29 lines
648 B
Markdown
29 lines
648 B
Markdown
|
# Validity estimators
|
||
|
|
|
||
|
|
For a general concept description see ../03-documentation/terms_of_reference-key_terms.md#validity
|
||
|
|
|
||
|
|
From Maitrot2017 -> Section 4, Figure 3 and Appendix table notes
|
||
|
|
They rank *only* quasi-experimental/experimental
|
||
|
|
|
||
|
|
Do we too?
|
||
|
|
|
||
|
|
## internal
|
||
|
|
|
||
|
|
|
||
|
|
0-5 ranking using study design/method:
|
||
|
|
|
||
|
|
OLS=2 (ordinary least squares)
|
||
|
|
DID=3 (difference-in-difference)/(-in-difference)
|
||
|
|
DM=3 ()
|
||
|
|
PSM=3.5 (propensity score matching)
|
||
|
|
IV=4 (instrumental variable)
|
||
|
|
RD=4.5 (regression discontinuity)
|
||
|
|
RCT=5
|
||
|
|
|
||
|
|
## external
|
||
|
|
|
||
|
|
non-representative surveys=2
|
||
|
|
sub-nationally representative surveys=3
|
||
|
|
nationally representative surveys=4
|
||
|
|
surveys based on census data=5
|