Experiments are needed to quantify the main causes of insect decline.

Sparked by reports of insect declines of unexpected extent, there has been a surge in the compilation and analysis of insect time series data. While this effort has led to valuable databases, disagreement remains as to whether, where and why insects are declining. The 'why' question is particularly important because successful insect conservation will need to address the most important drivers of decline. Despite repeated calls for more long-term data, new time series will have to run for decades to quantitatively surpass those currently available. Here we argue that experimentation in addition to quantitative analysis of existing data is needed to identify the most important drivers of insect decline. While most potential drivers of insect population change are likely to have already been identified, their relative importance is largely unknown. Researchers should thus unite and use statistical insight to set up suitable experiments to be able to rank drivers by their importance. Such a coordinated effort is needed to produce the knowledge necessary for conservation action and will also result in increased monitoring and new time series.