A guide to help resolve between Voteview data and Congress data

[adamboche]

There are a couple of differences in how we report votes compared to the Congress website, and for older votes it can take some work to match them up. They often email us having trouble with this. We should have a guide with some points to consider.

To kick it off:

• Our "rollnumber" field on the website excludes vacated votes and quorum calls (in order to maintain continuity with earlier versions of Voteview, which began this practice long ago). Whenever it is present, our "clerk_rollnumber" field should match the Congress website's number.

• The dates should always agree between sources.

• Voteview uses CQ Roll Call's assessment of presidential positions to estimate ideal points for presidents. The president may be counted with a "pseudo-vote" as if he had been a member of the Senate and voted on a motion.

• There are several different ways to cast a vote -- Yea (1), Paired Yea (2), Announced Yea (3), Announced Nay (4), Paired Nay (5), Nay (6), Did not vote (9), -- and different sources may include or exclude paired votes. Our raw "member vote" files have the specific vote codes 1-9 when you need to see the specific type.

[JeffreyBLewis]

This is a good idea. One source difference we should just rid of. It is any easy fix to simply count the 1s and 6s in the Y and N totals which (once the president pseudo votes are removed) should give the totals reported in other sources. We will still need to describe paired votes, but it should help folks search for specific roll calls and reduce the number of emails. Is it right the .dtl files for the olde votes provide no info about clerks number? If not, maybe I should put some undergrads on appending those (also invites the question as to whether they even exist back to the dawn of Congress).

…

On Fri, Jan 22, 2021 at 7:30 AM Adam Boche ***@***.***> wrote: There are a couple of differences in how we report votes compared to the Congress website, and for older votes it can take some work to match them up. They often email us having trouble with this. We should have a guide with some points to consider. To kick it off: - Our "rollnumber" field on the website excludes vacated votes and quorum calls (in order to maintain continuity with earlier versions of Voteview, which began this practice long ago). Whenever it is present, our "clerk_rollnumber" field should match the Congress website's number. - The dates should always agree between sources. - Voteview uses CQ Roll Call's assessment of presidential positions to estimate ideal points for presidents. The president may be counted with a "pseudo-vote" as if he had been a member of the Senate and voted on a motion. - There are several different ways to cast a vote -- Yea (1), Paired Yea (2), Announced Yea (3), Announced Nay (4), Paired Nay (5), Nay (6), Did not vote (9), -- and different sources may include or exclude paired votes. Our raw "member vote" files have the specific vote codes 1-9 when you need to see the specific type. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#15>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABEKOZL4DYLKF7UJOQU6QC3S3GKVVANCNFSM4WOUSXAQ> .

-- Jeffrey B. Lewis Chair of the Faculty, College of Letters and Science Professor, Department of Political Science University of California, Los Angeles BOX 951472, 4289A Bunche Hall Los Angeles, CA 90095-1472

[adamboche]

There is a sequential field in the DTLs but I don't remember what it represents. They can be viewed with db.voteview_rollcalls.find({congress:106}, {clerk_rollnumber:1, dtl_str:1}).pretty() On Fri, Jan 22, 2021 at 9:41 AM Jeffrey Lewis <notifications@github.com> wrote:

…

This is a good idea. One source difference we should just rid of. It is any easy fix to simply count the 1s and 6s in the Y and N totals which (once the president pseudo votes are removed) should give the totals reported in other sources. We will still need to describe paired votes, but it should help folks search for specific roll calls and reduce the number of emails. Is it right the .dtl files for the olde votes provide no info about clerks number? If not, maybe I should put some undergrads on appending those (also invites the question as to whether they even exist back to the dawn of Congress). On Fri, Jan 22, 2021 at 7:30 AM Adam Boche ***@***.***> wrote: > There are a couple of differences in how we report votes compared to the > Congress website, and for older votes it can take some work to match them > up. They often email us having trouble with this. We should have a guide > with some points to consider. > > To kick it off: > > - > > Our "rollnumber" field on the website excludes vacated votes and > quorum calls (in order to maintain continuity with earlier versions of > Voteview, which began this practice long ago). Whenever it is present, our > "clerk_rollnumber" field should match the Congress website's number. > - > > The dates should always agree between sources. > - > > Voteview uses CQ Roll Call's assessment of presidential positions to > estimate ideal points for presidents. The president may be counted with a > "pseudo-vote" as if he had been a member of the Senate and voted on a > motion. > - > > There are several different ways to cast a vote -- Yea (1), Paired Yea > (2), Announced Yea (3), Announced Nay (4), Paired Nay (5), Nay (6), Did not > vote (9), -- and different sources may include or exclude paired votes. Our > raw "member vote" files have the specific vote codes 1-9 when you need to > see the specific type. > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#15>, or unsubscribe > < https://github.com/notifications/unsubscribe-auth/ABEKOZL4DYLKF7UJOQU6QC3S3GKVVANCNFSM4WOUSXAQ > > . > -- Jeffrey B. Lewis Chair of the Faculty, College of Letters and Science Professor, Department of Political Science University of California, Los Angeles BOX 951472, 4289A Bunche Hall Los Angeles, CA 90095-1472 — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#15 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEXJBC3Z7W4CXQPN76OLET3S3G2DBANCNFSM4WOUSXAQ> .