CFPS 2020 — Analysis-sample codebook
(internet use & labour income; n = 6948)

IDENTIFIERS
  pid           individual id (CFPS)
  commid        community id  (= fid_cid20)  — used to build the instrument
  provcd        province code (= provcd20)   — province fixed effects

OUTCOME
  emp_income    annual labour income, CNY (raw, kept > 0)
  ln_income     ln(emp_income)               — main dependent variable
  incomeb       primary-job annual income, CNY (robustness)
  ln_incomeb    ln(incomeb)                  — robustness dependent variable

TREATMENT
  internet      =1 if online via mobile (qu201) OR computer (qu202)
  internet_mob  =1 if online via mobile  (qu201)
  internet_pc   =1 if online via computer (qu202)

USE TYPE (defined for internet users; raw indicators kept full-sample)
  use_learn     =1 online learning  (qu94)   — productive use
  use_shop      =1 online shopping  (qu92)   — mixed use
  use_game      =1 online gaming    (qu91)   — consumptive use
  use_time      daily minutes online (qu201a, >=0)
  ln_time       ln(use_time + 1)

MEDIATORS
  info_internet importance of internet as information source (qu802, 1-5)
  contract      holds a formal labour contract (qg5, 0/1)

INSTRUMENT
  iv_comm       community leave-one-out internet penetration rate

CONTROLS
  age           age in years (16-60)
  age2          age squared
  gender        male = 1
  eduyear       years of education (= cfps2020eduy)
  health        self-rated health (qp201, 1 best .. 5 worst)
  urban         urban residence = 1 (= urban20)
  married       currently married = 1 (marriage_last == 2)

SUBGROUP HELPERS
  edu_high      =1 if eduyear >= 12
  age_group     1 = 16-30, 2 = 31-45, 3 = 46-60
  employ_now    =1 if currently employed (employ == 1)
