Decide on sampling procedure

[timo-b-roettger]

See measured variables from Hardwicke et al. (2021).

Sampling is done via Scopus

Following the psych journal (changed years to reflect two points in time (a) prior to replication crisis and (b) most recent).

• sourcetype: journal (j)
• document types: articles (ar) + review (re)
• publication year: 2008-2009 + 2018-2019
• language: English

suggestion for subject-terms following ASJC Code:

• SUBJTERMS(1203) = Language and Linguistics
• SUBJTERMS(3310) = Linguistics and Language

Corresponds to Scopus query:
SRCTYPE ( j ) AND ( SUBJTERMS ( 1203 ) OR SUBJTERMS ( 3310 ) ) AND ( LIMIT-TO ( DOCTYPE,"ar" ) OR LIMIT-TO ( DOCTYPE,"re" ) ) AND ( LIMIT-TO ( PUBYEAR,2008) OR LIMIT-TO ( PUBYEAR,2009) OR LIMIT-TO ( PUBYEAR,2018) OR LIMIT-TO ( PUBYEAR,2019) ) AND ( LIMIT-TO ( LANGUAGE,"English" ) )

A possible problem with this approach is that there are journals that clearly aren't linguistic (e.g. "Notes & Queries").

SOLUTION: Manually check paper and if not on language, then put back on "questionable" pile and cross-validate with someone else.

A possible opportunity for exploration comes with other the discipline codes. Our journals overlap with the following macro disciplines (top 5):

• Psychology (14%)
• Health Professions (9%)
• Computer Science (8%)
• Neuroscience (5%)
• Medicine (3%)

[melroethlisberger]

• Selection of documents to search for: OK
• Search query in scopus: OK, I tried it with the tools I had at hand, not using LIMIT-TO but just OR and AND:

Query: SRCTYPE ( j )  AND  ( SUBJTERMS ( 1203 )  OR  SUBJTERMS ( 3310 ) )  AND  LANGUAGE ( english )  AND  ( PUBYEAR  =  2008  OR  PUBYEAR  =  2009  OR  PUBYEAR  =  2018  OR  PUBYEAR  =  2019 )  AND  ( DOCTYPE ( ar )  OR  DOCTYPE ( re ) )

= Same number: 48'066 results.

• Export: there is a maximum of 20'000 articles to be exported and only information on citation can be exported.