# Meeting Miguel 2024-02-09

achieved:

- [ ] remaining study batch screened
    - added active labour market policy studies
    - starting to wrangle synthesis:
        - better categorization for some and did dual categorizations, e.g. job search assistance or subsidized employment/wages partly into ALMP
        - Moved away again a little from strict framework adherence to make more sense for the synthesis -> workfware programmes

- [ ] focus now:
    - table reviewing main findings
    - screen the remaining 200 studies
    - improve the writing on the synthesis itself
        - to adhere a little closer to

- [ ] questions:
    - logistical:
        - create extraction matrices with main/statistical findings
            - are we going to put those in the main Appendix?
              or do we just pass on the CSV extraction as the data?
        - is there a specific length I am going for for the end result?
        - is it fine if for now I leave intro/conclusion more vague to fit in with rest of the document
    - writing:
        - categories overlap sometimes, of course
            - a study on NREGS in India which has minimum wage effects as primary channel
            - a study on vocational training which fuses it with workfare program (ALMP)
        - I would put the focus on 'per-area' conclusions, not the 'per-study' 
    - validity visualizations
        - built a simple scoring rubric of 0-not fulfilled, 1-implicitly fulfilled, 2-explicitly fulfilled
        - internal: (design + method) - (selection bias + measurement bias + confounding bias + performance bias)
        - external: (representativeness) - (generalizability concerns + population invalidity + ecological invalidity)
        - internal:
            - design (observational/qualitative, correlational, causal(quasi-experimental/experimental))
            - confounding variable control (no mention, limits mentioned, limits mentioned and explicitly discussed)
            - data collection (not mentioned, sources mentioned, sources/sample size/observation-length given)
            - statistical control (no robustness checks, data robustness mentioned, explicit robustness checks)
            - biases (no mention, possible biases mentioned, explicitly controlled for/influcence explained)
        - external:
            - generalizability (no mention, implicitly generalizable, explicitly generalizable)
            - representativeness of sample (no mention/sub-national, national, regional/global)
            - ecological validity (no mention, implicit policy implication, explicit policy implication)

- [ ] TODO
    - main findings table still remaining
    - write it out 
    - shorten?