1. 智能绿色助推：通过数字足迹与因果机器学习减少商品退货

原标题：Smart Green Nudging: Reducing Product Returns Through Digital Footprints and Causal Machine Learning

期刊：Marketing Science

作者：Moritz von Zahn, Kevin Bauer, Cristina Mihale-Wilson, Johanna Jagow, Maximilian Speicher, Oliver Hinz

出版时间：2024

DOI：10.1287/mksc.2022.0393

摘要：在电子商务领域，商品退货已成为零售商面临的一个昂贵且日益严重的问题。除了对企业带来的经济影响外，商品退货还导致温室气体排放增加和自然资源的浪费。传统方法（例如对退货收取费用）在抑制退货方面效果有限，这凸显了需要采用更精细化的策略来解决这一问题。本研究通过与一家欧洲领先时尚零售商在线商店合作开展的大规模随机实地实验（样本量为117,304），探讨向消费者告知商品退货对环境带来的负面影响（即“绿色助推”）在减少商品退货方面的有效性。研究结果表明，实施绿色助推可将商品退货率降低2.6%，且不会对销售额产生负面影响。随后，本研究开发并评估了一种因果机器学习模型，用于识别干预效果的异质性并实现绿色助推的个性化（即使助推更加“智能化”）。非政策性评估的结果表明，这种个性化方法可以将绿色助推的效果提高近一倍。本研究表明，借助微妙的营销干预和因果机器学习实现的个性化，可以有效减少对环境和经济有害的商品退货。这突出了在数字化环境中采用“更好的营销促进更美好的世界”方法的可行性。

Abstract：In e-commerce, product returns have become a costly and escalating issue for retailers. Beyond the financial implications for businesses, product returns also lead to increased greenhouse gas emissions and the squandering of natural resources. Traditional approaches, such as charging customers for returns, have proven largely ineffective in curbing returns, thus calling for more nuanced strategies to tackle this issue. This paper investigates the effectiveness of informing consumers about the negative environmental consequences of product returns ("green nudging") to curtail product returns through a large-scale randomized field experiment (n = 117,304) conducted with a leading European fashion retailer's online store. Our findings indicate that implementing green nudging can decrease product returns by 2.6% without negatively impacting sales. We then develop and assess a causal machine learning model designed to identify treatment heterogeneities and personalize green nudging (i.e., make nudging "smart"). Our off-policy evaluation indicates that this personalization can approximately double the success of green nudging. The study demonstrates the effectiveness of both subtle marketing interventions and personalization using causal machine learning in mitigating environmentally and economically harmful product returns, thus highlighting the feasibility of employing "Better Marketing for a Better World" approaches in a digital setting.

2. 食品价格变动对消费者购买行为的影响：一项随机实验研究

原标题：The effect of food price changes on consumer purchases: a randomised experiment

期刊：Lancet Public Health

作者：Wilma E. Waterlander, Yunnan Jiang, Nhung Nghiem，Helen Eyles, Nick Wilson, Christine Cieghorn, Murat Genc, Boyd Swinburn, Cliona Ni Mhurchu,  Tony Blakely

出版时间：2019

DOI：10.1016/S2468-2667(19)30105-7

摘要：关于健康相关食品税收和补贴的大多数证据主要依赖于观察性数据，并聚焦于单一营养素或食品的影响，而非整体饮食。本研究旨在通过随机分配的食品价格变化，测量消费者购买行为的影响，模拟了一些常见的食品税收和补贴政策，包括对水果和蔬菜的补贴、含糖饮料税，以及基于食品中糖分、钠含量和饱和脂肪含量的税收。在本研究中，新西兰的成年参与者（年龄≥18岁）在虚拟超市中完成最多五次每周购物。每次购物随机分配至控制组（价格无变化）或一种或多种价格选项，模拟以下政策：水果和蔬菜补贴（20%）、含糖饮料税（20%或40%）、饱和脂肪税（每100克饱和脂肪2新西兰元或4新西兰元）、盐税（每100毫克钠0.02新西兰元或0.04新西兰元）以及糖税（每100克糖0.40新西兰元或0.80新西兰元）。研究结果显示，在五项政策中，有三项显著增加了健康食品购买的比例，但增幅较小；含糖饮料税和水果蔬菜补贴分别导致非显著的增长，增幅为0.18%和0.41%。此外，饱和脂肪税和盐税产生了以下重要的替代效应：一方面，水果和蔬菜购买量在食品总重量中所占比例有所增加；另一方面，总能量中糖的比例也有所上升。政策组合的交互作用项大多不显著，表明政策组合的效果主要是加性效果。

Abstract：Most evidence on health-related food taxes and subsidies relies on observational data and effects on single nutrients or foods instead of total diet. The aim of this study was to measure the effect of randomly assigned food price variations on consumer purchasing, where sets of prices emulated commonly discussed food tax and subsidy policies, including a subsidy on fruit and vegetables, a sweetened beverage tax, and taxes on foods according to sugar, sodium, and saturated fat content. In this study, adult participants (>= 18 years) in New Zealand completed up to five weekly shops in a virtual supermarket. Each shopping occasion was randomly allocated to control (no change in prices) or one or more pricing options simulating the following: a fruit and vegetable subsidy (20%), a sweetened beverage tax (20% or 40%), a saturated fat tax (NZ$2 per 100 g or $4 per 100 g saturated fat), a salt tax ($0.02 per 100 mg or $0.04 per 100 mg sodium), or sugar tax ($0.40 per 100 g or $0.80 per 100 g sugar). The results show that three of the five policies increased this proportion by a small, but significant amount. The sweetened beverage tax and fruit and vegetable subsidy resulted in non-significant increases of 0.18% and 0.41%, respectively. Both the saturated fat tax and salt tax resulted in the following important substitution effects: an increase in fruit and vegetable purchases as a percentage by weight of all food purchases；but also an increase in sugar as a percentage of total energy. Interaction terms for combined policies were mostly non-significant, consistent with additive effects of policy combinations.

3. 事先反思提升助推可持续饮食选择的效果

原标题：Sustainable dietary choices improved by reflection before a nudge in an online experiment.

期刊：Nature Sustainability

作者：Sanchayan Banerjee, Matteo M. Galizzi, Peter John, Susana Mourato

出版时间：2023

DOI：10.1038/s41893-023-01235-0

摘要：当前的食品选择碳足迹较高，与气候目标不相符，因此向更环保的饮食方式转变尤为重要。行为科学中的“助推”（nudges）已被广泛用于减少对肉类食品的需求，通过微妙地改变选择的呈现方式，而非通过禁令或提高成本来实现。然而，将助推措施大规模推广存在挑战，并有时引发伦理争议。为应对这一问题，行为科学提出通过促使个体反思其选择来赋权，从而推动有意义且更环保的行为改变。在一项包含3074名英国参与者的实验研究中，作者比较了三种增强个体主动性的工具（“提升”（boost）、“思考”（think）和“助推+”（nudge+））与传统助推策略（如默认选项和标签标识）在促进可持续饮食意图方面的效果。结果显示，所有行为干预措施均提高了参与者对可持续食品的选择意图，但在默认选择绿色饮食前鼓励反思饮食偏好效果最佳。特别是，“助推+”策略（在默认选项前增加承诺环节）显著减少了参与者预期订购餐食所产生的碳排放量达40%。研究表明，食品企业可以通过在助推消费者选择更可持续食品之前，引导其进行思考，从而增强其可持续发展努力的成效。

Abstract：Current food choices have a high carbon footprint and are incompatible with climate goals. Transitioning to more environmentally friendly diets is therefore important. Behavioural 'nudges' have been widely used to reduce meat-based food demand, subtly altering choice presentation without banning or raising costs. However, scaling up nudges has proven challenging, sometimes raising ethical concerns. To address this, behavioural science proposes empowering individuals to reflect on their choices, fostering meaningful and more environmentally-friendly behavioural changes. In an experimental study with 3,074 UK participants, we compared three agency-enhancing tools ('boost', 'think' and 'nudge+') with classic nudges (opt-out default and labelling) to promote sustainable dietary intentions. All behavioural interventions increased intentions for sustainable foods but encouraging reflection on dietary preferences before defaulting people into greener diets yielded the best results. Adding a pledge before the default nudge, as in nudge+ (pledge+ default), additionally reduced emissions from intended orders of meals by 40%. Our research suggests that food companies can enhance their sustainability efforts by prompting customers to think before nudging them into consuming more sustainable food.

4. 选择架构在在线食品配送应用中促进可持续选择

原标题：Choice architecture promotes sustainable choices in online food-delivery apps

期刊：Pnas Nexus

作者： Paul M. Lohmann, Elisabeth Gsottbauer, James Farrington, Steve Human, Lucia A. Reisch

出版时间：2024

摘要：食品系统的温室气体排放约占全球总排放量的三分之一，因此在这一领域采取减排措施是研究和政策的重要目标。本研究通过使用包含4008名参与者的激励兼容在线随机对照试验，实证检验了不同干预措施在减少食品配送应用中食品选择碳足迹的有效性。实验利用一个互动网页平台，模拟了流行的在线食品配送平台（如Just Eat），并包括三种干预条件：标示肉类税、碳足迹标签以及选择架构干预，通过调整菜单顺序，将低碳餐厅和菜肴排在最前面。研究结果表明，只有选择架构的助推显著减少了平均每餐碳足迹——每单减少0.3千克二氧化碳当量（减少12%），这主要是由于高碳餐食选择减少了5.6个百分点（减少13%）。此外，研究还发现了显著的健康和福祉协同效益。菜单调整使得平均每餐订单的营养价值更高，卡路里含量更低，同时显著提高了参与者对餐食选择的自我报告满意度。简单的初步计算表明，如果在大规模实施，菜单重新排序将是一个非常具有成本效益的政策工具，预计每减少1吨二氧化碳排放的投资回报率将在1.28英镑至3.85英镑之间，具体取决于实施成本。

Abstract：Greenhouse gas emissions from the food system constitute about one-third of the global total, hence mitigation in this sphere of human activity is a vital goal for research and policy. This study empirically tests the effectiveness of different interventions to reduce the carbon footprint of food choices made on food-delivery apps, using an incentive-compatible online randomized controlled trial with 4,008 participants. The experiment utilized an interactive web platform that mimics popular online food-delivery platforms (such as Just Eat) and included three treatment conditions: a sign-posted meat tax, a carbon-footprint label, and a choice-architecture intervention that changed the order of the menu so that the lowest carbon-impact restaurants and dishes were presented first. Results show that only the choice-architecture nudge significantly reduced the average meal carbon footprint-by 0.3 kg/CO2e per order (12%), driven by a 5.6 percentage point (13%) reduction in high-carbon meal choices. Moreover, we find evidence of significant health and well-being co-benefits. Menu repositioning resulted in the average meal order having greater nutritional value and fewer calories, whilst significantly increasing self-reported satisfaction with the meal choice. Simple back-of-the-envelope calculations suggest that menu repositioning would be a highly cost-effective policy instrument if implemented at scale, with the return on investment expected to be in the range of 1.28 pound to 3.85 pound per metric ton of avoided CO2 emissions, depending on implementation costs.