Strategies for guiding urban residents’ water-saving behavior based on field experiments
Received date: 2022-04-06
Revised date: 2022-07-03
Online published: 2022-10-25
In recent years, information intervention based on field experiment has been widely used in the field of environmental protection. In order to study the impact of information interventions on urban household water consumption, an eight-month field experiment was carried out in Huangdao District, Qingdao City, Shandong Province. Combined with the empirical model, the water-saving effects of four information intervention methods were analyzed and compared: water-saving skills; water-saving skills + water-saving education; water-saving skills + water use comparison within the community; and water-saving skills + water use comparison in the same city. The results show that: (1) The four types of information interventions all have a significant impact on the reduction of urban residential water consumption; (2) The water-saving effect of intra- and inter-community comparison information provision is affected by the geographical distance of those to compare with—a closer distance to the counterparts is more conducive to saving water than comparing with distant counterparts; (3) The water-saving effect of information intervention is more effective for people with high water consumption in the short term, but more effective for people with low water consumption in the long term. (4) The water-saving effect of information campaigns attenuate over time. Through the experiments, this study confirmed the validity of field experiment methods for household water saving in the context of China. The conclusions of this research provide a reference for the application of information intervention methods in the field of household water saving campaigns.
LI Yongbo , WANG Juan . Strategies for guiding urban residents’ water-saving behavior based on field experiments[J]. Resources Science, 2022 , 44(8) : 1663 -1678 . DOI: 10.18402/resci.2022.08.10
表1 实验干预分组和信息内容设计Table 1 Experiment groups and information intervention designs |
实验干预 | 收到短信用户数 | 信息内容示例 |
---|---|---|
节水技巧(A组) | 370 | ①家中预备水盆收集废水,废水冲厕,节约清水;② 淘米水、煮面条的水用来洗碗筷,去油又节水;③淘米水浇花,能促进花木生长 |
节水技巧+节水教育(B组) | 368 | 随着城市化发展,生活用水逐年增加。居民参与节水,能够有效缓解水资源短缺问题。节水小贴士:①家中预备水盆收集废水,废水冲厕,节约清水;②淘米水、煮面条的水用来洗碗筷,去油又节水;③淘米水浇花,能促进花木生长 |
节水技巧+同社区用水比较(C组) | 368 | 本社区6、7两个月合计户均用水量约为XX方,您的用水号6053XXX用水量为XX方。节水小贴士:①家中预备水盆收集废水,废水冲厕,节约清水;②淘米水、煮面条的水用来洗碗筷,去油又节水;③淘米水浇花,能促进花木生长 |
节水技巧+同城用水比较(D组) | 369 | 城区6、7两个月合计户均用水量约为XX方,您的用户号6055XXX用水量为XX方。节水小贴士:①家中预备水盆收集废水,废水冲厕,节约清水;②淘米水、煮面条的水用来洗碗筷,去油又节水;③淘米水浇花,能促进花木生长 |
表4 不同信息干预策略对用水量影响的基准回归Table 4 Benchmark regression results of water-saving effects of different information intervention strategies |
(1) | (2) | (3) | (4) | (5) | (6) | |
---|---|---|---|---|---|---|
Treat-Pooled | -0.569*** (-5.94) | |||||
TreatA | -0.558*** (-4.180) | -0.547*** (-3.960) | ||||
TreatB | -0.656*** (-5.130) | -0.653*** (-5.040) | ||||
TreatC | -0.758*** (-5.570) | -0.732*** (-5.560) | ||||
TreatD | -0.319** (-2.010) | -0.272* (-1.820) | ||||
Ycontrol | 0.449*** (4.700) | 0.448*** (4.660) | 0.528*** (5.560) | 0.528*** (6.280) | 0.571*** (6.120) | 0.291*** (2.950) |
Yspring | 0.425*** (4.420) | 0.426*** (4.44) | 0.411*** (3.780) | 0.409*** (4.180) | 0.347*** (3.260) | 0.641*** (6.950) |
常数项 | 1.353*** (5.793) | 1.352*** (5.890) | 0.910*** (4.76) | 0.919*** (5.200) | 1.024*** (5.350) | 1.049*** (5.160) |
R2 | 0.652 | 0.652 | 0.691 | 0.707 | 0.682 | 0.644 |
观察值数 | 1 831 | 1 831 | 974 | 985 | 968 | 992 |
注:括号内的数值表示相应估计系数的t统计值,***、**、*分别表示在1%、5%、10%的水平显著,下同。 |
表2 指标及样本的描述性统计Table 2 Descriptive statistics of the indicators and samples |
控制组 | A组 | B组 | C组 | D组 | 单位 | |
---|---|---|---|---|---|---|
筛选前 | 763 | 370 | 368 | 368 | 369 | 户 |
筛选后 | 696 | 278 | 289 | 272 | 296 | 户 |
Y | 7.124 | 6.290 | 6.359 | 6.159 | 6.562 | m3 |
(3.346) | (3.133) | (3.398) | (3.166) | (3.445) | ||
实验前后变化率 | 5.043 | -4.116 | -5.189 | -4.946 | -0.304 | % |
Ycontrol | 6.782 | 6.560 | 6.707 | 6.470 | 6.580 | m3 |
(3.034) | (2.963) | (3.324) | (2.976) | (3.773) | ||
Yspring | 6.404 | 6.001 | 6.230 | 6.247 | 6.042 | m3 |
(2.988) | (3.429) | (3.755) | (4.202) | (2.984) | ||
P值 | 0.197 | 0.309 | 0.730 | 0.152 | 0.388 | - |
注:P值为2020年6—11月实验组与对照组双边t检验的显著性;括号内为相关变量的标准差。 |
表3 不同信息干预手段对用水量的影响模型变量的多重共线性检验Table 3 Multicollinearity test of the variables |
Ycontrol | Yspring | TreatA | TreatB | TreatC | TreatD | |
---|---|---|---|---|---|---|
VIF | 2.100 | 2.100 | 1.190 | 1.190 | 1.190 | 1.200 |
1/VIF | 0.476 | 0.476 | 0.841 | 0.839 | 0.843 | 0.836 |
表5 不同信息干预策略节水效果差异检验Table 5 Tests of differences in water-saving effects of different information intervention strategies |
(1) | (2) | (3) | (4) | |
---|---|---|---|---|
Treat-Pooled | -0.572*** (-5.56) | -0.539*** (-5.17) | -0.509*** (-5.00) | -0.657*** (-6.72) |
Treat-Pooled×TreatA | 0.014 (0.100) | |||
Treat-Pooled×TreatB | -0.117 (-0.920) | |||
Treat-Pooled×TreatC | -0.249* (-1.900) | |||
Treat-Pooled×TreatD | 0.337** (2.220) | |||
Ycontrol | 0.449*** (4.700) | 0.45*** (4.690) | 0.448*** (4.690) | 0.449*** (4.670) |
Yspring | 0.425*** (4.410) | 0.425*** (4.410) | 0.426*** (4.440) | 0.425*** (4.430) |
常数项 | 1.353*** (5.793) | 1.351*** (5.170) | 1.354*** (5.940) | 1.352*** (5.890) |
R2 | 0.652 | 0.653 | 0.653 | 0.654 |
观察值数 | 1 831 | 1 831 | 1 831 | 1 831 |
表6 不同信息干预策略对用水量影响的分位数回归结果Table 6 Quantile regression results of water-saving effects of different information intervention strategies |
10%分位点 | 50%分位点 | 90%分位点 | |
---|---|---|---|
TreatA | -0.490* (-1.790) | -0.415*** (-4.970) | -0.732*** (-2.970) |
TreatB | -0.566** (-2.390) | -0.402*** (-3.420) | -0.931*** (-4.050) |
TreatC | -0.842*** (-4.990) | -0.499*** (-5.300) | -1.148*** (-4.400) |
TreatD | -0.208 (-1.460) | -0.328*** (-3.120) | -0.501* (-1.830) |
Ycontrol | 0.319*** (3.730) | 0.399*** (7.170) | 0.369*** (5.600) |
Yspring | 0.389*** (3.680) | 0.586*** (10.260) | 0.707*** (11.970) |
常数项 | 0.556*** (3.920) | 0.602*** (8.280) | 1.926*** (7.660) |
R2 | 0.343 | 0.539 | 0.495 |
观察值数 | 1 831 | 1 831 | 1 831 |
表7 不同信息干预策略节水效果的持久性Table 7 Persistence of water-saving effects of different information intervention strategies |
第一个实验周期 | 第四个实验周期 | ||||||
---|---|---|---|---|---|---|---|
总体回归 | 高于中位数 | 低于中位数 | 总体回归 | 高于中位数 | 低于中位数 | ||
TreatA | -1.262*** (-7.200) | -1.604*** (-5.670) | -0.954*** (-4.700) | -0.484*** (-2.850) | -0.081 (-0.270) | -0.844*** (-4.800) | |
TreatB | -1.659*** (-9.320) | -2.071*** (-7.390) | -1.217*** (-5.640) | -0.400** (-2.130) | 0.111 (0.340) | -0.899*** (-4.640) | |
TreatC | -1.313*** (-7.080) | -1.700*** (-5.500) | -0.953*** (-4.490) | -0.427** (-2.100) | -0.109 (-0.330) | -0.762*** (-3.150) | |
TreatD | -1.051*** (-5.260) | -1.862*** (-6.760) | -0.263 (-0.910) | -0.017 (-0.060) | 0.025 (0.080) | -0.114 (-0.280) | |
Y202007 | 0.625*** (22.340) | 0.539*** (9.120) | 0.515*** (8.380) | ||||
Y202101 | 0.643*** (8.210) | 0.450*** (3.320) | 0.775*** (12.930) | ||||
常数项 | 2.989*** (15.170) | 4.287*** (8.190) | 2.971*** (9.710) | 2.296*** (4.830) | 4.169*** (3.520) | 1.659*** (5.690) | |
R2 | 0.399 | 0.234 | 0.129 | 0.344 | 0.167 | 0.128 | |
观察值数 | 1 831 | 912 | 919 | 1831 | 860 | 971 |
表8 不同信息干预策略对用水量影响:缩尾处理和WLS回归结果Table 8 Impact of different information intervention strategies on water use: Results of tailing treatment and weighted least squares (WLS) regression |
(1) | (2) | |
---|---|---|
TreatA | -0.462*** (-3.750) | -0.499*** (-3.350) |
TreatB | -0.616*** (-5.420) | -0.603*** (-4.300) |
TreatC | -0.625*** (-5.130) | -0.742*** (-5.310) |
TreatD | -0.347*** (-2.770) | -0.445*** (-3.020) |
Ycontrol | 0.386*** (4.760) | 0.120*** (9.800) |
Yspring | 0.373*** (4.620) | 0.743*** (43.940) |
常数项 | 1.948*** (9.760) | 1.624*** (11.700) |
R2 | 0.657 | 0.659 |
观察值数 | 1 831 | 1 831 |
表9 平行趋势检验Table 9 Parallel trend test |
混合处理组 | A组 | B组 | C组 | D组 | |
---|---|---|---|---|---|
PRE_4 | 0.255 (1.410) | 0.216 (0.950) | 0.041 (0.210) | 0.254 (1.240) | 0.502 (0.970) |
PRE_3 | -0.065 (-0.510) | -0.051 (-0.240) | -0.148 (-0.790) | -0.062 (-0.350) | 0.001 (0.010) |
PRE_2 | -0.044 (-0.340) | -0.200 (-1.230) | -0.141 (-0.980) | 0.391 (1.170) | -0.201 (-1.540) |
CURRENT | -0.974*** (-9.110) | -0.900*** (-4.720) | -1.237*** (-7.410) | -0.977*** (-6.620) | -0.785*** (-5.570) |
TIME_1 | -0.937*** (-6.350) | -0.737*** (-3.060) | -0.875*** (-3.840) | -1.288*** (-6.460) | -0.862*** (-4.380) |
TIME_2 | -0.807*** (-6.260) | -0.963*** (-4.520) | -0.981*** (-5.010) | -0.751*** (-3.980) | -0.542*** (-3.080) |
TIME_3 | -0.367*** (-2.620) | -0.497** (-2.480) | -0.548** (-2.500) | -0.431** (-2.150) | -0.012 (-0.050) |
Weather | 0.006 (0.590) | 0.006 (0.590) | 0.006 (0.590) | 0.006 (0.590) | 0.006 (0.590) |
常数项 | 6.486*** (76.220) | 6.473*** (64.760) | 6.653*** ( 66.230) | 6.459*** (67.340) | 6.485*** (66.980) |
时间固定效应 | Yes | Yes | Yes | Yes | Yes |
个体固定效应 | No | No | No | No | No |
R2 | 0.032 | 0.064 | 0.065 | 0.048 | 0.035 |
观察值数 | 14 648 | 7 792 | 7 880 | 7 744 | 7 936 |
表10 DID回归结果Table 10 Difference-in-differences (DID) regression results |
混合处理组 | A组 | B组 | C组 | D组 | |
---|---|---|---|---|---|
Treat-Pooled×Time | -0.808*** (-8.040) | ||||
TreatA×Time | -0.766*** (-5.630 ) | ||||
TreatB×Time | -0.849*** (-6.640) | ||||
TreatC×Time | -1.001*** (-6.600) | ||||
TreatD×Time | -0.626*** (-3.250) | ||||
Weather | 0.028*** (5.360) | 0.040*** (7.190) | 0.041*** (6.860) | 0.044*** (6.930) | 0.038*** (6.930) |
常数项 | 6.379*** (62.480) | 6.112*** (68.700) | 6.114*** (69.400) | 6.059*** (60.230) | 6.241*** (37.610) |
时间固定效应 | Yes | Yes | Yes | Yes | Yes |
个体固定效应 | No | No | No | No | No |
Cluster | Yes | Yes | Yes | Yes | Yes |
R2 | 0.03 | 0.063 | 0.063 | 0.045 | 0.032 |
观察值数 | 14 648 | 7 792 | 7 880 | 7 744 | 7 936 |
[1] |
梁缘毅, 吕爱锋. 中国水资源安全风险评价[J]. 资源科学, 2019, 41(4): 775-789.
[
|
[2] |
|
[3] |
|
[4] |
申津羽, 王煜琪, 赵正. 公众参与城市水环境治理行为的影响因素: 基于北上广三地的调查数据[J]. 资源科学, 2021, 43(11): 2289-2302.
[
|
[5] |
朱永彬, 史雅娟. 中国主要城市水资源价值评价与定价研究[J]. 资源科学, 2018, 40(5): 1040-1050.
[
|
[6] |
吴正, 田贵良, 胡雨灿. 基于开放式水资源嵌入型CGE模型的税改政策经济影响与节水效应[J]. 资源科学, 2021, 43(11): 2264- 2276.
[
|
[7] |
张巍, 韩军, 周绍杰. 中国城镇居民用水需求研究[J]. 中国人口·资源与环境, 2019, 29(3): 99-109.
[
|
[8] |
|
[9] |
廖显春, 夏恩龙, 王自锋. 阶梯水价对城市居民用水量及低收入家庭福利的影响[J]. 资源科学, 2016, 38(10): 1935-1947.
[
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
陈凯. 信息干预对绿色消费行为的影响分析[J]. 中国特色社会主义研究, 2020, (Z1): 126-131.
[
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
[29] |
|
[30] |
|
[31] |
刘霁瑶, 贾亚娟, 池书瑶, 等. 污染认知、村庄情感对农户生活垃圾分类意愿的影响研究[J]. 干旱区资源与环境, 2021, 35(10): 48-52.
[
|
[32] |
虞佳玲, 王瑞, 袁勤俭. 社会认同理论及其在信息系统研究中的应用与展望[J]. 现代情报, 2020, 40(10): 159-167.
[
|
[33] |
|
[34] |
王建明, 王秋欢, 吴龙昌. 城市居民节水行为的启动、形成和持续机制: 双重视角下的探索性研究[J]. 消费经济, 2016, 32(4): 10-16.
[
|
[35] |
|
[36] |
|
[37] |
|
[38] |
刘满芝, 石明珠, 史梦茜. 近邻不如远邻? 比较性信息反馈对节能的影响:空间距离视角[J]. 中国矿业大学学报(社会科学版), 2020, 22(5): 131-148.
[
|
[39] |
|
[40] |
|
[41] |
芈凌云, 丁超琼, 俞学燕, 等. 不同信息框架对城市家庭节电行为干预效果的纵向实验研究[J]. 管理评论, 2020, 32(5): 292-304.
[
|
[42] |
|
[43] |
|
[44] |
|
/
〈 |
|
〉 |