Crystal Clear: Investigating Databases for Research, the Case of Drone Strikes

Round 1

Reviewer 1 Report

Using a scientometric approach, the paper investigates the validity and reliability of two databases, in order to quantify the impact and consequences of drone strikes in Pakistan, Yemen and Somalia.

To confirm the empirical data, the authors analyzed the datasets using two indicators, their evolution being studied between 2002 and 2017. The objective was that, in this way, the degree of convergence of the obtained data to indicate the validity of the results.
The study was published in October 2018 as a pre-print. Comparing with 2018 version, we can observe that there are only minor differences in certain phrases. The analyzed data remained at the level of 2017.     In conclusion, the paper is well constructed, but research should and could be improved by including updated data.

Author Response

Please see attachment 

Author Response File: Author Response.pdf

Reviewer 2 Report

Decision: "mayor Review"

1) Some of the conclusions are presented without supporting results, for instance in the following lines

-Line 422: "Since the two databases are  highly correlated" How do you support this? Has been computed the actual correlation between the 2 data sets ?

-Line 436: the databases are likely to become even more central to any future research on drone strikes"  It looks like this is a subjective observation. How do you reach this conclusion, based on the results?

-Line 443:"there are now ample opportunities for social scientists to improve the granularity and quality of their research" What do you mean "improve the granularity" ? Granularity is fine or coarse?

-Line  447: Starting in this line, the second half of the paragraph discussed aspects that are not part of the  paper results

-Line 396: Furthermore, the ‘strike by strike’ matrix shows a strong accordance between the datasets. This claim is unsupported, as the aforementioned matrix is not presented in the manuscript

2) I would say that the strike data for each of the analyzed cases ( Pakistan, Yemen, Somalia)  should be calles a data set  instead of a "database"

3) Methodology section: In line 312, it is stated that "the chosen format comprehends 10 columns". But only 6 column names are mentioned

 

Author Response

please see attachment 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Thanks for taking care of most of the observations.

However, there is still an observation that has not been addressed. In my first review, I had stated:

3) Methodology section: In line 312, it is stated that "the chosen format comprehends 10 columns". But only 6 column names are mentioned

Please correct this issue.

Thank you.