The nice man from iSelect Corporate Affairs has just rung me to inform me that he disagrees that the media release purports that there is something more to read. The proposition is that "the research has been released through the media release".
I am grateful for the information. He has at least pointed me to the full release on the company website. It is interesting how this varies from the versions on various newsfeed sites like Press King. I am also aware that at least one media organisation only received a version through one of these sites that did not include the footnotes on how the research was conducted.
These footnotes make interesting reading. The third footnote provides details on how the "research" was conducted (naming the firm and as a sample of 1000 households). Three of the footnotes though look like the additions of a zealous fact checker. Three amounts in the release ($141M overspend, 368 million gigabytes and $200 for data) are carefully reproduced as actual numbers to the last cent or Gigabyte below. It is hard to know which is really more misleading - given that the representation to the cent represents an assertion of accuracy beyond the bounds of the survey methodology. The correct way to footnote the data points from a survey is to quote the data point and the estimated sampling error.
I will stand by my view that the first and third paras of the release assert that a thing called the "iSelect Broadband Report" has actually been released, not just a set of data drawn from that research.
But my real gripe is with the media outlets that churned out the data from this release without asking to see the basis of the report. I am still at a loss to understand the basis of the assertion of the $141M overspend. My guess is that the research asked people what their current plan and usage was, looked up the best plan on the iSelect site and made the straight comparison. But we don't know because we are not told.
I will save for another post a comment on the economics.
Novae Meridianae Demetae Dexter delenda est