IJPDS International Journal of Population Data Science

https://doi.org/10.23889/ijpds.v7i4.1764.review.r1.reviewa

Thank you for giving me the opportunity to review this manuscript, which describes an analytical plan to link US census data with other administrative registers. The current infrastructure includes censuses from 1940, 2000 and 2010, and the manuscript presents a plan to link censuses from 1960 to 1990. The manuscript is well written, and the plan is described in detail. I have some comments that might help the authors improve the manuscript.

1. What is the most important contribution of this manuscript to the scientific literature? I think a paper describing the process after the process is done could be more relevant. Then, specific estimates about linkage rates, etc. could be included (and not only expectations as they are now). For example, specific subgroups for which linkage might be more difficult are not described, but a proper analysis could be performed after the linkage has taken place.
2. I think a final paragraph describing the potential of such dataset could be useful. What type of research questions could this data answer? Or how could it be used?
3. In page 2, in the middle of 2^nd paragraph, there is a sentence starting with “Our approach”. Please check the sentence, as I think it is not correct.

https://doi.org/10.23889/ijpds.v7i4.1764.review.r1.reviewb

This manuscript describes an approach to create a link among routinely-collected data through the US Census to happen some time in the future (no timeline is provided). No "non-routine" database is actually considered in this manuscript.

It is not clear to me whether this is a protocol paper for project, or a description of an aspirational project. There is no research question or a goal for achievement of a degree to linkage that would show that this method is indeed better than those used before and overcomes the challenges described very briefly in the introduction. Even title is misleading, as the paper does not describe how the administrative records *are* used but how they *may* be used in the future. The paper structure is somewhat confusing as elements are somewhat arbitrarily placed in sections - especially baffling is inclusin of Results, considering that this appears to be a future project so there are no results as yet?

The authors refer to success linking, reaching “even greater proportion” (p9), but do not specify what are their expectations in terms of what these proportions will be, and how will they judge that their method is a success. How low is it currently, and what this new method will bring, and will it be cost-effective in terms of labour involved and data potential? It would also be helpful to explain whether the Census is (and has been) mandatory in the US, and what is the % of population that completes it; and who might be missing.

The first paragraph of Results really belongs in the Introduction as it sets the study goal – with judicious use of the existing references (as authors say “work that we and others have done” but there are no accompanying references). And many following paragraphs belong to Methods – see below.

Introduction, 1^st paragraph – is this issue (low rates of coverages due to changes in names etc.) specifically a challenge in the US Census data, or is it a pervasive issue in other countries as well? The Introduction could be expanded to fuller illustrate the magnitude of the problem or how it was dealt with elsewhere.

P2 “Our approach will be to creating linkages between censuses will be to...” please revise

“Our work builds on methods…” should this be supported by a reference if those methods have already been established?

P3 Last introduction paragraph – was there no research or at least a validation question?

P4 first paragraph – I am a linked datasets’ user rather than creator, but the method described here seems to me a very standard one not necessarily novel?

Second paragraph – could the authors describe briefly the records they mention (such as SSA Master Beneficiary Record etc.) for international audiences? Are SSA essentially tax data? I can see that they are described later – maybe add section titles so that this information could be easily found

“Though the SSN links are expected to be high quality…” why?

P7 “The resulting file…” if the file will have one row per person, how will family members be cross-referenced?

P8 …” even though it will only be available for some cases” – should this protocol not include a detailed record of matches using each type of variables, especially for those that will only be available for some cases – should they not be used to evaluate the effectiveness of the procedure?

“we would have success linking…” Please define criteria for “success” of the linkage using this method – is there a percentage that you are striving for?

P8 – does it follow that the success rates will be better for those US citizens who completed the long form census?

P9 “…used to produce research in top-tier journals across the social sciences” could you specify what kind of research exactly? Highlighting a few specific topics would certainly increase the reader’s engagement.

“For a large majority of the population…” such as? Do you have any evidence what particularly makes this “large majority” different and why it is so warranted to seek this additional “even greater proportion” match?

P9/10 How many such additional passes the authors expect to make? Why would it be efficient?

P11/12 Pilot studies are described in Results, even though they appear to have been already published – if that is the case then they should have been described in the Introduction not Results. As far as I can gather the activities are to take place in the future, so there really aren’t any results per se?

P12 – I finally found the information on people who are likely to be missed using the standard methods… (2^nd paragraph) again this is something that should be up front as a rationale for the work. Will they not be also missed in this more detailed approach though?

P13 2nd paragraph – also finally there is some information on linkage quality – I would have expected that in the Methods – and at least mention at the end of Introduction. These paragraphs describe methodology yet are placed in results…

P16 “when the digitization is complete…” so there really is no concrete timeline for this work?

https://doi.org/10.23889/ijpds.v7i4.1764.review.r1.dec

Dear Joseph Alexander, Katie Genadek:

We have reached a decision regarding your submission to International Journal of Population Data Science, "Using Administrative Records to Support the Linkage of Census Data".

Please address the attached reviewers' comments and return to us: one clean and one tracked changes version of your revised manuscript, plus a point by point letter of response/rebuttal, by 31st August 2022.

Our decision is to:

Resubmit for Review

As well as addressing the reviewers’ comments, please explain what this work adds to the knowledge base e.g. is the planned strategy informative to others? It needs to be badged and structured it more clearly as a protocol paper, with anticipated challenges and benefits, etc. Also, is there an indicative timeline? The results given as are of earlier studies – these need to be moved as per reviewer’s comments. Are any of the datasets non-routine? If not, although the manuscript would not fit in the special issue, it could go in a standard issue, providing it’s accepted. I hope this is helpful.

Kind Regards

Kerina

https://doi.org/10.23889/ijpds.v7i4.1764.review.r2.reviewb

I do appreciate the authors taking a lot of effort to address many of my detailed comments. I have to admit that it would have been easier to track the changes if, in their document on responses to review, the authors included the new page numbers where these revisions were made. However, they were mostly ssatisfactory.

I have just a few comments below.

1. I think that the main issue mentioned by the Editor has not been addressed. Even protocol papers need to include literature review background that situates the new project in the context of the existing state of knowledge and justify the new study. The existing background section really only describes this particular census database. I therefore believe that Introduction needs to be somewhat revised to make the case that this project and resulting research will bring new knowledge.

2. With the exception of 2 sentences on page 3 there is no information about what kind of studies could be conducted using this database. I will leave it to the Editor to decide whether it is sufficient.

3. I do not believe that the last paragraph on page 4 is required.

4. My other minor point is regarding my request for more information about the Census – I was asking on behalf of all readers unfamiliar with it, so this information, as well as relevant references, needs to be included *in the paper,* not only in the response to review. (my comment and response copied below)

It would also be helpful to explain whether the Census is (and has been) mandatory in the US,and what is the % of population that completes it; and who might be missing. The U.S. Census is a mandatory survey with estimated coverage rates ranging from 93.5% in 1880 to 98.6% in 1980. See Miriam King and Diana Magnusson, “Perspectives on Historical U.S. Census Undercounts”, Social Science History 1995 (19:4) 455-466.

5. P17 – please explain (in the paper) what the “ground truth” means.

6. Finally, there are many acronyms that are not intuitive (at least for this reviewer). It might be useful to have a list of them as the first table.

https://doi.org/10.23889/ijpds.v7i4.1764.review.r2.dec

Dear Joseph Alexander, Katie Genadek,

We have reached a decision regarding your submission to International Journal of Population Data Science, "Using Administrative Records to Support the Linkage of Census Data".

Our decision is: Revisions Required

Please address the attached reviewers' comments and return to us: one clean and one tracked changes version of your revised manuscript, plus a point by point letter of response/rebuttal, by [editor to insert date here].

Kind Regards

https://doi.org/10.23889/ijpds.v7i4.1764.review.r3.dec

Dear Joseph Alexander, Katie Genadek:

We have reached a decision regarding your submission to International Journal of Population Data Science, "Using Administrative Records to Support the Linkage of Census Data", and are delighted to inform you that our decision is to: Accept Submission.

We look forward to working with you through the next stages towards final publication.

Please get in touch if you have any queries going forward. Thank you.

Kind Regards