search

QuickPay-Operational-Performance / Data-and-code

Data and code for econometric analysis
0 stars 0 forks source link

Stage-dependent treatment effects #69

Closed JNing0 closed 1 year ago

JNing0 commented 3 years ago

I have written down the motivation and model for project stage-dependent treatment effects here. (I've also created a wiki page.)

This is something for us to reflect on for the paper. As explained in the "motivation" part of the write-up, I feel that this direction highlights the distinguishing feature of projects: that it has a life span. The stage of a project's life probably matters when the vendor makes a decision. In fact, I feel that even a null result might be of interest, as it says that, contrary to the expectation, project stage actually does not matter. Payment acceleration has a uniform influence, regardless of the stage of the project in its life.

vibhuti6 commented 2 years ago

Hi everyone, I just thought it may be better to post the results here instead of opening a new issue.

Quintiles of project stage

Screen Shot 2022-04-08 at 6 23 22 PM Screen Shot 2022-04-01 at 11 54 04 AM

Deciles of project stage

Screen Shot 2022-04-08 at 6 23 16 PM Screen Shot 2022-04-01 at 11 54 11 AM

Update (April 15, 2022): Fixed the error in outlier of max stage for last quintile/decile.

vob2 commented 2 years ago

Looks good, thanks! Vlad Babich

Professor of Operations and Information Management
McDonough School of Business
Georgetown University
37th and O Streets, NW
531 Rafik B. Hariri Building
Washington, DC 20057

Phone:  202-687-4782 <http://voice.google.com/calls?a=nc,%2B12026874782>
E-mail:  ***@***.***

On Fri, Apr 1, 2022 at 9:30 PM Vibhuti Dhingra @.***> wrote:

Hi everyone, I just thought it may be better to post the results here instead of opening a new issue.

  • Continuous version: Treatment effect increases with project stage. Please see Table 3 on Page 11 here https://github.com/QuickPay-Operational-Performance/Data-and-code/blob/master/qp_first_implementation/qp_first_pc_delay-2.pdf .
  • Discrete version: I plotted the treatment effect for quintiles and deciles of project stage. Please see the graphs below. The estimates that are significant are shown in black, insignificant ones are in gray.
    • Takeaway: The treatment effect is significant only for projects that are at least halfway through. Specifically, delays only increase for small projects that have completed at least 44% of their lifespan.

Quintiles of project stage

[image: Screen Shot 2022-04-01 at 11 54 30 AM] https://user-images.githubusercontent.com/47764977/161358136-21c19035-bd72-4e8d-94ac-f557160d7eb0.png

[image: Screen Shot 2022-04-01 at 11 54 04 AM] https://user-images.githubusercontent.com/47764977/161358140-8e281cba-bc07-4576-853e-fca3f36605fb.png

Deciles of project stage

[image: Screen Shot 2022-04-01 at 11 54 47 AM] https://user-images.githubusercontent.com/47764977/161358817-6745f5f2-ef9f-4689-aed3-d1ac21de1df5.png

[image: Screen Shot 2022-04-01 at 11 54 11 AM] https://user-images.githubusercontent.com/47764977/161358912-9eac6c6e-4140-48f3-b5ef-9475261aada4.png

— Reply to this email directly, view it on GitHub https://github.com/QuickPay-Operational-Performance/Data-and-code/issues/69#issuecomment-1086460523, or unsubscribe https://github.com/notifications/unsubscribe-auth/AN73EK372GAMVD2HPEW2J6DVC6PKZANCNFSM43RL5ACQ . You are receiving this because you were assigned.Message ID: <QuickPay-Operational-Performance/Data-and-code/issues/69/1086460523@ github.com>