Determining the value allocated to players is a key problem in online cooperative games. However, the theoretical definitions of fairness often fail to align with social norms. Certain value-sharing policies even produce outcomes in specific contexts that contradict with widely accepted notions of fairness. To address this conflict, we conducted human-subject experiments on this topic. Through questionnaires, participants were asked to allocate payoffs in online cooperative games with different characteristic functions, thereby exploring public perceptions of fairness in such games. Then, the experimental results were qualitatively analyzed to verify if they are consistent with certain theoretical axioms, including symmetry, incentives for early arrival, online individual rationality and monotonicity. Following the analysis, we show that additivity and strong monotonicity are not necessarily required in social norms. Based on these results, we designed a value-sharing algorithm for online cooperative games, ensuring compliance with these identified properties.