Table 1 Average treatment effect (ATE) estimates

From: Eliciting attachment security with social norm messages is linked to reduced energy consumption in extreme heat in the United Arab Emirates

 

I

II

III

IV

(Social Norm) × Post

−0.502***

−0.502***

−0.492***

−0.502***

 

(0.064)

(0.064)

(0.063)

(0.064)

(Social Norm + Attachment) × Post

−0.820***

−0.820***

−0.789***

−0.820***

 

(0.113)

(0.113)

(0.109)

(0.113)

Difference between the two arms

−0.318***

−0.318***

−0.298***

−0.318***

p value

0.001

0.001

0.002

0.001

Weather

No

No

No

Yes

Apartment fixed effects

Yes

Yes

Yes

Yes

Day of the week fixed effects

Yes

Yes

Yes

Yes

Week fixed effects

No

Yes

Yes

Yes

Building × Week fixed effects

No

No

Yes

No

Observations

5544

5544

5544

5544

R2 (Overall)

0.816

0.817

0.820

0.819

F-statistics

28,032

53,557

65,366

58,130

  1. All models examining the risk of experimental stimuli’s effect on total daily household energy usage (kWh) during the 4-week intervention period. In all four columns, we report regression coefficients with Driscoll and Kraay standard errors as estimated by Eq. (1). Errors are not clustered, given that the randomization is performed at the apartment level. The third row shows the difference in ATEs between the two treatment groups: the group received a standard social norm message, and the group received an enhanced social norm message. Standard errors in parentheses; p value < 0.10*; p value < 0.05**; p value < 0.01***.
Back to article page