The use of coloured pan traps (bee bowls, Moericke traps) for sampling bees (and other pollinators) has continuously increased over the last two decades. Although a number of methodological studies and conceptual frameworks offer guidance on standardised sampling, pan trap setups vary widely in characteristics even when optimised for capturing bees. Moreover, some uncertainty persists as to how local flower abundance and diversity influence sampling. We systematically reviewed peer-reviewed studies that used pan traps for bee collection and that were listed in the Web of Science core collection. To gauge methodological variation, we identified a set of relevant methodological criteria and assessed the studies accordingly. For obtaining evidence that pan trap samples and floral environment around traps are correlated, we screened the relevant studies for such correlations. While some aspects of pan trapping (e.g., trap coloration and elevation) were similar in the majority of studies, other aspects varied considerably (e.g., trap volume/diameter and sampling duration). Few studies used floral abundance and/or diversity as an explanatory variable in their analyses of bee samples. Among these studies, we found a considerable variation in key aspects of floral survey methods, such as time and space between vegetation surveys and pan trap sampling, abundance measures (quantitative, semi-quantitative and presence-absence), and processing of raw data prior to analysis. Often studies did not find any correlation between the floral environment and bee samples. Reported correlations varied markedly across studies, even within groups of studies applying a similar method or analysing a similar group of bees. Our synthesis helps to identify key issues of further standardisation of pan trap methodology and of associated floral surveys. In addition to the few aspects that have been standardised over the past decades, we suggest methodological direction for future research using pan traps as a better standardised method for the collection of wild bees. We encourage further studies to illuminate if and how varying floral resources around traps bias bee samples from pan traps. More generally, our synthesis shows that trapping methodologies should be reviewed regularly when their use increases to ensure standardisation.