ECON $1629$

Field Experiment Assignment

$.$

Submit to Canvas your Field Experiment proposal. You can work in pairs or individually.

$1)$ What is the question$/$hypothesis$?$

$2)$ Why is it important to know the answer? Think about this when answering: Once we

know the answer to your question, what might we do differently?

$3)$ Why can$$t you use observational data to test your hypothesis? Be very clear about your

threats to identification and the probable direction of bias in an OLS regression based

on observational data.

$4)$ What is the design of your experiment? Be sure to define:

a$.$ Who are your subjects? Where$/$how will you recruit them?

b$.$ How will you define your treatment group? Your control group?

c$.$ What is your treatment? How will it be administered?

d$.$ How does your treatment relate to the main variable of interest indicated in

question $1?$ Eg$,$ are they the same? If not, how are they related?

e$.$ What data will you be collecting? When and how?

$5)$ How will you check whether your assignment of treatment$/$control status will result in

balanced groups? What data do you need for that? Be specific.

$6)$ Who are the compliers? The always takers? The never takers?

$7)$ What will be the intent to treat effect? Explain and provide an estimating equation. How

do you interpret your coefficient of interest?

$8)$ What will be the first stage for estimating the treatment on the treated effect? Explain

and provide an estimating equation. Interpret your coefficient of interest.

a$.$ How do you know that exclusion and relevance have been met in this case?

$9)$ What is the second stage for estimating the treatment on the treated effect?

a$.$ What will determine if it is different from the intent to treat effect?

$10)$ How generalizable will your results be$?$ Explain your reasoning $.$