2018 Howard County missing precinct level data

[crass]

The file 2018/20181106__tx__general__howard__precinct.csv does not actually contain the results for each precinct, the precincts are combined.

[alanhuang122]

This is, unfortunately, a deficiency in the data that was provided to us by the Secretary of State. Both the Democratic and Republican primary results spreadsheets contain only six precinct groupings. As a side note, the Texas Legislative Council offers statewide precinct results, but their data does not break votes down by early voting / election day / absentee, nor do I think their data includes uncontested races.

[crass]

So why don't we get the Texas Legislative Council's data? I think its better and more expected to have true precinct level numbers without early voting / election day / absentee breakdowns, than to have that but not actually being precinct level data. What about contacting the county directly for the precinct level data?

[dwillis]

@crass it's a fair question, and you're welcome to use the TLC data for your needs. the reason we try to obtain and convert the results from the counties is because we place a premium on having the vote by mode breakdowns and, as @alanhuang122 says, the TLC data is missing some races. We have contacted the counties directly - Texas takes months to do and I can show you records of emails, faxes and phone calls we've made, and hundreds of dollars we've spent to acquire these records. Not every county is responsive to our requests.

[crass]

@dwillis Thanks, for the explanation. I hope my question didn't come off as accusatory. I'm fairly new to the project and not familiar with what data is a premium, hence the question. I haven't seen that information anywhere, perhaps it should be added to the CONTRIBUTING.md file. (Also, perusing TLC's site, I don't see any election data, maybe its on request?).

I believe that the data collection is a costly and time consuming aspect of this projects. I'm not blaming anyone for not doing a good enough job. In fact, I admire the work you guys have been doing. I also think that the results files are misleading when records contain cumulative precincts. Perhaps the precinct should be empty and a combined_precincts field used instead (and the field be formatted for easy and consistent parsing, eg. use | to delimit multiple precincts). Perhaps this should be a new issue if you agree with making a change. Regardless, I think this issue should stay open until we get the precinct level data.

[dwillis]

@crass No worries; we're glad you're here and are interested in the project. The TLC recently revamped its site, and in doing so changed the way they present data, but here is the section with precinct data. The biggest change is that results seem to be broken out by contest.

Howard isn't necessarily alone in combining precincts, although it's usually more common in primaries, where many fewer voters results in that being done in counties with smaller populations. Howard did the same in 2016, too. So our choices here aren't great, but our goal is to turn the records that we get into data, and that's what we try to do, even if the records we get are not perfect. I've asked Howard if they have results that correspond to each individual precinct and will update when/if I hear back.