feat(script): Display strength of findings evidence in binned format

- for weak evidence, + for substantiated evidence, ++ for strong evidence.

Currently separates between internal and external, and uses the bins:
0.0-2.9 for weak
3.0-5.9 for substantiated
6.0-... for strong evidence

02-data/supplementary/findings-institutional.csv

@@ -0,0 +1,26 @@
+area of policy,findings,channels,studies
+minimum wage,mixed evidence for short-/medium-term income inequality impacts,can lead to income compression at higher-earner ends,Wong2019;Sotomayor2021;Alinaghi2020;Gilbert2001
+,some evidence for long-term inequality decrease,job loss offsets through higher wages,Sotomayor2021;Chao2022;SilveiraNeto2011
+,,some spatial transfer from urban manufacturing sectors to rural agricultural sectors
+,bad targeting can exacerbate existing inequalities,negative effect on women's hours worked if strong household labour divisions,Alinaghi2020;Wong2019;Militaru2019
+,,low-earners sometimes secondary high-income household earners while low-wage households have no earners at all
+,"potential impact larger for single parents, rural/disadvantaged locations",women more affected if they make up large share of low-wage earners,Alinaghi2020;Gilbert2001;SilveriaNeto2011
+paid leave,evidence for significant increase in rtw after childbirth,esp. disadvantaged women benefit due to no prior employer-funded leave,Broadway2020;Dustmann2012;Davies2022
+,some evidence for positive rtw effects to occur with medium-/long-term time delay,short-term exit but no long-term increase to hiring pattern discrimination,Broadway2020;Dustmann2012
+,,can exacerbate existing household labour division
+,mixed evidence for fixed-/short-term contracts counter-acting effect on rtw,fixed-term contracts often insufficiently covered by otherwise applicable labour regulation,Davies2022;Mun2018
+collective bargaining,evidence for decreased income inequality with strong unionisation,stronger collective political power vector enables more equal redistributive policies,Alexiou2023;Cardinaleschi2015
+,,"increased probability for employment on formal, standard employment contract"
+,marginal evidence for increased income/representation of women/minorities in workforce/management,internal heterogeneity due to predominantly affecting median part of wage distribution,Ferguson2015;Ahumada2023
+,,self-selection of people joining more unionised enterprises/organisations/sectors
+,,"depending on targeting of concurrent policies can bestow more benefits on men, increasing horizontal inequalities"
+protective environmental policies,evidence for decrease in spatial inequality,increased employment probability through large-scale rural energy projects,Kuriyama2021
+,mixed evidence for increase of existing inequalities,elite policy capture can exacerbate existing social exclusion & disadvantages,Kuriyama2021;Stock2021
+workfare programmes,evidence for decrease of vertical inequality,,Whitworth2021;Li2022
+,evidence for possibility of increased spatial inequalities,bad targeting increases deprivations for already job-deprived areas,Whitworth2021
+,evidence for effective outcomes dependent on on prior material equalities,prior inequalities such as land ownership can lead to political capture and less effective policies,Li2022
+social protection,evidence for conditional cash transfers producing short- and long-term inequality reduction,production of short-term cash influx,Debowicz2014
+,,conditioning on school attendance can decrease educational inequalities over long-term
+,mixed evidence for childcare subsidies decreasing gender inequalities,lifting credit constraints greater effect on low-income households,Hardoy2015;Debowicz2014
+,evidence for stagnating income replacement rates exacerbating existing vertical inequalities,benefit levels unlinked from wages can widen division between income groups,Wang2016
+,healthcare subsidy impacts strongly dependent on correct targeting,dependence on non-participation in labour market may generate benefit trap,Carstens2018

scoping_review.qmd

@@ -631,12 +631,15 @@ g = sns.PairGrid(validities[["internal_validity", "external_validity", "identifi
 
 
 ```{python}
+#| label: tbl-findings-institutional
+#| tbl-cap: Main findings summary institutional policies
 findings_institutional = pd.read_csv("02-data/supplementary/findings-institutional.csv")
 from src import prep_data
 import math
 
-validities = prep_data.calculate_validities(by_intervention)
+EVIDENCE_STRENGH=["\-","\-","\-","\+","\+","\+","++","++","++","++","++","++","++","++","++","++"]
-valid_subset = validities[["internal_validity", "external_validity", "citation"]].fillna(1.0).drop_duplicates(subset=["citation"]).sort_values("internal_validity")
+
+valid_subset = prep_data.calculate_validities(by_intervention)[["internal_validity", "external_validity", "citation"]].fillna(1.0).drop_duplicates(subset=["citation"]).sort_values("internal_validity")
 def combined_validities(df_in, column: str = "internal_validity"):
     if not isinstance(df_in, str):
         return
@@ -646,14 +649,14 @@ def combined_validities(df_in, column: str = "internal_validity"):
         if len(new) > 0 and not math.isnan(new.iat[0]):
             combined += new.iat[0]
     if combined:
-        return combined
+        return EVIDENCE_STRENGH[int(combined)] + f" ({str(combined)})"
-    return 0.0
+    return "\-"
 def combined_external(df_in, column: str = "external_validity"):
     return combined_validities(df_in, column)
 
 findings_institutional["internal_validity"] = findings_institutional["studies"].apply(combined_validities)
 findings_institutional["external_validity"] = findings_institutional["studies"].apply(combined_external)
-md(tabulate(findings_institutional[["area of policy",  "internal_validity", "external_validity", "findings", "channels"]], showindex=False, headers="keys", tablefmt="grid"))
+md(tabulate(findings_institutional[["area of policy",  "internal_validity", "external_validity", "findings", "channels"]].fillna(""), showindex=False, headers="keys", tablefmt="grid"))
 ```
 
 {{< landscape >}}