The Lincoln Plawg - the blog with footnotes

Politics and law from a British perspective (hence Politics LAW BloG): ''People who like this sort of thing...'' as the Great Man said

This page is powered by Blogger. Isn't yours?
Sunday, April 03, 2005

Powell Amendment and the Civil Rights Bill 1956: summary of voting

[Following up my piece of March 30.]

The maths dictates that, if you only consider Nay and Yea votes [1], the number of arrangements [2] to be considered in the outcomes of n roll call votes (RCVs) is 2n. Here, I'm considering six RCVs - 64 arrangements [3].

Three votes (the first in time) were on the School Construction Bill HR 7535:
  • 122: Powell Amendment (PA), to condition school aid on compliance with Brown.

  • 123: Motion to recommit, with instructions to substitute the administration program.

  • 124: Vote on passage.

The second three were on the Civil Rights Bill HR 627:
  • 134: Motion to table a motion to dispense with routine proceedings.

  • 135: Motion to recommit to provide means of further securing civil rights.

  • 136: Vote on passage.

Given that RCs 122-124 took place on July 5 1956, and RCs 134-136 only a few days later, you'd expect some consistency in voting: in fact, there's a great deal of apparent inconsistency.

The polisci profs seek to identify in 122-124 an example of sophisticated voting: closet opponents of desegregation or opponents of Federal school aid (Republicans mostly in each category) who supposedly voted for the PA knowing any bill including it would not be passed.

My angle is the surprising shift in voting patterns between 122-124 and 134-136.

The result of crunching all 64 arrangements (spreadsheets) identifies two clear patterns:
  1. the Confederacy, as expected, voted (more or less solidly - 99 votes) against everything with desegregation in it; and

  2. a block of 265 members (D108:R157) solidly supported HR 627 and opposed attempts to delay it: an NNY vote on 134-136.

That accounts for the bulk of House members - 364 out of 435.

Of the NNY voters, a liberal block also voted a straight liberal ticket on HR 7535: 87 members (D70:R17) voted YNY on 122-124.

That leaves 181 NNY voters to account for. Of these, 59 (all GOP) adopted an apparently paradoxical stance - voting YYN on the school bill: in favour of Powell, and recommittal but against the bill. The PA vote was a tad cynical, given that this group could have swung the vote on passage the other way!

A further group of 39 (D32:R7) voted NNY on the school bill: a symbolic PA vote (7 Dems were from border states, the rest of the group from non-slave sections) followed by a reluctant grab for the pork, perhaps?

A group of 37 (D1:R36) voted YYY on the school bill: a real chicane: going with Powell; getting cold feet; then finally scrambling on board.

A (penultimate, for my purposes) group of 18 (D3:R15) went NYY on HR 7535, four coming from slave sections. Again, a hard group to fathom on the school bill.

Finally, 15 (all GOP) went YNN on the school bill: voting for Powell, and then against the bill makes little obvious sense.

(The way the numbers fit together is rather more easily fathomed with the spreadsheet in front of you!)

It's one thing to identify the votes and aggregate the variations: the real problem comes, as I've mentioned before, at trying to explain the motivations of particular combinations or individuals.

For example, to take a still-famous name, a Google search on "civil rights bill" 1956 "robert byrd" produces 45 of 73 items, none of which seem germane.

Not quite time to throw in the towel, though - if only on the Micawber Principle.

  1. Voteview considers pairs to have actually voted. Though it produces a summary of actual votes cast (net of paired votes), you cannot tell whether any particular member has actually voted in a roll call or merely been paired. That fits with the intention of Voteview's makers (measuring preference, no providing a historical record); and since theirs is the only game in town...

  2. I'm avoiding permutations and combinations which have special meanings not relevant here!

  3. A mere 25 RCVs generate 34 million! The number of rows in an Excel spreadsheet is 216, I believe.


First rule of voting analysis (I've found!): tabulate all votes first. A lot of time is wasted if you try to pick what you think are the most interesting combinations - and then find that others are also interesting!

Second rule: take your votes in strict date order from the start. Swopping later and earlier votes once your spreadsheets are finished may seem easy in theory (isn't that what spreadsheets are supposed to do?), but soon drops you in the quicksand. (You being me, in this instance.)

However, we're now back to apple-pie order - fingers crossed!

free website counter Weblog Commenting and Trackback by