OK, I need to get two things onto the table right at the beginning of this post:

1. Yes, this is an extremely geekish and polity wonkish post, but that’s what interests me and this analysis is the one I have really wanted to do since the 219th General Assembly adjourned last July.  I do think there is something important about the PC(USA) in here so if you want to skip the data analysis and jump to the end you will find my discussion there.
   
2. I posted a preliminary result on Twitter on Saturday but got the variables confused.  Sorry about that. I posted a correction on Twitter and will point out the error when I come to it in this post.

So, the question that has had me on the edge of my seat is the degree to which each of the three high-profile amendments is correlated with the other two.  I took an initial pass at this question a couple of weeks back and found a strong correlation.  That correlation has weakened a bit but is still present, stronger in some relationships than others.  While it still may be a bit premature to make strong conclusions from the data at this point in time, I think I’ve got enough data to do a preliminary analysis.

Now, if you are looking for just the vote results after last Saturday here is the “word on the street.” Belhar is still not getting the 2/3 it needs with 32 yes and 28 no.  The New Form of Government continues to have weak support and still trails, currently at 25 to 31.  The story of the last week is that support for Amendment 10-A continues at the pace we have seen throughout the month and with three more presbyteries switching their votes a total of 12 presbyteries have shifted to “yes” with only one shifting to “no.”  At this point enough presbyteries have shifted (a net of nine was needed) that with all the rest of the presbyteries voting as they did in the last round Amendment 10-A will be approved. At the end of the weekend the vote stood at 55 to 41.  No further analysis of that today, I’ll come back to that in another week or two. (Particularly in light of the question about the vote totals that is raised at the end of the next paragraph.)

First, the usual details regarding data:  For my data I have aggregated numbers from Twitter as well as vote counts at the Covenant Network, Yes on 10-A, Reclaim Biblical Teaching and the Layman. This aggregation is available in my spreadsheet.  I have also updated my cross-vote spreadsheet through Saturday’s reports.  The analysis below is more sensitive to the exact vote count and where the tally sheets sometimes differ a bit I have used either a majority among them, the Twitter reports, or a consistency in total votes to select a preferred number.  This is also probably a good place to add that the voting is not finished yet and this analysis is only preliminary based on the current data. And in a very interesting development today as I was finishing this up, the official vote tally from the Office of the General Assembly was posted.  It has caught the attention of several of us because it has numbers significantly different than the unofficial sites — nFOG 38 to 25, Belhar 38 to 18, and 10-A 47 to 33.  The difference is presumably due to reports by presbytery stated clerks not reflected in the unofficial counts.  Hopefully with time the two sets of lists will converge.

So, let’s take the three comparisons from strongest to weakest (and if you want to see the graphs in more detail they are larger in their original form and you can open them individually):

Belhar to nFOG
The strongest relationship between the issues is between the votes on the Belhar Confession and the New Form of Government. (This is the one I should have pointed out in the tweet on Saturday.)  So far 33 presbyteries have voted on both of these issues, and 27 of those have recorded vote numbers on both votes.  Looking at the numbers you can see the strength in both the cross-tabulation and the linear regression:

n=33	Belhar yes	Belhar no
nFOG yes	10 30%	2 7%
nFOG no	1 3%	20 60%

Bottom line: The strength of a presbytery’s vote on nFOG is going to be very close to the strength of a presbytery’s vote on Belhar.  The fit of the linear line is good with an R² = 0.73  (a number very much like correlation that I talked about in a previous post with 1.0 as a good and 0.0 as not correlated, but this number is always positive), and a slope pretty close to 1 (the two vote percentages increase in the same proportion).  This is seen in the yes/no comparison where 30 presbyteries have voted the same way on both issues and only 3 (10%) have voted opposite on them.

Belhar to Amendment 10-A
The next strongest relationship between the issues is that between the votes on the Belhar Confession and Amendment 10-A.  (This is the one I incorrectly pointed to in the tweet.) So far 35 presbyteries have voted on both of these issues, and 25 of those have recorded vote numbers on both votes.  Here is what the numbers look like:

n=35	Belhar yes	Belhar no
10-A yes	17 49%	3 9%
10-A no	3 9%	12 34%

Bottom line: The strength of a presbytery’s vote on Amendment 10-A is going to be related to the strength of a presbytery’s vote on Belhar, but not as strongly as for the last case and not in 1:1 proportion.  In this case, the fit of the linear line is not as good, but still moderate, with an R² = 0.62 and a slope 0.51. There is also a significant upward shift in the trend line of almost 20%.  What this means is that for presbyteries not strongly in favor of Belhar, on average there is a 20% “base” in favor of Amendment 10-A.  On the other end, a presbytery strongly in favor of Belhar has, on average, a 30% “base” opposed to Amendment 10-A.  The yes/no comparison also shows that the linkage is not as strong and direct where 29 presbyteries have voted the same way on both issues and six (18%) have voted opposite on them.  From these results, the association of these two issues is only partial and the attitudes on one are not driving the other as strongly as might be suspected.

nFOG to Amendment 10-A
The weakest relationship is between the votes on the nFOG and Amendment 10-A. So far 36 presbyteries have voted on both of these issues, and 23 of those have recorded vote numbers on both votes.  Here is what the numbers look like:

n=36	nFOG yes	nFOG no
10-A yes	12 33%	5 14%
10-A no	5 14%	14 42%

Bottom line: There is a weak, positive relationship between a presbytery’s voting strength on nFOG and the vote strength on 10-A.  However, as can be seen in the scatter of the data on the graph, especially at the higher end the relationship is weak.  The scatter in the data is evident with R² = 0.39 and the lower slope of 0.46 also suggestive of a weaker linkage. The yes/no comparison supports  that the association is not as strong and direct with almost 1/3 of the presbyteries voting opposite ways on the two issues.

Discussion and Conclusions
I must admit that the strength of the Belhar/nFOG association was a bit of a surprise to me.  With the on-going discussion of the synergy between Belhar and 10-A I was expecting to that to have the strongest correlation. And the very nearly 1:1 association means that they two issues probably elicit the same response from any given commissioner.  One thought that occurred to me is the similar nature of these two issues in regards to their impact on PC(USA) polity.  While the impact of each is still being debated and is, to a certain degree, unknown, if approved they each would leave a significant mark on the constitutional documents.  There could also be a less tangible factor in the willingness to preserve the status quo — since these two amendments have similar impacts on the established order of things it is reasonable to presume that if a commissioner had a particular comfort level with changing one of them, they would have a similar comfort level changing the other. But whether it is related to those explanations, or other factors, the data appear to show that even if presbytery commissioners don’t necessarily explicitly link them, they still seem to think about them in the same way.

Having said that, and recognizing the vote tally differences from today’s announcement, I need to point out that it appears point twice as many presbyteries have voted against both of them as have approved them.  This raises a couple of questions when we look at the voting trends for the issues by themselves since the votes overall are more even.  The first thing is that as the double-issue voting catches up the close agreement could go away.  But if the close agreement continues, and considering that one currently has a majority and the other does not, we might expect the vote margins to narrow.  We also open up the possibility that Belhar might not even receive a majority vote if nFOG continues to not receive a majority.  The opposite could also be true – that nFOG will be pulled up by future positive voting on Belhar.

We could also ask the question about the strength of Belhar from the 10-A relationship.  Doing a back of the envelope calculation and extrapolating out the 10-A voting based on current proportions a 99 to 75 final vote (56.6% yes vote) would be a reasonable conclusion.  If we then mix apples and oranges and ignore whichis the the dependent and which the independent variables, plugging 56.6% yes vote on 10-A into the regression formula gives a 73% yes vote on Belhar.  Fun to speculate but I just violated too many mathematical and data rules to really believe that.  A more valid approach would be to take the presbytery yes/no vote cross-tabulation as a guide where we see that at the present time the opposite voting categories would off-set each other.  This would suggest that for presbyteries (apples to apples) the Belhar final vote could would be very close to the 10-A final count, in which case 56.6% won’t get it approved.

I’m not sure there is much to say about the weak correlation between nFOG and 10-A.  This is more of what I was initially expecting since the two issues do not have a lot in common polity-wise.  The weak linkage seen could be some polity point I am overlooking or a desire to preserve the status quo.  Either way, there is not enough strength in that correlation to risk making any conclusions about one from the outcome of the other.

So that is what I see at this point.  I will point out again that this is truly preliminary since at this time for each pairing only around 1/5 of the presbyteries have voted on both amendments.  I look forward to seeing how this progresses as the voting continues filling in the missing data.  Stay tuned…