challenge to get enough resources to get to a sustainable level of programming skills.”
.jpg)
The Shift to Modern Tools
Blackboards are history, and even PowerPoint has given way to interactive coding environments. Kristian now teaches R and Python for statistical production, as the NSOs are moving away from proprietary software like SPSS and Stata. This shift isn’t just technical-it’s strategic. Open-source tools mean countries aren’t locked into expensive licenses.
Another change also reflects a broader trend: statistical offices increasingly expect statisticians to program themselves. In the past, there was a clear divide between those who produced statistics and those who did the programming. Now, everyone involved in producing statistics must be able to use the software, and only the most advanced parts are developed by IT experts. “It’s the same evolution we have had in Norway,” Kristian notes. “The countries we work with are on the same path as us-we are just a bit ahead. That’s how we have something to contribute.”
Lessons learned
Kristian has taken a couple of pedagogy courses through Statistics Norway. Adding experience, he has a long list of lessons learned.
“We have started doing systematic IT assessments at the beginning of projects, so that we know in advance what level to start from. This counts for both infrastructure and skills. Another lesson learned is that connecting the courses to the concrete tasks participants are working on is more sustainable than holding general courses. When people have a concrete task, they continue using what they have learned. We still often need to start with the basics, though now the learners can also find much of the general content online.”
People also need follow-up and to keep practicing after a training, just as in Statistics Norway. Online meetings are now making it possible to circle back after the trainings with short sessions where participants share progress and ask questions.
“It’s easier now, but it still requires commitment,” says Kristian.
Recycling work
You can’t just flip through the same stack of PowerPoints when you’re teaching six different programs to be used for a variety of statistical domains in 18 different countries. Some work can however be reused.
A main goal in official statistics is to produce it in the same way across countries to get comparability. In 2019, Kristian Lønø and head of price statistics at SSB, Espen Kristiansen, travelled to Mali with Statistics Sweden to assist the national statistical office develop a system for calculating price indices. The project was ambitious: create a robust method for foreign trade price statistics that could be maintained locally. Espen designed a framework, and together they simplified it as much as possible to match the programming skills available. “No matter how simple you make it, there’s always complexity in statistical production,” Kristian reflects.
Afterwards he saw an opportunity. He rewrote the system in Python to make the tool reusable and this decision paid off. The Mali solution became the foundation for a new, open-source version now used in Namibia, Malawi is working on it, and Mozambique is considering it. “It’s the same concept, adapted to local conditions,” Kristian explains. “Moving from proprietary software to open-source makes it sustainable. The NSOs can keep developing the tools themselves.”
SSB has published a document by Kristian comparing five of the most used software to help people to decide which one to use and how to move from one software to another. Ironically, ChatGPT was launched at the same time. However, the days of this knowledge are not over.
“Some people have this idea that AI will soon write the syntax for us, but it’s not quite like that. AI is a very useful tool for finding code and suggest solutions. It can save us time, but you need to understand the code and critically assess what’s happening. Also, we should use it in a way that makes us smarter, not dumber. Otherwise, you can end up with seriously wrong statistics.”
