As shown in Table 1, all the KW tests support significant (p < 0.05) changes in people’s bag usage. Facing tougher measures, the consumers reduce their purchasing of charged carrier bags by almost 46%. Two options have been made to meet people’s carrying demand. First, they choose to bring significantly more reusable bags (χ² = 124.500; p = 0.000) and old plastic bags (χ² = 124.300; p = 0.000). The usage of reusable bags has seen an increase of 36%, while that value for old plastic bags surprisingly reaches 117%. Second, they choose to use significantly (χ² = 40.705; p = 0.000) more free inner bags, but the extent of increase is just 0.4%.

  On average, both grocery and non-grocery markets show reduced usage of charged carrier bags by over 40%. Regarding the usage of inner bags, the two types of markets vary in changing directions. As shown in Table 2, the usage of inner bags drops by over 50% within grocery markets but increases by 2.7% within non-grocery markets. As inner bags are rarely available in grocery markets, packaging style may primarily account for the different changes of inner bag usage. Thus, source control is inferred to be effective in curbing the usage of inner bags.

We use the techniques of logistic regression to learn how each influential factor modifies people’s usage of different types of bags. We classify people’s bag usage into a binary situation, i.e., to use or not to use. We then use this binary choice as the dependent variable, and the factors of gender, generation, packaging style, and tougher measures as independent variables. We first examine the potential influence of each factor by the KW test and only those significantly influential on people’s bag usage are further analyzed in logistic regression models. As shown in Table 3, the VIF value of each factor is around 1.0 in the regressions models, indicating a multi-collinearity-free status of the independent variables ^[7]. For charged carrier bags, generation is the most powerful impacting factor, then followed by tougher measures. Regarding old plastic bags, its usage is dominated by tougher plastics ban measures. As for reusable bags and inner bags, the most powerful impacting factors are gender and packaging style, respectively.

Generally, the tougher 2021 plastic ban measures work effectively in modifying people’s usage of different bags by significantly reducing people’s purchase of charged carrier bags, and by significantly increasing people’s usage of reusable bags and old plastic bags. Unavoidably, random data collection errors or other factors may account for the change. However, this valuable change is explained best by the execution of the tougher 2021 measures. The reasons are two-fold. Firstly,  the data was collected at the same markets and at the same time of a day during our two field studies. The only difference lies in the execution of tougher plastic ban policies. Secondly, the regression results also support the significant impact of the execution of tougher policies on the usage of all types of bags. With the tougher 2021 policies, on the one hand, plastic charged carrier bags are banned, which compels consumers to change; while, on the other hand, only expensive cloth or nylon charged carrier bags are available, which could again force those price-sensitive consumers to seek other cheaper options. Consequently, the consumers reduce their usage of charged carrier bags by nearly 46%, which is similar to the 49% decrease in carrier bag usage observed previously ^[8], while, as for inner plastic bags, their usage increases significantly as some consumers would take advantage of policy loopholes and turn to free inner plastic bags as alternatives for carrier bags. A similar loophole exploiting behavior is also observed in Portugal where the consumption of plastic garbage bags increases dramatically thanks to the introduction of plastic ban policies ^[9]. That is to say, as long as loopholes exist, people will exploit them to their best interests. However, what surprises us is the tiny increase in inner bag usage, i.e., 0.4%, which is quite different from observations from existing studies ^[10][11]. This is partially due to the complete ban of free inner bags within some grocery supermarkets. However, even within non-grocery supermarkets, the increased amount is also just 2.7%. That is to say, the majority decreased usage of charged carrier bags is offset by the increased usage of old bags and reusable bags rather than that of inner bags, which is positive for plastics management. For these positive results, by no means are they achieved simply through the tougher 2021 policies. As shown in the previous study ^[12], the 2020 policies could boost people’s usage of reusable bags by forbidding the usage of non-biodegradable plastic carrier bags and imposing high prices for biodegradable plastic carrier bags. The tougher 2021 policies continue to strengthen people’s behavioral change that was already shaped by 2020 policies. For example, no-bag shopping and bag-carrying behaviors start to prevail after the introduction of national regulations and the behaviors become more popular and widely accepted in 2021. Therefore, the strengthening effect could in turn explain the tiny increase in inner plastic bag usage in 2021.

Regarding this study, measures for plastics policies in China are all penalty-oriented. Incentives are not available for businesses or consumers to comply with the plastic bag ban policies. To improve the effectiveness of plastics management, future studies are encouraged regarding how various incentives could shape people’s usage of plastic bags.

Spill-over effects are found within the two no-tougher-measure supermarkets by observing similar changes in the usage of different bags. It is possible that the similar changes arise from data collection errors; however, the spill-over effects could explain better. The reasons are three-fold. To begin with, within the two rounds of data collection we keep bag counting rules the same and repeat the data collection for each site three times to keep random collection errors as few as possible. Secondly, except for old plastic bags, the amount of usage change of other bags is much smaller within no-tougher-measure markets compared to that within tougher-measure-executing supermarkets. Thirdly, people are generally consistent in their habits and behaviors ^[13]. When consumers are forced to bring reusable bags or old plastic bags for shopping in tougher-measure-executing supermarkets, bag-carrying habits could be formed and reinforced.