Tuesday, April 7, 2009

OPINIONS, NOT OPINION POLLS

Watching the "Opinion Polls" on NDTV and CNN-IBN often leaves one with a feeling that these polls reflect not the opinion of the people who are going to vote but of "opinion makers" who are using the platforms of these channels to project a Congress victory possibly in the hope that such projections will push undecided voters into backing the winning horse and thereby make these polls appear credible.

Recently, both the channels aired the results of the Opinion Polls carried out by them in respect of the critical state of Andhra Pradesh that sends as many as 42 MPs to the Lok Sabha. An analysis of what was shown to viewers has revealed that statistics were cleverly highlighted/hidden to fit what appeared to be an inflated projection of the seats that the Congress party will win in the state.

Before we get to the opinion polls, a look at the vote share in the 2004 Lok Sabha elections is necessary.

2004 Elections Vote share

Congress- 41.56%
TRS - 6.83%
Left - 3.61%
----------
Total -51%
----------

TDP - 33.15%
BJP - 8.41%
-----------
Total 41.56%
----------
Others -7.44%

Alliances for 2009 Elections

In these elections, the Congress has lost both the TRS and the Left to the TDP. In lieu, it has got the AMIM whose influence is limited mostly to the Hyderabad seat. Therefore, if voters vote exactly as they did in 2004 for the alliances as they exist today, the vote share will be as follows:

Congress+ - 42%
TDP+ - 43.56%
BJP - 8.41%
Others - 7.44%

In 2004, there was some sort of a wave against Chandrababu Naidu. This time nothing of the sort, for or against anyone, has been spotted by any pollster. This election, however, sees the emergence of a new party, the Prajarajyam Party (PRP), floated by Telegu superstar Chiranjeevi. He is no NTR but has a strong following in some parts of the state outside Telengana. Naturally, both the CNN-IBN and NDTV pollsters are not able to put their finger on exactly what percentage of the voters will vote for PRP. Nevertheless, they have made their assessments and after factoring them in, have given out their verdicts on the likely results that Andhra Pradesh will throw up.

CNN-IBN Opinion Polls

CNN-IBN's first opinion poll this year was telecast in February and was based on a survey that was carried out before the Satyam scandal surfaced. It may be recalled that Andhra Chief Minister YSR Reddy has been accused by almost everyone of being involved with the Raju family in this huge corruption scandal running into thousands of crores of rupees.

This poll was also carried out before the TDP, the TRS and the Left formed an alliance. The party wise vote share then was:

Congress - 45%
TDP - 30%
BJP - 9%
PRP - 7%
Left - 2%
Others - 7%

The vote share of the TRS was not seperately given. But, if one were to go by what that party got in 2004, its alliance with the TDP along with Left means that this Grand Alliance, as per the figures of CNN-IBN, was getting almost 39% of the vote share in February. Notice the following anomalies in this projection:
  • The vote share of the Congress party has actually gone up by around 4% since 2004 despite the loss of key allies, and despite being in power for five years in the state.
  • The vote share of the the TDP has declined by 3.51%, even though there is no wave against the party this time.
  • The vote share of the Left has declined by 1.61%.
  • The vote share of the BJP is almost the same.
  • The vote share of TRS plus Others has declined from 14.27% to 7%. This means that either the TRS or the others have been decimated.
  • The 7% vote share that the PRS has created for itself comes from the TDP+Left (5.12%) and the Others, including the TRS (1.88%). It clearly gets nothing at all from the Congress. On the contrary, the Congress itself has got additional voters from elsewhere.
These projections clearly suggest that there is a real "wave" in favour of the Congress party.

There is little doubt that Rajdeep Sardesai and Yogendra Yadav knew that these statistics did not reflect the real situation on the ground. That is probably why in their latest Opinion Poll telecast on April 7, they have simply done away with the pretense of giving party wise vote share altogether! Seats predicted have, therefore, been called "simulation", not "projection". As per this "simulation", the basis of which remains unclear, the Congress (plus AMIM) is doing very well with 26 seats, of which the AMIM is not expected to get more than one seat. The TDP has 16 seats. The PRP has nil, despite its vote share now being between 8% and 15%, against 7% in January. But, says Yogendra Yadav, if there is a wave for PRP and it gets at least 25% of the vote share, then the Congress will get 22 seats, the TDP 15 and the PRP five.

No statistical data obtained during the latest Opinion Poll has been shared with the public to support this projection by channel. This must surely be a first in the history of opinion polls anywhere in the world.



NDTV Opinion Poll

Prannoy Roy and Dorab Sopariwala, in the Opinion Poll telecast by NDTV 24/7 on April 7, did not dispense with statistics totally but used them very creatively to arrive at their own seat projections. This was the projected vote share shown by them:

Congress+ - 40%
TDP+ - 38%
PRP - 14%
Others - 8%

The breakdown of Others, down sharply to 8% from 15.85% was not given at all. It must be mentioned that out of these Others, the BJP alone had 8.41% vote share last time. This time half of them have inexplicably vanished into one party or another!

But more important was the manner in which Prannoy Roy erroneously compared the present 40% vote share of the Congress party which is now alone, except for AMIM, with the 51% vote share of the big alliance that the Congress had last time, rather than with the 41.56% it had on its own. Thus, although as per NDTV's poll, the Congress party has lost only around 1.5% of the vote share that it had in 2004, Roy made it appear as if it had lost all of 8% to PRP and only 3% to TDP, when actually Congress+ has lost 10.51% to the TDP by the desertion of its erstwhile allies to that party. Similarly, the new TDP alliance had a 43.66% vote share based on the figures of 2004 ( TDP 33.15+TRS 6.83+Left 3.61). But NDTV showed it to be only 39%, down further by 1% this time. If NDTV is to be believed, this was due to 4% being taken away by PRP , offset by 3% taken by TDP+ from Congress!

Simple arithmetic tells us that if the Congress party had actually lost 11%, then it could not have lost only 3% to TDP+. Similarly, if PRP has got 14%, as the survey says it has, then it means that it actually has actually got net only 1.5%, not 8%, from the Congress, 5.66% from TDP+ and about 7% from other parties. This means that while PRP has eaten heavily into TDP plus, the BJP and other small parties, it has not been able to touch the Congress. Is this believable? So, this "Chiranjeevi Effect" that Prannoy Roy says has damaged the Congress by as much as 8%, is actually not true at all ! Surely, the team at NDTV knows that.


Thanks to this clever jugglery, NDTV has given the Congress 24 seats, the TDP 16 and PRP 2.

If indeed the PRP has taken away 8% and the TDP 3% of the vote share of the Congress party, as Prannoy Roy was emphatically telling viewers, then the vote share of the latter is actually down to 30.5%! How devastatingly does that translate into number of seats? Will Dorab Sopariwala be kind enough to tell everyone?

Summary


As is evident from the analysis above, both the opinion polls have attempted to project the Congress as doing very well in the key state of Andhra Pradesh. CNN-IBN, in one of its polls, actually shows a near 4% swing in favour of the Congress despite Chiranjeevi getting 7% vote share. NDTV, on the other hand, slyly slips in the unnoticed fact that despite Chiranjeevi getting a whopping 14% of the votes, double that shown by CNN-IBN, the vote share of the Congress remains almost untouched at the same high level that it was in 2004!

We will get to know what the real voter really wants only on May 16, 2009. But, as the analysis of the aforementioned opinion polls shows, it is becoming increasingly apparent that such polls are becoming less and less about the real opinions of the voters they are supposed to represent and more and more about the political predilections of those who commission them. That is why, perhaps, such polls-turned-opinions are going increasingly off the mark, even though they should be getting more and more accurate. Credibility and trust are taking a hit. But who cares?


UPDATE 10 PM APRIL 14

NDTV's opinion poll on UP has just been telecast. Although the Congress is shown as getting only 8 out of 80 seats, NDTV's Rahul worship was in full display. Within this poll on UP, there were three polls about Rahul alone! First was whether Rahul can revive the Congress in UP. 70% said "Yes". On a different question about the most popular leaders in the state, Mayawati was first, Mulayam second and Rajnath Singh third. Rahul was nowhere in picture. Despite that and despite the virtual decimation of the Congress there despite Rahul's efforts for seven years, he was voted as the hope of the Congress for UP. I don't know what to make of it.

The second question was whether Rahul will beome PM one day. If I recall correctly, around 78% said "Yes". The third question, hold your breath, was whether he is ready to be PM now. 52% said "Yes"! Why was that question asked in the first place? Are we heading for a Kashmir type surprise should the Congress come to power?

Even DD was never so sycophantic! So much for a free and honest media. These guys seem to have no qualms!

The real IPL T20 is on in UP! -2
-------------------------------------------------------------------------------
Readers may also read:
1. India Today-ORG-MARG-AC Neilsen Poll
2. Hi-Tech Vs People's Rule - Andhra Elections
3. Andhra Pradesh region-wise seat wise predictions
4. Swift message from swing states
5. Covering up the mother of all corruption scandals
6. NDTV makes a mockery of exit polls.
7. Real IPL T20 is on in UP!
8. NDTV sees Congress steaming ahead after Phase 2

No comments: