Sometimes you can have a research idea that is promising in theory, but it is hard to say if it is possible to actually execute it. Thijs Lindner, PhD candidate at the Erasmus School of Social and Behavioural Sciences, had such an idea. To find out how realistic his plans were, he approached the ODISSEI Social Data Science Team (SoDa).
Lindner currently works on a project that is already well on its way: he researches ideas about welfare among citizens. ‘While we already know a lot about political opinions on welfare among citizens, we know less about the different meanings people attribute to it,’ Lindner narrates. Through focus groups, he spoke to people from a variety of socioeconomic backgrounds. Based on these conversations, he and his colleagues asked further questions through the LISS panel, of which he is now analyzing the outcomes. The purpose and execution of that research are clear. His qualitative background was of much value, and throughout the research project he has further developed his quantitative skills.
But what do you do when you get inspired by another study, that you cannot fit immediately within your current project? ‘Inspired by this research conducted abroad, I’m developing a research plan to use comparable Dutch data to see how that particular issue has developed here. I want to use inductive cluster analyses for this. The method is innovative, new, and therefore unknown. Because of this, there are few methodological examples in the literature, where you of course look first. Because examples do not exist, it is very helpful that it is possible to approach SoDa’, Lindner explains.
His question was about the computational power needed to execute the new research. ‘If you use data from 28 waves, with 1500 respondents and 200 items per wave, then this results in potentially millions of combinations of items that correlate with each other. I was wondering if a personal computer would be up to that, or if you would need a supercomputer. In my direct environment, there is not someone who could answer such a question, which is why I contacted SoDa’.
The SoDa Team aims to be easily available to help researchers with a variety of questions around computational social sciences. SoDa Team Lead Erik-Jan van Kesteren walked through the most important challenges with Lindner. Within their one hour meeting, they looked at ways the analysis could be executed and how this could be programmed efficiently. Based on this, they concluded that a personal computer would be more than sufficient to execute this research idea. ‘I’m very happy with the help that I got, and if I would have any more questions in the future, I already know who to turn to,’ Lindner concludes.
Do you have a data-related question for SoDa? Join them at the next Data Drop-in on 16 December.