Review ch4

[mariaruth]

After re-reading the chapter, i agree with Fede's comment that it is odd that we have left out enumerator training and field management. Those do both have important implications for data quality and research ethics. I think the best way to address this would be to add new short subsections to the survey section. i put placeholders in but wanted to check what others thought before writing.

[mariaruth]

Most of the reviewer comments are addressed. I've dealt with all I can without further inputs. See https://paper.dropbox.com/doc/DRiP-Ch4-Revisions-Peer-Review--BDVcwXe7Xe~zPKbDqzhC11RyAQ-mPc9G8URJXXlQYr9nqdd3 for a summary of what's pending and what inputs are needed.

[bbdaniels]

@mariaruth If you want to add enumerator training and field management I would say keep those sections pretty short -- it will be hard to cover in sufficient detail for all the possible arrangements of who is managing data collection and might not come down to much more than "have a plan for this".

[mariaruth]

That’s exactly why I’m not sure if this makes sense to include. Is something that short useful? Maybe just explicitly say we aren’t going to treat this in detail and point people to other materials we have on it? But it is important for data quality for sure ...

Sent from my iPhone

On Jan 19, 2021, at 2:10 PM, Benjamin Daniels notifications@github.com wrote:

 @mariaruth If you want to add enumerator training and field management I would say keep those sections pretty short -- it will be hard to cover in sufficient detail for all the possible arrangements of who is managing data collection and might not come down to much more than "have a plan for this".

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[kbjarkefur]

EDIT: wrote a comment first that I re-wrote after finishing the whole section

Regarding questionnaire design and survey management. I think that these sentences describes what this this book will focus on. And I think it is ok for this book to focus more on electronic/digital data collection as other great resources already exists. I think we can add in this intro to the section that if primary data is not used then this section can be skipped.

https://github.com/worldbank/dime-data-handbook/blob/fe09efce174ab846781c3c735302c3c002c8d0b4/chapters/4-acquisition.tex#L251-L254

Questionnaire design

On the other hand, I think we are covering questionnaire design in as much detail as most other topics in this book. We talk about resources like LSMS where one can start. Then we discuss the difference between questionnaire design and questionnaire programming. Then we talk about including outcome variables necessary for our research design and theory of change. And we have a paragraph about content focused piloting and a link to an checklist. It is not the topic we dive the deepest into, but I think we cover it well since we also point to a very well developed wiki page.

Field supervision/survey management

We do not cover field supervision and survey management. Does that include something else than enumerator training? If not, then I think we can have a short section on this where you included a title where we list the most important things to remember but then point to this exhaustive wiki page for details: https://dimewiki.worldbank.org/wiki/Enumerator_Training I think that would be a great way to combine the advantages of both the book and the wiki.

Data quality checks specific to primary data collection

I'd say that both back checks and HFCs would fit here, but on the call this day we said that HFCs was a part of data cleaning. I do not think that back checks is a big enough topic to fill a full sub-section. If we do not want to move HFC then I do not think we should have a section on data quality here in Ch4, and then fit back checks somewhere next to HFCs anyways.

[luizaandrade]

Field supervision/survey management

I think this is related not only to enumerator trainings, but also to quality checks. It includes, for example, having a dedicated quality control person in the team, having the supervisor follow interviews and check-in with enumerators that are having issues, communicating issues found on back checks and HFCs to field teams and deciding how to follow up. This is what I could come up with, but I think that a couple of paragraphs indicating that these are things to think about and pointing to resources should be enough.

Data quality checks specific to primary data collection

The reason why we added quality checks in chapter 5 is that tidy data is needed for it. I think in this chapter we can say what back checks and HFCs are, that when collecting primary data the team should also do enumerator checks and stress the need to communicate issues with the team. Then everything that is not survey-specific can go on Ch 5.

[luizaandrade]

I think these lines should be moved to chapter 4:

• https://github.com/worldbank/dime-data-handbook/blob/ba0105d6a9a3f779abbb7026e723db8bdecaf792/chapters/5-processing.tex#L329-L341
• https://github.com/worldbank/dime-data-handbook/blob/ba0105d6a9a3f779abbb7026e723db8bdecaf792/chapters/5-processing.tex#L408-L460

[mariaruth]

@kbjarkefur all your requested changes have now been addressed

[mariaruth]

@bbdaniels your comments are all resolved, thanks for the proofread. @kbjarkefur this is ready to merge but there are conflicts...