Sebastian Tello
  • Home
  • CV
  • Contact
  • Research
  • Resources
  • RMDA
  • APP
InstagramBluesky

Quiz 7

  1. Suppose you’re using an RD design to see how tax benefits affect firm profits. Firms receive with 50 full time workers or less receive a 10% tax relief relative to firms with more than 50 full time employees. Firms are responsible for reporting employment measures. Which of the following might be a concern?
    1. This reduces sample size near the cutoff
    2. A concern would be the accurate reporting of full time workers.
    3. A concern would be that this is a fuzzy RD
    4. This implies the treatment is measured with error, which RD can always ignore
      5. ‣
      Answer

      B

  2. Suppose firms also receive disaster recovery funding when they have below 50 workers. Is this a problem? Why?
    1. It’s not a problem because it’s a separate treatment
    2. It is a problem because it means we can’t attribute a change in outcomes at the cutoff to the treatment we want to examine.
    3. It’s not a problem because disaster recovery funding has a different effect than tax benefits
    4. It is a problem because we can’t control for whether firms receive disaster recovery funding
      5. ‣
      Answer

      It is a problem because it means we can’t attribute a change in outcomes at the cutoff to the treatment we want to examine.

  3. HIV is the second largest pandemic in world history (behind only the black plague) and hence an importance focus for public health. One of the main tools to “fight” the pandemic has been a retroviral therapy called ART. This drug helps maintain the virus from spreading more quickly for the people with the virus. Since the availability of this drug hasn’t reached demand, some countries focus on providing ART only to certain people. One of the measures they use to decide who to offer this drug to is people with certain level of white blood cells. CD4 is a measure of the white blood cells someone has– the lower the blood cells, the more immunocompromised. Hence, the policy stated that people with CD4 counts below the cutoff became eligible for ART. A paper uses a RD design to measure the effect of anti-retroviral therapy (ART) on mortality for those with HIV/AIDS in South Africa. Answer the following questions with this context in mind
    1. (3.1) What is the outcome?
      1. HIV Rate
      2. HIV related mortality
      3. Taking up ART
      4. White blood cell counts
        5. ‣
        Answer

        HIV related mortality

    2. (3.2) What is the treatment?
      1. HIV Rate
      2. HIV related mortality
      3. Taking up ART
      4. White blood cell counts
        5. ‣
        Answer
        1. Taking up ART
    3. (3.3) What is the running variable?
      1. HIV Rate
      2. HIV related mortality
      3. Taking up ART
      4. White blood cell counts
        5. ‣
        Answer

        White blood cell counts

    4. (3.4) Researchers started their paper with the graph below. What is the purpose of the graph?
        5. image
      1. to demonstrate that there are fewer people with high CD4 counts than low CD4 counts
      2. to demonstrate that people are continuously and randomly distributed around the cutoff
      3. to demonstrate that the cutoff value for ART is 200 cell/µL
      4. to demonstrate that ~4500 participants were involved in the study
        5. ‣
        Answer

        to demonstrate that people are continuously and randomly distributed around the cutoff

    5. (3.5) Based on the following graph, the section of the graph where D=1 is: (Note: blue dots represent the probability of ART initiation within 3 months, while red is the probability of starting within 12 months)
        6. image
      1. Left of the cutoff
      2. Right of the cutoff
        3. ‣
        Answer

        left of cutoff

    6. (3.6) Looking at the graph below, which of the following models would be the most appropriate model to start with?
        7. image
      1. Mortality=β0+β1I(CD4i<200)+β2(CD4i−200)+ϵiMortality=\beta_0+\beta_1I(CD4_i<200)+\beta_2(CD4_i-200)+\epsilon_iMortality=β0​+β1​I(CD4i​<200)+β2​(CD4i​−200)+ϵi​
      2. Mortality=β0+β1I(CD4i<200)i+β2(CD4i−200)+β3D(CD4i−200)+ϵiMortality=\beta_0+\beta_1I(CD4_i<200)_i+\beta_2(CD4_i-200)+\beta_3D(CD4_i-200) +\epsilon_iMortality=β0​+β1​I(CD4i​<200)i​+β2​(CD4i​−200)+β3​D(CD4i​−200)+ϵi​
      3. Mortality=β0+β1I(CD4i<200)i+β2(CD4i−200)+β3(CD4i−200)2+ϵiMortality=\beta_0+\beta_1I(CD4_i<200)_i+\beta_2(CD4_i-200)+\beta_3(CD4_i-200)^2 +\epsilon_iMortality=β0​+β1​I(CD4i​<200)i​+β2​(CD4i​−200)+β3​(CD4i​−200)2+ϵi​
      4. Mortality=β0+β1I(CD4i<200)i+β2(CD4i−200)+β3I(CD4i−200)+β4(CD4i−200)2+β5(I(CD4i<200))(CD4i−c)2+ϵiMortality=\beta_0+\beta_1I(CD4_i<200)_i+\beta_2(CD4_i-200)+\beta_3I(CD4_i-200 )+\beta_4(CD4_i-200)^2+\beta_5(I(CD4_i<200))(CD4_i-c)^2+\epsilon_iMortality=β0​+β1​I(CD4i​<200)i​+β2​(CD4i​−200)+β3​I(CD4i​−200)+β4​(CD4i​−200)2+β5​(I(CD4i​<200))(CD4i​−c)2+ϵi​
        5. ‣
        Answer

        C

    7. (3.7) Using the graph from 3.6, if we were to use model C or D from question 3.6, what sign would we expect the value of β1\beta_1β1​ to be
      1. postive
      2. negative
        3. ‣
        Answer

        negative