Blog

Questions 16-187 additionally rely on the following table of subgroup averages for the outcome metric from the experiment.   Subgroup Average outcome metric Below median age (all) 25 Below median age , control 24 Below median age, savings rate prompt 26 Above median age, all  43 Above median age,  control 40 Above median age,  savings rate prompt 48    

The organization is planning a future experiment. They calculate that, with 80% power, 5% significance, and their anticipated sample size, the standardized minimum detectable effect is 0.14. If they believe the intervention increases contributions by 10% of the baseline mean, should the experiment be implemented as designed? Hint: you need to use the numbers in Savings Experiment Table 1.

We have data from an experiment and we want to use subgroups to improve precision. Which subgroups would be most helpful to focus on?

consider the following python code and output from the Artea experiment: f1 = ‘revenue_after ~ C(test_coupon):C(shopping_cart)’m1 = smf.ols(data=data, formula=f1).fit()======================================================                                              coef——————————————————Intercept                                   5.4817C(shopping_cart)[T.1]                       7.6657C(test_coupon)[T.1]:C(shopping_cart)[0]    -0.7395C(test_coupon)[T.1]:C(shopping_cart)[1]     1.3806========================================================= What is the effect of test_coupon on revenue_after, for the subgroup that previously had an item in their shopping cart?

What output is produced by the below command?   output = smp.tt_ind_test_power( alpha=0.05, nobs1=1000, power =0.90)

Describe a way in which you would measure a person’s temporal/duration discrimination ability. (Bonus point: If you could use/apply any of the psychoacoustic method that you learned from your discussion posts. You should name the psychoacoustic method appropriately)

Every year Amazon sends out winter gift catalogs, which are full of toys for kids. They would like to send them to families with young children, but they don’t know which addresses have young children at them. Identifying addresses with young children is a…

In python, we have a data frame  with outcome variable y and treatment variable treat, which is categorical variable which takes on the values 0, 1, and 2. Which code sets up the formula for estimating the effect of treatment on our outcome?    

 What’s a confound?

An organization conducts an A/B test with adequate power. Upon analyzing the data, they find that condition B performs statistically significantly better than condition A. What of the following would be most important for them to consider before deciding whether to scale up the intervention?