Add CPS with h_seq and ffpos Variables

[andersonfrailey]

This PR corresponds with TaxData PR #123. I've added the new variables to records_variables.json and replaced the CPS file.

cc @MattHJensen

[andersonfrailey]

It looks like the appveyor tests failed because they never even ran. This is the message:

Build started
if "%ARCH%" == "64" set MINICONDA=C:\Miniconda36-x64
if "%ARCH%" == "32" set MINICONDA=C:\Miniconda36
set PATH=%MINICONDA%;%MINICONDA%/Scripts;%MINICONDA%/Library/bin;%PATH%
git clone -q https://github.com/open-source-economics/Tax-Calculator.git C:\projects\tax-calculator
fatal: early EOF
fatal: The remote end hung up unexpectedly
fatal: index-pack failed
error: RPC failed; curl 18 transfer closed with outstanding read data remaining
Command exited with code 128

Could this have to do with the new version release?

[codecov-io]

Codecov Report

Merging #1635 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #1635   +/-   ##
======================================
  Coverage     100%    100%           
======================================
  Files          37      37           
  Lines        2857    2857           
======================================
  Hits         2857    2857

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 431af2e...602b10c. Read the comment docs.

[martinholmer]

@andersonfrailey said:

It looks like the appveyor tests failed because they never even ran.

But the other tests ran and passed. Probably a problem with appveyor.
I just now reran the tests to see if we have better luck with appveyor.

After a few minutes: We see that all the tests passed. Now that you're starting to do a lot of Tax-Calculator development work, you are getting a sense of how flaky the GitHub testing environment can be.

[andersonfrailey]

Thanks, @martinholmer. Looks like everything worked this time.

[MattHJensen]

Can anyone think of common mistakes or sources of false confidence that users are likely to encounter when they start using these variables to tab taxdata/tax-calculator variables against variables from the raw CPS (including demographic variables)? I’m wondering if we need some documentation about proper use of these variables before we merge them.

cc @andersonfrailey @martinholmer @Amy-Xu @hdoupe @GoFroggyRun @codykallen

[martinholmer]

@andersonfrailey, Is pull request #1635 ready to be merged? Or are you still working on it?

[MattHJensen]

Can anyone think of common mistakes or sources of false confidence that users are likely to encounter when they start using these variables to tab taxdata/tax-calculator variables against variables from the raw CPS (including demographic variables)? I’m wondering if we need some documentation about proper use of these variables before we merge them.

@martinholmer, I think this issue still needs to be addressed ^^

It seems like we should at least mention that we don't have targets to maintain demographic accuracy in the reweighing procedure. Shouldn't that be a problem if someone wants to tab by race and say, for example, that the average tax cut for whites is X and for non-whites is Y, @Amy-Xu @andersonfrailey?

When we add variables like this to the CPS file, I think we should add discussion to the tax-calculator documentation since we are shipping the cps file with taxcalc, or at least link to documentation elsewhere. Talk about giving people rope to hang themselves...

[martinholmer]

@MattHJensen said in #1635:

It seems like we should at least mention that we don't have targets to maintain demographic accuracy in the reweighing procedure. Shouldn't that be a problem if someone wants to tab and say, on average the tax cut for whites is X and for non-whites is Y, @Amy-Xu @andersonfrailey?

When we add variables like this to the CPS file, I think we should add discussion to the tax-calculator documentation since we are shipping the cps file with taxcalc, or at least link to documentation elsewhere. Talk about giving people rope to hang themselves...

I haven't been involved in generating the cps.csv file, so I don't have any idea what you're referring to here.

I'd be happy to include in the Tax-Calculator user documentation language that somebody else drafts, somebody that knows more than I do about how the cps.csv file was generated.

[andersonfrailey]

@martinholmer this PR is ready to go.

I agree with @MattHJensen that we should note that maintain demographic accuracy. Would the index file that @martinholmer edited in PR #1648 be the best option for these notes or do y'all have another file in mind?

[MattHJensen]

index.html is what I was thinking. I'll merge this now and then that can follow in a separate PR.

[MattHJensen]

Thanks for the contribution @andersonfrailey!

[martinholmer]

@andersonfrailey and @MattHJensen,
Never edit the index.html file because it is generated by the make.py script using the skeleton index.htmx file as input. In order to change text in the user documentation you have to edit the index.htmx file.

[MaxGhenis]

Was documentation ever added for these fields? I don't see it in the core user documentation.

[martinholmer]

@MaxGhenis asked in PR #1635:

Was documentation ever added for these fields?

Yes.

I don't see it in the core user documentation.

Try your browser's search/find capability.

[MaxGhenis]

The only documentation for h_seq and ffpos is their standard field docs:

Input Variable Name: h_seq
Description: CPS household sequence number (not used in tax-calculation logic)
Datatype: int
Availability: taxdata_cps
IRS Form Location:
2013-2016: sample construction info

Input Variable Name: ffpos
Description: CPS family identifier within household (not used in tax-calculation logic)
Datatype: int
Availability: taxdata_cps
IRS Form Location:
2013-2016: sample construction info

Is there documentation along the lines @MattHJensen suggested, i.e. how to best use these fields and potential pitfalls with weighting and targets?