Tuesday, December 8, 2009

Mr Barbarosie should consider a new profession

Arcadie Barbarosie has published his latest opinion poll.  I won't comment on the results, as I simply don't believe them.  Consider the following:
  • Barbarosie has AMN coming in at 1.8%, way below the threshold.  His polls prior to the 29 July election were giving similar results, and yet AMN comfortably exceeded the threshhold and secured 7 seats.
  • Barbarosie gives the PSD 2.7%.  This is a party that has suffered through the loss of a co-leader (Musuc) and the recent dissensions relating to the proposed fusion with the PD.  There is no way they enjoy 2.7% support.
  • According to Barbarosie, 43.7% would want to rejoin the Soviet Union, while 16.3% would think about it.  60% of Moldova's population prepared to countenance a return to totalitarianism?  Crazy.
Barbarosie's polls have a tendency to overly favour leftist views.  In part this is accomplished by leading questions (e.g. the question about the USSR above came immediately after one which asked "Are you living better than you did under the Soviet Union?"), however I suspect there is also some sampling bias (or worse, manipulation of the data) which needs to be addressed urgently.

Otherwise, perhaps Mr Barbarosie should take up a career writing fiction.

4 comments:

  1. Zimbru, there where other two polls (from IDIS and IMAS) in November that showed similar results. This is not something Barbarosie does on purpose - AMN voters decide just before the election (and around 30% are undecided).

    IMAS gave 2% to PSD. So what? The error margin is 3%.

    About the Soviet nostalgia: this a really sad truth, unfortunately. It is again proven by many, many polls. You didn't think the communist voters are very smart people, didn't you?

    And yes, IPP's poll has a leftist bias, because the commies are less undecided than others. But it is the most reliable poll we have, and the fact that it gives now less than 40% to PCRM is quite remarquable. Actually it might be the first time in many, many years.

    ReplyDelete
  2. The only part that seemed odd to me was the results for the questions regarding social issues and foreign relations. In particular, it wasn't just the absolute numbers, but the trends that were quiet odd, for example in some cases younger people being the most enthusiastic about rejoining the USSR and more ethnic minorities considering the Moldova's main foreign partner should be Romania than ethnic Romanians!!! These trends lie in direct opposition to previous polls and most importantly to common sense!

    On the other hand, as Corneliu said the results for the politics section are quite reasonable. So I don't really know what to make of this poll, if maybe people didn't take the other categories too seriously or if there was just some statistica anomaly...

    ReplyDelete
  3. Thanks to both of the readers above. In your comments you have provided the sort of analysis that is so often missing from news reports.

    And yes, I may have been a bit tough on Mr Barbarosie...

    ReplyDelete
  4. I just got hold of the report. What makes me suspicious is that they publish answers by age, sex, nationality etc, but that nowhere in the report can I find the actual demographic breakdown of the respondents.

    I am unable to check that around 50% are male, that around 80% are Romanian / Moldovan, or that the age structure approximately matches that of the general population.

    Hence my suspicions regarding sampling error.

    ReplyDelete