Talk:Results of the 2008 Democratic Party presidential primaries/Archive 10

Proposals to Change Presentation of Results
Now that the vote has concluded in favor of resuming event-based rows, we can talk about how best to present the results of those events. Please create subsections below.Northwesterner1 (talk) 00:45, 20 March 2008 (UTC)


 * Thanks for doing this, Northwesterner. I was fiddling around with drafting a section much like this. Great minds, I guess. --Bryan H Bell (talk) 02:04, 20 March 2008 (UTC)

Proposal: Percentages for popular vote only
(Cut and pasted from the above discussion)

I don't think percentages should be included unless they reflect a popular vote. In cases where just delegate totals are released, having the percentages makes the table more confusing. ~ Paul T +/C 21:46, 17 March 2008 (UTC)

That is, the above example should be changed to this:


 * Support Paul's fix. Northwesterner1 (talk) 22:07, 17 March 2008 (UTC)


 * I agree as well, though my preference would be to work out the details of whichever option we select after the vote has concluded. I'd like to ensure that discussion here doesn't cause this vote to stray from the core issue we're trying to decide because it's been so difficult bringing this issue to a conclusion. --Bryan H Bell (talk) 22:09, 17 March 2008 (UTC)


 * Support Jon (talk) 13:18, 20 March 2008 (UTC)


 * Support Subver (talk) 14:57, 20 March 2008 (UTC)


 * Support — Zntrip 17:33, 20 March 2008 (UTC)


 * Support Gelbza (talk) 05:31, 21 March 2008 (UTC)

I implemented the other part of Paul's fix -- the brackets around estimated delegates produced by the first event in states with multiple events -- which I guess we didn't actually vote on here but which I assume is also accepted. However, I ran into some questions: Northwesterner1 (talk) 17:51, 26 March 2008 (UTC)
 * 1) How do we deal with "estimates" that have been struck through? (IA Jan 3, MI, FL) I left these struck through, with no brackets.
 * 2) Then, how do we deal with the Iowa county convention results, which are still estimates. I put brackets here. Is this the best fix?
 * 3) How do we deal with states where some but not all delegates were bound at the first event? (ID, KS, NE, WY) I put brackets around these numbers to indicate that the total is an estimate, but that may be misleading, as some of those delegates are actually bound. One potential solution is to have two rows of delegate numbers in these states, the first giving the total estimates (in brackets) and the second giving the number of delegates actually bound at that event (without brackets).
 * 4) Some caucuses appear to have all delegates bound at the first event. Is this correct? (AS, MN, NM, HI)
 * 5) Dems Abroad is its own special case, as they appear from this table to have bound some delegates and not reported estimates on the others. Thus, no brackets. Is this correct?


 * Here's my take on Northwesterner's questions:
 * I don't see a problem with bracketing estimates that have been struck through. MI and FL are primary states, so their numbers aren't estimates anyway and shouldn't be bracketed. I do understand Subver's issue (expressed here) that using strikeout could be confusing on outdated estimates (especially considering that we use strikeout for MI and FL as well). I just can't think of a better way to display it.
 * I think Northwesterner's choice of bracketing Iowa's county convention results is good. Even though they're updated numbers, they're still estimates.
 * States in which some but not all delegates are bound does present a problem. I like the idea of having two rows of delegate numbers in those states (estimates in brackets and bound numbers without brackets). It will complicate the presentation a bit, but maybe it's worth it to provide more accurate information.
 * Yes, according to their delegate selection process documents, even though AS, MN, NM, and HI are caucus states, all delegates are bound after the first event. Subsequent events only determine who will serve as delegates.
 * Dems Abroad is indeed a strange case. According to their delegate selection process document, no delegate votes are bound at the first event, 4½ are bound at the second event, and 2½ at the third event. But according to their press releases, they seem to indicate that they consider 4½ of the delegate votes to already be bound. They offer no estimates for the remaining delegates, so perhaps we should obtain estimates for the remaining delegates from another reliable source.
 * --Bryan H Bell (talk) 19:59, 27 March 2008 (UTC)

Proposal: Delete the "actual pledged delegates row"
I propose to delete this row. Reasons: Northwesterner1 (talk) 00:45, 20 March 2008 (UTC)
 * It adds to the confusion regarding event-based results, and the note that explains it "The data contained in the row entitled Actual pledged delegates is a subset of the data in the row entitled Estimated pledged delegates" may be confusing to the general reader. The fewer notes needed to explain the presentation of results, the better.
 * "Actual pledged delegates" are really just "pledged delegates estimated to a greater degree of certainty," as any pledged delegate can change her/his vote up to the convention itself. I don't think we should be in the business of decided how likely delegates are to switch their votes, so we should just stick with the estimates reported by the NYT and other reliable sources.
 * Having a single "estimated pledged delegates" row encourages standardization across this article, the main primary article, and the individual state articles.


 * We wrestled with this issue quite a bit back here. A few proposals were kicked around, but ultimately we failed to arrive at a resolution. I don't think simply deleting the "actual pledged delegates" row is necessarily the best solution here, but I need to think about this a bit before I can offer an alternative. --Bryan H Bell (talk) 02:43, 20 March 2008 (UTC)


 * Thanks for the link to the prior discussion -- very helpful. One thing I would note is that the prior discussion predated the decision to standardize the main primary article with the results article and the individual state articles. At the current point in time, I see the "estimated pledged delegates" as something equivalent to "Wikipedia's official estimated delegate count." I've added a bullet point above noting article standardization as a point in favor of the proposal.Northwesterner1 (talk) 02:52, 20 March 2008 (UTC)


 * Support Subver (talk) 14:57, 20 March 2008 (UTC)


 * Ok, here's what I would offer as an alternative to the proposal: eliminate the Total estimated pledged delegates row instead of the Total actual pledged delegates row. I would name the rows thusly:
 * Total pledged delegates + superdelegates
 * Total estimated superdelegate endorsements
 * Total pledged delegates
 * The third row would be a sum of all delegates from the events rows that have actually been bound to candidates (no estimated delegate counts). The second row would use the DemConWatch numbers as before. The first row would be a sum of the second and third rows. In addition, I would replace the "Actual pledged delegates" line in the table notes with a line that says "Estimates appear inside square brackets ( [ ] ) and are not included in the totals." My basic idea here is to make this article focus on the results of the events, rather than on estimates that give us the current status of the race. We should inlcude estimates in event rows (in square brackets) only when no delegates are bound to candidates. We should do so not to give readers a total picture of where the race currently stands, but to show a historical picture of what the results of the event signified at the time. I don't believe that the totals on this article must remain in lock-step with the totals on the main article. For Wikipedia readers looking for a current snapshot of the race, they can refer to the main article.
 * Another alternative would be to eliminate the first two rows I list above and instead place an event at the bottom of the table called "Total superdelegate votes" dated August 25. We could fill it in with DemConWatch estimates (in square brackets) until that date and then fill in the actual results afterward. --Bryan H Bell (talk) 05:05, 21 March 2008 (UTC)


 * I strongly oppose this alternative. In my view, at this point in time, the reporting of estimated pledged delegate totals is one of the most important functions of both this article and the main article.Northwesterner1 (talk) 05:14, 21 March 2008 (UTC)


 * I can understand your view. I agree that it's an important function to report the estimated delegate totals, I'm just don't feel as strongly that the function must be duplicated on both articles. I think it's enough that the main article reports this. Let me suggest a more radical change to this article that might be a compromise between your proposal and my alternatives as well a compromise on the sortable list idea discussed below.
 * What if we divided the Overview of results section in two: Overview and Details. The Overview section would contain a table with the totals rows and the Details section would contain a table with the event rows. This would permit us to easily make the Details table sortable. This would also give us the space to make the Overview table a bit more complex, perhaps showing two possibilities: an "agressive" grand total that includes estimated pledged delegates and estimated superdelegates, and a "conservative" grand total that includes only bound pledged delegates and no superdelegates. Perhaps we could even divide such a table into two tables. --Bryan H Bell (talk) 07:24, 21 March 2008 (UTC)


 * I'm open to it. I'm going to be out of town and can't really weigh in much. Whatever you all decide is cool with me. Northwesterner1 (talk) 07:30, 21 March 2008 (UTC)


 * I oppose the Bryan counterproposal. I repeat my support on original proposal. "Actual" is a misleading term, the row create a lot of confusion, the source (NYT) is not neutral and not very reliable (DCW decided to strip NYT from its comparatives, for what it's worth). Moreover the confidence of those "definitive" numbers is debated. A conservative way is to consider all numbers as estimates or unofficial until every state conventions, so this row become totally useless. --Subver (talk) 10:03, 21 March 2008 (UTC)

(undent) I agree with several of your points, Subver. The term "actual" is indeed an inexact term, a more precise term might be "bound", meaning that the allocation of delegates to candidates is binding and cannot change. How many delegates are bound to candidates and at which events this occurs differs from state to state. I also agree that we should move away from using the New York Times as a source for the total number of delegates currently bound to each candidate. We already get the official raw data from each state's Democratic Party or Secretary of State. All we have to do is total up the bound delegates from each state. Your final statement, however, that all numbers are estimates until the state conventions is not correct. In most (if not all) primary states, the delegates are bound to the candidates at the very first event (the primary). In caucus states, it is sometimes true that some of the delegates are not bound until the state convention, but this is not always (or even mostly) true. In Iowa, for example, 29 delegates are officially bound to candidates at the 3rd event (the district conventions), though the remaining 16 are bound at the state convention. The "conservative" total is valuable because it shows us numbers that are the most accurate and free of any bias (because they simply reflect the facts of how many delegates have been officially bound to each candidate), though it is true that these numbers remain somewhat (but not completely) "behind the times" until the Democratic National Convention in August. The "agressive" total shows us numbers that are sometimes wrong and reflect the unavoidable bias of the sources used for the estimates in each state, but it is also valuable because it provides the most up-to-date snapshot of the current state of the race. --Bryan H Bell (talk) 11:11, 21 March 2008 (UTC)

Proposal: Delete the language about "events won" at the top of the article
I propose to delete the sentences beginning, "There are 65 total events listed below..." Reasons: Northwesterner1 (talk) 00:45, 20 March 2008 (UTC)
 * It seems to imply that Event X should be given equal weight to Event Y. As I think we'd all agree that "winning" the initial popular event (primary or precinct-level caucus) is more meaningful than winning subsequent county conventions, we shouldn't give them equivalence here.
 * It artifically inflates the totals when a candidate wins multiple events in one state vis-a-vis states that have only one event.
 * There are so many numbers here, many editors don't take the time to update them, and these numbers can get out of date easily.
 * There are so many numbers here, it gets jumbled and is hard to understand. Not what we need at the top of the article.


 * Support. It might be useful (and notable) to say something about any states that switch from supporting one candidate in the majority of its delegates to another. I don't think this has happened yet (or is likely to happen), but it's something to keep in mind. --Bryan H Bell (talk) 02:14, 20 March 2008 (UTC)


 * Support. Athough in terms of delegate count it's whichever event(s) that finalizes the delegates that's the more meaningful. Precint-level Iowa was a virtual three way tie in terms of projected delegates. The county convention has a clear first, second, and third place finisher in terms of delegates. The other state to watch for projected delegation shifting is Nevada. Jon (talk) 13:27, 20 March 2008 (UTC)


 * Support Subver (talk) 14:57, 20 March 2008 (UTC)


 * Support — Zntrip 17:33, 20 March 2008 (UTC)

DONE. Northwesterner1 (talk) 22:56, 20 March 2008 (UTC)

Proposal: Explaining events won
We need a good, clean, clear paragraph at the top of this article, explaining how results are presented on this table on the basis of "events," as we're trying something here that is different than what is generally reported in the media and what is expected by the reader. How can we explain this most clearly? Anybody want to take a stab....?Northwesterner1 (talk) 00:45, 20 March 2008 (UTC)

Proposal: Ordering rows alphabetically by state rather than chronologically
I propose to order event rows alphabetically by state rather than chronologically. Northwesterner1 (talk) 00:54, 20 March 2008 (UTC)
 * This may clear up confusion in states with multiple events. Look at how easy Paul's example with Washington is to understand when all the Washington events are grouped together. When these results are spread throughout the table, it makes it much, much, much harder to understand what's going on in multiple-event states.
 * Now that most events are past us, a chronology of events is no longer as useful as an understanding of what's going on in individual states.
 * A chronological order of events is useful mostly for looking at upcoming contests. But this is a results article, so it shouldn't be optimized for looking at upcoming contests. That chronology can be found in the main primary article.
 * If we make this change, we could still make this a sortable table in case anyone wants to sort by date.


 * Isn’t that basically what we voted not to do? – Zntrip 02:12, 20 March 2008 (UTC)


 * The vote was to keep multiple rows for states with multiple events (rather than summarizing multiple events in one row). My proposal maintains the multiple rows we voted on. I just have a different suggestion for how they should be ordered. They would be grouped by state and ordered alphabetically.Northwesterner1 (talk) 02:30, 20 March 2008 (UTC)


 * Oppose. Those are some intriguing arguments, but I prefer that this article presents a sweeping and detailed history of the nominating process from a national perspective, rather than presenting a series of smaller, discrete, localized timelines. The latter is already presented in the state articles, albeit not on a single page. A sortable list isn't a bad idea for a compromise, though. We'd have to restructure the table a bit to make it work. It might be tricky getting the events to line up chronologically by state if an alphabetical sort is chosen. If we did this, my preference would be for the article to present the rows sorted by date as a default. --Bryan H Bell (talk) 02:28, 20 March 2008 (UTC)


 * Let's see what others think. If we go to a sortable list as a compromise, it's easiest to group alphabetically by states and then have an option to sort chronologically. But it would be possible to start chronological and have an alphabetical sort if we name the individual events something like Washington (1), Washington (2), to keep them in the right order. Then again, that's just adding more clutter to the tables. Northwesterner1 (talk) 02:44, 20 March 2008 (UTC)


 * Help:Sorting offers some interesting options (particularly "Sorting with hidden sortkey") that might help. So far I've only skimmed that article. --Bryan H Bell (talk) 02:52, 20 March 2008 (UTC)


 * I think that fixes it. If this is the preferred solution, we would just have to code the tables as "Washington 1 ," "Washington 2 ," etc. The 1 and the 2 would not be displayed in browsers. I assume we'd also have to adjust the table by giving each row a date, instead of letting, for example, "February 5" span multiple rows. Northwesterner1 (talk) 03:02, 20 March 2008 (UTC)


 * Sortable list seems best to me. Jon (talk) 13:29, 20 March 2008 (UTC)


 * Thanks for the input, but which default ordering: chronological by date or grouped by states? Northwesterner1 (talk) 13:56, 20 March 2008 (UTC)


 * It is impossible have sortable tables when "colspan" and "rowspan" are used. --Subver (talk) 14:51, 20 March 2008 (UTC)


 * We would need to take out the rowspan coding and put a separate date with each row. We would also need to separate the top part of the table (the summary) from the bottom part (the event rows) to get rid of the colspan red line. Not a big problem. Northwesterner1 (talk) 15:05, 20 March 2008 (UTC)


 * Oppose — Zntrip 17:33, 20 March 2008 (UTC)


 * Oppose Couldn't the the states with multiple events be hotlinked (wrong term probably, but where the url uses a # to link to a different section on the same page) to the later same-state contests? So the 1st Washington results would have a link leading to the second event, similar to the reference arrows? Gelbza (talk) 05:38, 21 March 2008 (UTC)

Proposal: Removing pie charts, possibly other graphics
I propose to remove the pie chart graphics from the page based on the following reasons:
 * As of this moment, March 21, the graphics are extremely outdated, showing results only through Feb 19th. Additionally, trying to keep up with modifications accurately gets seems a daunting task.
 * These graphs in particular don't really seem to have much bearing on this article. They do not reflect actual results such as delegates earned but rather a sociological survey of the population.
 * The graphs are based on exit polls (which do not have guaranteed statistical accuracy), and not actual specific polling data.

On a side note, the other graphs and charts should be rethought as well, as they don't necessarily add any new information, take up large quantities of space, and in some the wording at least is decidedly biased.

Gelbza (talk) 05:57, 21 March 2008 (UTC)


 * I made most of the graphs, but I'm not attached to any of them. Just wanted to check in, as I'm going to be out of town & off Wikipedia. Feel free to keep or delete what you like. I'm willing to update the maps with every new result, and the delegate/superdelegate bar charts every two weeks or after significant new events. I'm done updating the exit poll pie charts; it's too hard to composite the exit polls, and they're not worth it. So I'd recommend deleting those. Northwesterner1 (talk) 06:51, 21 March 2008 (UTC)

Proposal: Delete "(", ")" on Edwards delegates
I think those brackets are useless and misleading. I don't understand the meaning. If the meaning is that Edwards is out of race, they are not correct, because Edwards is out and not his delegates. --Subver (talk) 13:23, 22 March 2008 (UTC)


 * Support. I, too, see no need for the parenthases around Edwards' delegate numbers. --Bryan H Bell (talk) 20:41, 22 March 2008 (UTC)


 * Support Jon (talk) 22:48, 23 March 2008 (UTC)


 * Support. Andareed (talk) 23:59, 23 March 2008 (UTC)

How do we handle Nevada's county conventions?
I've been working my way from the top of the Overview of results table adding events for states. I'd completed Iowa, New Hampshire, and Michigan when I ran into trouble with Nevada. Nevada was the first caucus state to hold its 2nd nominating event (not counting Washington's non-binding primary), which was its county conventions on February 23. However, the county convention in Clark County (where Las Vegas is located) was so poorly planned that attendance overwhelmed the ability of the event to continue and they had to recess the convention until April 12 (see this article for the sordid details). The other county conventions proceeded normally. Clark County has the largest share of delegates among the counties in Nevada. This means that the majority of delegate votes for this event (2,463) have not yet been counted, and won't be for another month. The result is that listing the updated pledged delegate estimates for this event shows some really strange numbers. How do you think we should list the estimates for this event in the Overview of results table? I see the following options, but I'd be happy to hear other suggestions: Help! --Bryan H Bell (talk) 12:51, 21 March 2008 (UTC)
 * list the event as 94% reporting, list the estimates with a large "uncommitted" number, add a footnote explaining that the uncommitted number actually represents uncounted delegate votes, and avoid using strikeout for the estimates of Nevada's first event (this is the way I've left the table so far)
 * do the same as above except apply strikeout to the estimates for Nevada's first event (as we've done for Iowa's first event)
 * separate Nevada's county conventions event into two rows, one on February 23 called "county conventions (except Clark)" and another on April 12 called "Clark county convention" and re-compute the estimates accordingly


 * Once the source of the estimates stop using "uncommitted" to mean "uncounted" then we'll be able to handle it normally. But agreed that until then we have an issue with Clark county (which BTW is where 70% of the population of NV resides [two of the three CDs for NV are entirely in Clark county and the remaining one is also partly in Clark]). Do we have a reliable source segreating the Clark delegates from the first event away from non Clark? In that case we might be able to leave the original event line alone, add the Feb 23 event under the title sugested and place a footnote that the original projected delegate count from non clark was such and such. Jon (talk) 13:21, 21 March 2008 (UTC)


 * Another thing is did Clark start any counting on Feb 23 before going into recess? If they did start some counting before recessing and it was picked up by a reliable source then we have an additional complication. Jon (talk) 13:27, 21 March 2008 (UTC)


 * I say we do the same thing we did for Iowa. Just strikeout the caucus results, we'll update them after Clark County does there's. – Zntrip 17:39, 22 March 2008 (UTC)


 * The more I think about it, the more Zntrip's makes sense. Add to the date column a "-" and the date of Clark County concludes. But in the mean time I think the "94%" in is highly misleading; that is 94% of the counties of Nevada but only about 30% of the population. I think it should instead say in s "All but Clark County (Las Vegas)" Jon (talk) 13:21, 25 March 2008 (UTC)


 * Good suggestion. Essentially we'll treat Nevada's county conventions as if they are taking place over a range of dates and witholding results until after the end of the date range. This is how we handle other multiple-date events like the Democrats Abroad primary. I've therefore moved Nevada's county conventions from the February 23 position in the table to the April 12 position and displayed the date as February 23–April 12. I removed the "94%" as well as the partial delegate counts. I also added a footnote to the date explaining the situation. --Bryan H Bell (talk) 23:05, 25 March 2008 (UTC)

It is all extremely confusing even for us (and we are very experts of primaries by now). Elections mixed with conventions; popular events with assemblies; the (blue) winner of a state replicated 4-5 times, even if he really wins only once; parentheses "(" used both for Edwards (don't understand why) and for Iowa first event, Strike-through text used both for MI/FL and for Iowa (and further when available). I repeat that the summary-popular event based solution was much better, in particular for readers who don't want to be confused, but I know I'm in minority. A lot of times the simplest solutions is also the best, the problems with the adopted solution will be many. --Subver (talk) 12:58, 22 March 2008 (UTC)


 * While I don't think it's productive to re-hash recent consensus decisions, I do think we might be able to find ways to manage some of the concerns Subver raises:
 * "(blue) winner of a state replicated 4-5 times, even if he really wins only once" The blue/bold highlight doesn't indicate which candidate "won" each state, only which candidate has the most delegates in a given event. However, since each event in a state is really only deciding the distribution of the same delegates, maybe it would make sense to only use the blue/bold highlight in the most recent event in each state, leaving past events without highlights.
 * "parentheses "(" used both for Edwards (don't understand why) and for Iowa first event" This is now being discussed above. I support Subver's suggestion to remove the parenthases around Edwards' delegate numbers.
 * " Strike-through text used both for MI/FL and for Iowa (and further when available)" I think strikethrough text makes sense for MI and FL since those results have been nullified, but maybe we can find a better way to display estimates from past events (like Iowa's caucuses) than using strikethrough. I'm not sure yet what that better way might be, but I think it's worth some discussion to figure it out. --Bryan H Bell (talk) 21:07, 22 March 2008 (UTC)