PP279/MacCoun
PROGRAM EVALUATION PROPOSAL ASSIGNMENT
DUE DATES FOR THE PREPROPOSAL
The Preproposal
Prior to any serious work on the proposal, your team should submit a 1-page letter of intent--concisely describing the central policy issue you are addressing, and the specific research questions you plan to answer. (You needn’t explain your methodology here.) You can evaluate an existing program, a program that will soon exist, or a hypothetical program you’d like to exist.
The Proposal
The proposal must be submitted as an email attachment (maccoun@berkeley.edu) in MS Word format. This will allow me to embed my comments directly in the file and send it back to all of you.
Your proposal should include the sections described below; very approximate space-and-a-half page guidelines appear in brackets. If you wrote the maximum number of pages for each section, your proposal would be roughly 30 pages plus bibliography (we’ll skip the vitae); at minimum, you’ll probably need at least 20 pages. These pages are GUIDELINES; not strict rules. Use your own judgment about what your topic requires. When in doubt about anything, the fundamental deciding principle isn’t “what does Rob want” but rather “what would encourage a funder to decide it is worth risking a large sum of money on our organization?”
[1 page]: Brief summary of issues,
questions, and general strategy
[2-3 pages]: Explain the policy
issues that motivate your proposal. Assume the reader is a non-expert who is
fairly conversant in current policy issues. Briefly foreshadow the program you
plan to evaluate.
[4-6 pages]: What do and don’t we
know about your research questions? Be critical in your evaluation of the
existing literature; draw upon your vast and deep knowledge of experimental and
quasi-experimental methodology obtained in this course, including threats to
internal, external, construct, and statistical conclusion validity (e.g.,
power). Assume that while your reader may not be an expert on the policy
context, he or she (more likely, they--sometimes 10-15 different
reviewers) knows enough about methodology that you don’t have to ‘start from
scratch’ in explaining your critique.
[1-2 pages]
[2-4 pages]. Please describe it in
enough detail that the reader understands what it is you will be evaluating.
(a) Who will be the treatment provider? (b) What exactly is the treatment/
intervention? (c) What is the target population from which you’ll draw your
sample? Please be sensitive to ethical issues, and don’t propose anything that
(in your estimate) wouldn’t get passed a human subjects protection committee.
[4-6 pages]. (a) Under optimistic
conditions of full agency cooperation and minimal political controversy,
describe the experimental design you propose to use. (Use or adapt the
Campbell-Stanley notation, but you aren’t limited to their exact designs.) Be
clear about your unit of analysis. (b) Describe your planned sample size per
condition, and justify it. (c) Describe the different conditions and how you
plan to determine who gets in which condition. (d) Briefly list your major
dependent measures, but this time you don’t have to produce an actual
questionnaire. Be sure to include manipulation checks among your measures. (e)
I’d like to see at least a few sentences about how you plan to analyze the
data, but I won’t hold you accountable for any methods you have been trained to
use yet. (f) Assess the design with respect to threats to validity (all types
of validity)--be specific with respect to your particular topic and setting.
[4-6 pages]: (a) Identify possible
reasons why Design #1 might become infeasible. (b) Describe the
quasi-experimental design you plan to use as a "fall-back" option.
(c) Explain what threats to validity you anticipate as a result of this
fall-back option, and evaluate their potential consequences for your policy
conclusions. Mention any additional dependent measures you might use to
strengthen your inferences about the design. (d) In explaining your choice of
quasi-experimental design, feel free to mention other alternatives you
rejected. (Use or adapt the Campbell-Stanley notation, but you aren’t limited
to their exact designs.) (e) Describe your planned sample size per condition,
and justify it. (f) Again, I’d like to see at least a few sentences about how
you plan to analyze the data, but I won’t hold you accountable for any methods
you have been trained to use yet.
These are often redundant and/or
corny, but better than simply ending with purely technical material.
[whatever pages you need; probably 2-4?]
Same rules as before: (1) It is fair game to email me questions about your proposal. (2) It is also fair game to discuss this (unlike the homework) with others. Obviously, my hope is that you’ll discuss it with your teammates! In addition, you can discuss it with members of other teams. But I encourage you to see how far you can get without consulting experts (e.g., professors, survey professionals).