During my MSc in Behavioural Economics, I discovered that one of the biggest obstacles in the field isn’t designing experiments or understanding theory; it’s collecting meaningful data.
For my thesis, I struggled to recruit enough participants due to our institution's non-incentivised participation model and limited recruitment channels. Moreover, many participants did not complete the survey entirely, so I had to drop partial responses, further reducing my sample size. This left me with a smaller sample than what my study ideally required, and it exposed a challenge that many behavioural researchers quietly face.
Behavioural science relies heavily on primary data, often gathered directly from people through experiments, surveys, or tasks. But real human behaviour doesn’t always cooperate with research plans. People hesitate to participate without incentives, lose interest quickly, or skip questions. Even when they do participate, self-reported responses can be biased, inconsistent, or influenced by context, classic examples of human quirks that behavioural science aims to understand.
Unlike fields such as macroeconomics, where researchers can tap into large secondary datasets, behavioural studies need ethics approval, funding, proper recruitment mechanisms, and ways to keep participants engaged. Longitudinal research, essential for understanding how behaviour changes over time, becomes especially difficult without stable resources. This is why many behavioural studies end up cross-sectional, relying on quick snapshots rather than long-term insights. My experience made me realise that data collection is a fundamental bottleneck in behavioural research. If we want to study real behaviour more accurately and create long-term societal impact, the field needs more institutional support, accessible funding, and better infrastructure for participant recruitment.
Strengthening these foundations would not only help researchers but would also help behavioural science grow into an even more rigorous and representative discipline.