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

Quiz 9

  1. Researchers conduct a study to assess whether increasing the minimum wage leads to lower teen unemployment. They observe data aggregated at the state-year level, from 50 U.S. states over 10 years (2010–2019). Some states raise their minimum wage during this period; others do not. The outcome variable is the teen employment rate (ages 16–19) in each state each year. The basic regression is Employmentst=β0+β1MinWagest+ast+ϵstEmployment_{st} = \beta_0+\beta_1MinWage_{st}+a_{st}+\epsilon_{st}Employmentst​=β0​+β1​MinWagest​+ast​+ϵst​. Why would a researcher include state fixed effects in this model?
    1. To control for time-variant differences between states
    2. To control for time-invariant differences between states
    3. To ensure the sample is nationally representative
    4. To eliminate measurement error in the minimum wage variable
      5. ‣
      Answer

      B

  2. What might year fixed effects account for?
    1. Differences in state policies that change every year
    2. The effect of California’s specific wage laws
    3. Nationwide shocks that affect all states simultaneously, but do not differ across states
    4. Seasonal variation within each year
      5. ‣
      Answer

      C

  3. A critic argues: 'States that raise the minimum wage also tend to have stronger unions, and unions also boost teen employment.' Why is this not a threat to the fixed effects estimate?
    1. Union strength is different across each state, so the state-FE do not capture it.
    2. Union strength is not a threat because it’s impact is captured by the year-FE.
    3. If union strength varies across states but is stable within states over time, it is already accounted for by fixed effects
    4. The model controls for all omitted variables by construction
    5. The correlation between union strength and wages is too small to matter.
      6. ‣
      Answer

      C

  4. In the middle of their research, states receive an email containing additional data from 2020-2025. Should we add year fixed effects to this model now that we have additional data?
    1. Yes, to capture common shocks across states like the effects of the COVID-19 pandemic between 2020 and 2022.
    2. Yes, because we need a minimum of 10 years of data to run a regression with year fixed effects.
    3. No, because minimum wage policies are implemented in different years across states, so the year-FE would absorb the policy effect.
    4. No, because it makes our true regression too long.
      5. ‣
      Answer

      A