Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
Power analyses based on results from quantitative pretest
Hypothesis 1: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the control group.
Alpha = 0.05 Alpha = 0.001
Alpha 0,05 0,001
Power 0,999986 0,997783
N 1500 1500
N1 750 750
N2 750 750
Delta -0,3948 -0,3948
m1 2,952941 2,952941
m2 2,55814 2,55814
Std. Dev. 1 1,318344 1,318344
Std. Dev. 2 1,164218 1,164218
Hypothesis 2: Respondents are more likely to support the Co2 policy if they receive the generational treatment, as compared to the control group.
Alpha = 0.05 Alpha = 0.001
Alpha 0,05 0,001
Power 0,200147 0,014756
N 1500 1500
N1 750 750
N2 750 750
Delta -0,07489 -0,07489
m1 2,952941 2,952941
m2 2,878049 2,878049
Std. Dev. 1 1,280865 1,280865
Std. Dev. 2 1,318344 1,318344
Hypothesis 3: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the economic treatment.
Alpha = 0.05 Alpha = 0.001
Alpha 0,05 0,001
Power 0,998728 0,953673
N 1500 1500
N1 750 750
N2 750 750
Delta 0,319909 0,319909
m1 2,55814 2,55814
m2 2,878049 2,878049
Std. Dev. 1 1,164218 1,164218
Std. Dev. 2 1,318344 1,318344
Order effect hypotheses:
Hypothesis 4: Respondents are more likely to support the Co2 policy if they receive the generational treatment, as compared to the mixed treatment group.
Hypothesis 5: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the mixed treatment group.
Hypothesis 6: There is an order effect in whether respondents see the economic or the generational treatment first.
H4 H5 H6
Alpha 0,05 0,05 0,05
Power 0,999867 0,133363 0,051833
N 1500 1500 750
N1 750 750 375
N2 750 750 375
Delta 0,372167 0,052258 -0,01163
m1 2,505882 2,505882 2,511628
m2 2,878049 2,55814 2,5
Std. Dev. 1 1,249986 1,249986 1,241583
Std. Dev. 2 1,318344 1,164218 1,273558
Court treatment hypothesis:
Hypothesis 7: Respondents are less likely to support the Co2 policy if they receive the court treatment, as compared to the control group.
Alpha = 0.05 Alpha = 0.001
Alpha 0,05 0,001
Power 1 1
N 6000 6000
N1 3000 3000
N2 3000 3000
Delta 0,349827 0,349827
m1 2,54491 2,54491
m2 2,894737 2,894737
Std. Dev. 1 1,240417 1,240417
Std. Dev. 2 1,265156 1,265156
|
After
Power analyses based on results from quantitative pretest
Hypothesis 1: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the control group.
Power analyses based on pretests of our study design show that we can detect a change of half a point on our answer scale between framing groups with a power of 100% at the conventional statistical significance level of 5% and 1%.
Hypothesis 2: Respondents are more likely to support the Co2 policy if they receive the generational treatment, as compared to the control group.
Power analyses for this hypothesis shows that we can detect a change of half a point on our answer scale between framing groups with a power of 20% at the conventional statistical significance level of 5%.
Hypothesis 3: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the economic treatment.
Power analyses for H3 shows that we can detect a change of half a point on our answer scale between framing groups with a power of 99.8% (95.3%) at the conventional statistical significance level of 5% (and 1%).
Order effect hypotheses:
Hypothesis 4: Respondents are more likely to support the Co2 policy if they receive the generational treatment, as compared to the mixed treatment group.
Power analyses for H4 shows that we can detect a change of half a point on our answer scale between framing groups with a power of 99.9% at the conventional statistical significance level of 5%.
Hypothesis 5: Respondents are less likely to support the Co2 policy if they receive the economic treatment, as compared to the mixed treatment group.
Power analyses for H5 shows that we can detect a change of half a point on our answer scale between framing groups with a power of 13.3% at the conventional statistical significance level of 5%.
Hypothesis 6: There is an order effect in whether respondents see the economic or the generational treatment first.
Power analyses for H5 shows that we can detect a change of half a point on our answer scale between framing groups with a power of 5.2% at the conventional statistical significance level of 5%.
Court treatment hypothesis:
Hypothesis 7: Respondents are less likely to support the Co2 policy if they receive the court treatment, as compared to the control group.
Power analyses for H7 shows that we can detect a change of half a point on our answer scale between framing groups with a power of 100% at the conventional statistical significance level of 5% and 1%.
See documentation for full information and results from the power analyses.
|