Most on-line forums offer opinion polls on a wide variety of topics, as does OLO. I visited Andrew Bolts site to find out more about the launch of his half hour Sunday morning gig ( I know, I will have to wear sack cloth and ashes as penance). I noticed his poll which asked “will you watch this program”, Yes or No?

As of this morning the poll shows 36% Yes and 64% No.

Curious? This is AB’s web site, why would 64% of his faithful vote against him? This opens up a number of possibilities.

Firstly, that two thirds of those visiting his site are hostile. This seems unlikely although if true, he would appear to be a greater threat to some than first imagined.

Secondly, that those visiting his site prefer direct interaction with him on his blog, rather than seeing him on TV. Which seems a bit like “conditional support”?

Thirdly, is it possible that his opposition is hitting the “No” button in great numbers to skew the numbers in a hostile act founded on a perceived threat?

This is a very curious result, any explanations?

GY, do you have any concerns in relation on-line opinion polls and have you any suspicions of your polls being distorted?

Must be spin doc, surely not?
34% said they would watch him? rather eat worms!

I answer these polls all the time.

I think about what would be the popular response, and pick the opposite. I'm sure I'm not alone.

Maybe people read Andrew because they want their daily dose of outrage and self righteous indignation. That's why I read lots of stuff I don't agree with.

I remember that some years ago the ABC introduced a "right wing" radio talk show as a counter to allegations of leftist bias.

Like the long defunct Alan Jones TV show, it bombed in the ratings and sank like a stone.

Maybe some viewers are more interested in a show that promotes fair discussion and not as just another pedestal for a zealot to rant from.

You can get the same thing from any number of radio shock-jocks.

I guess Gina Reinhardt may have made a bad call in sponsoring him but he can still get his $800-odd appearance fees from Insiders if it doesn't work out.

Considering Spindoc's three questions in reverse order, I would begin by observing that any claimed online poll result would only be worth a glance if that poll had been conducted under a 'you may only vote once in this poll' rule. The 'you' referred to is, of course, the uniquely identifiable interaction of a particular computer ('yours') with the software of the site conducting the poll.

The tacit assumption is that few persons have access to multiple connections to the internet which they could, if desiring to skew a result, use to effectively vote multiple times in any one poll. I am not sure as to whether each computer on a LAN (local area network) having internet access is uniquely identifiable by the polling site's software should a vote be attempted to be cast by a single person having access to the multiple computers on that LAN. If each computer was uniquely recognisable, one would tend to think administrator privileges with respect to that LAN would be needed if one person was to use those multiple computers to cast recognised 'genuine' votes in the poll in question.

Notwithstanding the seeming practical unlikelihood and/or difficulty of one or a few motivated individuals overcoming the technical barriers to registration of multiple votes in an online poll, it is not hard to imagine that a commercial or organisational entity with an interest in skewing a result might mobilise multiple online 'identities' in the form of many different computers and their associated internet connections in order to do so. Boosting, or retarding, the apparent popularity of a media celebrity (or bete noir) as revealed by an online poll might be just the sort of occasion where the deployment of such resources could be considered worthwhile, provided, of course that it was considered likely that someone was going to believe, and act in some way upon, the (skewed) poll result.

Would such online poll results be likely to be used, for example, in deciding whether Andrew Bolt's TV slot would be retained or cancelled by a media organisation? You tell me.

spindoc's second question highlights the fact that a response in terms of the question asked is not necessarily fully revelatory as to whether the respondent is generally supportive of, or opposed to, in this case, the views expressed by Andrew Bolt. A negative response may mean nothing more than that the respondent is unwilling or unable (eg. through work or other commitments) to watch such TV program at that time of day.

A negative response may just as easily reflect an overriding dislike or distrust of the particular medium (eg. TV in general, or even the specific channel) upon which AB may be appearing, without there being any dislike of AB's views. It may indeed be the lack of interactivity implicit in a TV show that may be being voted against (presuming the votes to be all 'genuine')in this blog poll, not the commentator. spindoc refers to this latter possibility as reflecting 'conditional support' for AB.

All of which brings us to the question as to why, whether as a passive (?) audience, or in perceived interaction as with blog comments, some people follow such commentators? Is it because such commentators articulate views that their followers feel unable, or are too time poor, to express so well themselves? Is it because such views aren't seen to be being articulated competently, or at all, in what is expected as being the 'real world' of politics, where in due course there might be an expectation that some of those views may eventually be given legislative effect?

Finally, to address spindoc's first question as to whether it is possible that "two thirds of those visiting [AB's] site are hostile", I think it might have to be admitted that it is possible, if on other opinion or organizational sites hostile to AB's views there are links posted with encouragements to vote down, for example, a proposed or newly announced TV slot. The problem is that this would not fool the blog site operator, as trackbacks would show the links through which such votes had come.

Food for thought.

There is an Arlo Guthrie song about 'the last man', then man that we all need, to feel sorry for, to humiliate, to kick when they are down, to laugh at, to scorn.

That is the role of Andrew Bolt.

Once you've peed on Alan Jones and David Oldfield, there's always Bolt.

He makes us feel better, superior, better than him.... see, we all need a 'last man' to make ourselves feel better.

Forrest Gumpp,
Exactly and well expressed.

Well said TBC.

Every creature contributes to the whole.

:)

Too true Ammonite, even the slugs of the forest, the humble clegg (although maybe not the UK one), and the slithering serpents of commercial TV and radio, oh, and all the NewsCorp staff even, oh, and Gillard and Abbott, yes, even the subspecies have a role to play.

spindoc concludes his opening post by asking:

"GY, do you have any concerns in relation [to]
on-line opinion polls and have you any suspicions
of your polls being distorted?"

Obviously I cannot answer that question for GrahamY, nor do I seek to, but I think it is worth noting that OLO opinion polling is seemingly done, at least in part, if not exclusively, 'by invitation'. I don't think I am breaching any confidences in quoting from an email sent by Graham Young to me (among others), which says:

"I'm writing to you because you have contributed
to our surveys in the past and you have indicated
you are happy to receive invites from us to participate
in more. This current survey looks at the hung parliament
along with the usual political questions."

It seems the basis for being invited to participate in the OLO surveys is at minimum that one be a registered user of the site, and to have contributed comments. (There are registered users of OLO who have never commented according to their user history, for whom OLO would nevertheless hold an email address: I wonder whether any of these are asked to participate?) Whether there are any additional selection criteria applied, such as having posted to particular articles or topics, or total numbers of posts over time, used as the basis for invitation to participate, I don't know. GrahamY also has included this invitation to addressees:

"If you know anyone else who could be interested
in completing either questionnaire, please forward
this email on to them [using the link below]."

It would therefore appear unlikely that someone wishing to skew an OLO result by voting multiply (other than as an as-yet-undetected sockpuppet) would be able to casually access the survey incognito, the address for which is contained in the invitation email. There would exist an electronic papertrail in any event, should suspicions be aroused as to skewing.

I'm assuming that observations as to the perceived merits or demerits of AB would be considered off-topic by the OP. My position: http://forum.onlineopinion.com.au/thread.asp?discussion=4393#111731

spindoc's opening post describes an aspect of online polling that can easily go unnoticed.

That aspect is that the results displayed are in many cases PROGRESSIVE results of a poll still being taken, rather than FINAL results posted after a poll has closed.

It consequently has to be questioned as to whether the publication of progressive results, while the poll still remains open to further voting, facilitates or encourages any 'bandwagonning' effect amongst those who have not yet voted, or among those who, otherwise than as a consequence of learning of the progressive result, may not have intended to vote at all, to cast a vote. Should a 'bandwagonning' effect in fact exist, would it amount to a skewing of the poll results?

If a 'bandwagonning' effect is recognised amongst pollsters generally as existing, could the negative result on Andrew Bolt's site be one deliberately engineered by those running the poll in the belief that more viewers of the site might thereby be encouraged to (genuinely) vote? Those running the poll and displaying the progressive results would know at any point precisely the extent to which, and to what effect, they had inserted dummy votes. Those viewing the site (the 'mug public'), and its progressive poll results, would have no such information, and would be able to be kept in the dark as to the real result, and size, of the genuine sample of opinion taken.




To the outside view, a display of seemingly paradoxical unpopularity of AB's new TV show. To those really running the poll, secret knowledge as to the true result otherwise hidden in plain sight from the 'mug public', for what it may be worth, of the poll taken.




It indeed may be that what those conducting the online poll may be sampling is not AB's popularity, or the lack of it, but the extent to which AB's TV show's apparently flagging popularity in the poll elicits any response from the viewing public to attempt to either counter, or emphasise, that apparent result by voting in that poll.

Caveat elector!