Table 1 Effect sizes per key moderators

From: Meta-analysis of randomised controlled trials testing behavioural interventions to promote household action on climate change

Moderator

k

N

Effect size d (CI)

I2 (%)

POB (%)

Overall effect size

144

3,092,678

−0.093 (−0.160, −0.055)

64.6**

6.6

Sensitivity analysis

Sample type

Households

66

724,792

−0.112 (−0.221, −0.057)

73.1**

 

Individuals

78

2,367,886

−0.118 (−0.221, −0.060)

51.9**

 

Sample size per condition

≤100

82

5709

−0.335 (−−0.555, −0.190)

49.9**

 

]100, 500[

45

22,840

−0.141 (−0.280, −0.063)

51.4**

 

≥500

17

3,074,121

−0.028 (−0.106, −0.006)

25.6

 

Self-selection

Self-selected

79

12,550

−0.279 (−0.465, −0.161)

60.3**

 

Naïve

65

3,080,128

−0.040 (−0.103, −0.016)

53.6**

 

Region

Europe

43

2,333,441

−0.210 (−0.446, −0.093)

58.6**

 

US/Canada

78

750,854

−0.108 (−0.208, −0.054)

72.7**

 

Rest World

23

8383

−0.059 (−0.407, −0.013)

0

 

Behaviourª

Energy

 

47

719,059

−0.094 (−0.133, −0.055)

67.7**

6.6

Appliances

12

108,077

−0.036 (−0.129, 0.058)

22.6

2.5

Transportation

 

29

2,245,972

−0.136 (−0.183, −0.089)

98.4**

9.6

Car use

21

2,242,781

−0.036 (−0.039, −0.034)

0

2.5

Water

 

42

124,082

−0.052 (−0.079, −0.025)

40.1**

3.7

Towel

18

8909

−0.168 (−0.271, −0.064)

47.8**

11.9

Food waste

4

218

−0.231 (−0.518, 0.056)

21.6

16.3

Meat

7

666

−0.239 (−2.81, 0.008)

36.8

16.9

Recycling

23

2766

−0.457 (−0.595, −0.319)

69.9**

32.3

Intervention

Information

53

2,354,243

−0.048 (−0.075, −0.021)

34.7**

3.4

Social comparison

32

719,756

−0.077 (−0.108, −0.046)

72.2**

5.4

Engagement

 

38

10,486

−0.253 (−0.336, −0.170)

71.8**

17.9

Commitment

10

1446

−0.480 (−0.704, −0.255)

75.8**

33.9

Appeals

10

5952

−0.266 (−0.445, −0.086)

70.5**

18.8

Nudges

11

795

−0.352 (−0.492, −0.212)

0

24.9

  1. Note: k = #estimates; N = sample size; I2 = Heterogeneity; POB = probability of benefit (effect size d/√2)
  2. **p   < 0.05
  3. aThe total aggregate sample size per analysis of behaviour is 3,092,763—an additional 85 individuals than the overall 3,092,678. This difference is due to a single study (Kurz et al. 2005 in Supplementary References) testing the effect of an intervention in both water and energy and, thus, its sample (N = 85) was accounted in both behaviours