Pollsters in developed countries

In light of the election results, NACTA head Vishnu Bisram said

“I feel vindicated that the results have borne out what the NACTA poll predicted.”

As they say, even a stopped clock is right twice a day.  Yes, they got the numbers right (if you stand on one leg, cock your head, and squint you can just about see the match) but no, the results didn’t vindicate them.  But it’s the second part of Bisram’s statement that I find interesting.

Noting that one political party ran advertisements questioning his integrity as a pollster, the New York based teacher and political analyst said such an approach would never have happened in a developed country.

“I live in a developed society and in the United States of America, politicians do not attack pollsters,” he said, in a telephone interview. [Emphasis added]

Over the last few years of watching elections in the US, I have been spoiled by a wealth of data.  There are entire websites dedicated to the analysis of polling data (e.g., TPM Election Central).  Pollsters in developed countries?  Yep, I’m familiar with the breed.  You can go to their websites and download their poll results.  For a few dollars extra, you can get access to their raw data.  For the most part, there’s transparency.  Sure, you still have to trust their actual data collection.  But you can always double-check their results.  But it appears that NACTA doesn’t even have a web site.

To begin with, politicians in “developed countries” do attack  push polling and similar attempts by pollsters to manipulate the electorate.  If the links between Bisram and Ramesh Maharaj are true, then attacks on NACTA are fair game.  If they aren’t, then it’s fair to criticise a polling organisation that got things so badly wrong.

If NACTA really wanted “vindication” they could release their raw data.  I think a polling outfit in a “developed country” would have done so.

NACTA vindicated?

While most polls showed the race to be between the PNM and COP, one polling outfit, NACTA showed it to be a PNM-UNC A race, with only marginal support for COP. For this, NACTA was attacked by COP, and claims were made of links between Vishnu Bisram of NACTA and Ramesh Maharaj of the UNC A. Bisram is now asking for an apology:

Vishnu Bisram, the man behind the NACTA poll which has accurately predicted the outcome of General Elections since 1995, yesterday called for an apology from those who had attacked his integrity as a pollster during this year’s elections race.

“They owe me an apology,” Bisram said in a statement yesterday, noting that his “character and integrity” had been attacked by various parties. “I feel vindicated that the results have borne out what the NACTA poll predicted.”

Unfortunately for NACTA, the overall result was only part of their prediction. They also predicted the split of the popular vote:

“So overall,” the poll concluded, “the battle for control of the government is between the PNM and UNC A with COP making the difference in several seats and also having an outside chance of picking up a few seats.
Both the PNM and UNC A have seen a dip in their support from the last election. The PNM is projected to win about 43 percent, UNC A 35 percent and COP 19 percent of the popular votes.

The stated margin of error was “between four and six percent”. If this is correct, then the margin of victory of not only the marginal seats, but also many of the core UNC seats, was smaller than could have been detected by the NACTA pollsters.

If the NACTA poll had captured an accurate picture of the views of the electorate in these seats, they would have come up as “toss ups”. Results within the margin of error are impossible to interpret. The only way to square the NACTA poll with the actual result of the election was if there was a swing of more than 10 points away from the UNC A in the last days of the election campaign. It’s also possible for a polls to create their own reality. It’s obvious from the UNC A’s campaign that they were playing on concerns that “vote splitting” would allow the PNM to win. That factor alone would probably drive undecided voters to break for the UNC A. The idea that there was a 10-point swing swing toward COP isn’t credible. It’s far more likely that NACTA did help create the perception that COP couldn’t win, and push voters toward the UNC A.

NACTA vindicated? Not by these data.

Was there a plan?

I’m probably just a slave to tribalism and the cult of The Leader, but there’s a part of me that still believes that Bas must have had a plan, that there must be some reason (nefarious or otherwise) for his actions over the last couple years.  Why create a rump of a party?  Why drive all the talent out?

The obvious explanation is that he is willing to do whatever it takes to cling to a modicum of control – that it’s better to be ruler of your own little domain than to share power.  If you read Raffique Shah, you’d conclude that’s the way Bas has always been – in the ULF, in the NAR – he will always sacrifice the greater good for personal power.  But, for whatever reason, I have never been satisfied with that explanation.  It seems incomplete.  There’s got to be something more, right?  It’s thought processes like that (flawed as they may be) that lead me to ask “why?”  Why did Bas do what he did over the last couple years?  More importantly, why did he refuse to name a leader?  Sure, there’s a stain on his name.  But would being coy actually convince anyone to vote for the UNC who wouldn’t do so otherwise?  It struck me today that there might be an answer to both the question of “what is he doing?” (with respect to the whole power struggle with Dookeran in the UNC) and “what is he doing?” with respect to his failure to name a leader.  Mikela Panday. 

It’s entirely possible that Panday will simply hold onto the position of Leader of the Opposition, with the hope that he will still be able to fight the next election at the age of 80.  It would make more sense, of course, to allow the party to select a new leader, someone who could build a national reputation in preparation for 2012/13.  But who would that be?  The obvious choices have come and gone – Ramesh Maharaj, Winston Dookeran.  If he’d had Kamla Persad-Bissessar in mind he would not have turned her down in the run up to the election.  And I don’t see Jack Warner as being acceptable to the party.  Looking at incoming class of UNC MPs I see only a handful of possibilities: Tim Gopeesingh, Roodal Moonilal and Mikela Panday.  Of the others, Ramesh and Ramnath both returned to the in disgrace.  I don’t see Partap, Sharma, Baksh, Hamza or Subhas as having what it takes to lead the party.  Gypsy has the same problem Warner does.  Bharath is too much of an unknown for me to say much, except that I see someone with so little of a national profile being selected to be leader.

So my money’d be on Mikela, if not now, then within a year.

Still 22%?

During the 1971 No-Vote campaign the PNM won all 36 seats and received about 22% of the eligible vote. That marks their core support – about 22% of the electorate. According to preliminary figures, the PNM received 236,420 votes out of an electorate of 990,352. In other words, 23.9% of the electorate.

Update: Newer figures show that the PNM received 46% of the votes (299,813 votes) cast and the support of 30% of the electorate.  Turnout was a fairly normal 66%, not an unusually low 56%.

26-15-0

The PNM won the election, winning 26 of the 41 seats and 236,420 votes. The UNC A won 15 votes and received 103,247 votes. COP won no seats and received 119,007. About 56% of the electorate turned out. I estimated that the PNM received about 43% of the votes cast.

We’re basically back to the ONR in 1981. They won the second largest number of votes, but “not a damn seat”. About half the people whot voted support the PNM, about a quarter support Panday’s party, and about a quarter want something else. Ignoring, for the moment, the 44% of the electorate which didn’t vote (which includes people like me, who aren’t in the country and thus can’t vote), the inadequacy of the first-past-the-post system is just glaring. As in 1981, the second largest group of voters have no representation. Yes, we get to cast our votes. Yes, elections are mostly free and fair. No, this isn’t democracy.

Update: Newer figures today show that the PNM received 299,813 votes, the UNC A 194,425 votes and COP 148,041. (That’s what I get for taking Newsday seriously!) So I was wrong, COP didn’t receive more votes than the UNC A. Nonetheless, the basic fact remains that almost a quarter of the people who voted have no representation in Parliament.  Total turnout appears to be 66% rather than 56%.

Rationalising elections

I could say that the people didn’t act in their own best interests.  A UNC supporter could say that the COP supporter didn’t act strategically.  A PNM supporter could say…I don’t know, I suppose they could say “we win”.

Supporting a third party is always a risk.  Trying to get an electoral system to do something it isn’t designed to do is always a risk.  The commentators were saying that people support a party, not a person.  People hated Ramesh Maharaj in 2001 when he helped the PNM win.  The voted for him today, once he was back in the fold of the UNC.

If it’s true, and people will support “their” party (pretty much no matter what), how do you bring about change?  One of the commentators said that there were three case where individual beat party – Tubal Uriah Butler, Eric Williams and Basdeo Panday.  I don’t know about “ancient history” – Eric Williams, Butler.  I know Panday.  I also know that Panday has been talking about retirement for years, but even if he actually wishes to die with his boots on, will he really be able to face another election?  I have my doubts.  SO does he anoint a successor, pass the reins to someone else, or does he continue to hold until he dies and then lets those who survive him fight it out?  The people support the party, and the party supports the leader.  Patrick Manning showed that it doesn’t take any skill, just being in the right place at the right time.

And what of COP?  There’s a real constituency of people who want something different.  They are a third party, waiting in the wings, waiting to happen.  A third force in a two-party state, coming together once every few decades, then fading back, sitting uncomfortably in one party or the other.  Or do the people who who led COP stay there, and try to change the political culture?  Trinidadian culture is changing.  It takes forty years in the wilderness to forge a new tribe.  But what we want isn’t a new tribe.  We want something beyond tribe, something beyond maximum leaders, something more democratic.  Is that too much to ask?

Defeat

Manning won re-election for the first time. Three-party elections don’t work in a two-party system. So what next? Five more years of PNM mismanagement, five more years of worsening crime.

This really sucks. 😦

Update: I give up.  (My radio link went dead anyway). I’ll see what the final tally is in the morning.

results – PARTIAL

Place	PNM	UNC	COP
Arima
Arouca/Maloney
Barataria/San Juan	5516	4308	3133
Caroni Central
Caroni East	3864	8297	4235
Chaguanas East	6817	3367	3053
Chaguanas West	1499	11461	5104
Couva North	1631	2391	1499
Couva South	2764	3101	1724
Cumuto/Manzanilla	6944	7226	3079
D’Abadie/O’Meara
Diego Martin Central	4228	270	2143
Diego Martin North	2773	231	1661
Diego Martin West	9221	513	5600
Fyzabad	6021	6298	3088
La Brea
La Horquetta/Talparo
Laventille East/Morvant	11072	1201	1529
Laventille West	10637	483	1085
Lopinot/Bon Air	8535	3907	3190
Mayaro	2263	3200	837
Naparima
Oropouche East
Oropouche West
Pointe a Pierre	1132	746	383
Point Fortin	4820	765	1241
Port of Spain North/St Ann’s West	8359	304	3445
Port of Spain South
Princes Town North	5298	8231	3824
Princes Town South/Tableland	747	863	167
San Fernando East	10324	1630	3884
San Fernando West	2402	732	1568
Siparia	1418	3047	903
St Ann’s East	2512	215	1093
St. Augustine	4362	4392	5035
St Joseph
Tabaquite	5412	7099	5687
Tobago East	8664	980	1205
Tobago West	5771		231
Toco Sangre Grande	9433	3864	3214
Tunapuna	8303	3967	3896

note: Tobago West DAC 2632 DNA 0

Tobago East 412 DNA 0

Totals per constituency should end up being between 15,000 and 20,000

Note: I will update the totals once I get the numbers.  Seats won by the PNM are in red, those won by the UNC are in orange.  This graphic shows the breakdown geographically.

Results start trickling in…

As the first results start to trickle in, it looks bad. While there are just a handful of polling stations, it seems to be PNM or UNC first, COP second. But it’s consistently behind, and well behind. None of the results mean anything at the constituency-level, but there’s enough there to see a trend…not good for COP.

Update: COP has been ahead (at some point in time) in St. Augustine and Chaguanas East seats. UNC has been ahead in a few of the marginals, but there are others where PNM has about half the votes, UNC+COP about half. So the vote splitting may well leave the PNM holding several marginals…

Update II: Shockingly, the PNM is ahead in Tabaquite and Fyzabad…very slim leads, but that’s still amazing.  The PNM has Heeralal Rampertap in Tabaquite.  PNM 815, UNC 794, COP 665.

In Chaguanas East UNC is now in the lead, PNM in second place and COP in third place.  Sharon McNicol is actually doing respectfully (1000, to the PNM’s 2000 in PoS).  In San Fernando West Regrello (PNM) has 1416 votes, UNC has 423 and the COP’s Coudray has 804 votes.  In Princes Town South/Tableland, PNM has 747 votes, UNC 863 and COP 167.

Discussing the polls

As I mentioned earlier, I am suspicious of opinion polls with respect to Trinidadian elections. But I am even more suspicious of political commentators. One of the few exceptions is political scientist and columnist Hamid Ghany. The one problem is that he is too much of an academic – when asked about turnout he supplied actual data, and lots of it. It means that sometimes you have to wait a while until he gets to the point of saying “yes, so far turnout is high in Tabaquite”.

He is currently talking about Tobago and the revival of Tobago nationalism. Between 1976 and 2000 the DAC (and later the NAR) dominated politics there, but over the last decade the PNM made gradual inroads, to the point where they now dominate government in Tobago. There has been a rebirth in Tobago nationalism, which may have an impact on the electoral picture tonight.

Ghany suggested that the rebirth of Tobago nationalism was born of a number of things. The proposals for constitutional change left Tobagonians unimpressed. The EBC report increased the number of seats in the country by 5, but none of these were in Tobago. For the first time the smallest seat in Trinidad (San Fernando West) has fewer voters than either seat in Tobago. While Trinidadians are likely to see this as parity, Tobagonians will see this as dilution of their influence electorally.

Finally, Ghany said that some people think that Tobago gets a better deal when its seats are not held by the governing party. When Tobago is controlled by the governing party there is pressure to support that party. When Tobago is controlled by an opposition party there is no such need to toe the line. Some believe that Tobago gets a better deal in a situation like that.

Will this translate into a shift in power? Who knows.

Update: Doesn’t look so.  Preliminary results show the PNM candidates are very far ahead (in Tobago West Callendar has over 1000 votes; the other three parties probably have less than 200 votes between them; in Tobago East Stretch has a big lead).