downloadable pdf (1.03mb) "#MyBroadbandvReality submission to the Senate Select Committee on the National Broadband Network"
downloadable pdf (2.77mb) "MyBroadband vs Reality survey results" [excel available upon request, minus identifiable names of course]
What a week!
What can I say, I have been humbled and lucky to have had the opportunity to have spent a few days collaboratively creating a Senate Submission. Noely over at YaThinkN has a great summary post (as well as submission recreated if you don't want to read a PDF). In numbers:
- Over 800 Submissions
- 7 days to do it in
- 4 main editors
- 3 time zones
- 2 continents
- 1 submission
It's been a blast. The submission itself only has one graph, so heres a few more with breakdowns by State/Territory. A bit more writing will appear around these over the next few days.
Update: 1 March 2014 - Distribution, median, mean, and mode.
We collected 817 responses, of which we had 758 where I could graph (estimate v speed). The 65 I could graph a comparison were due to MyBroadband not having an estimate, or an difficulty in getting a reality speed reading. To calculate the distribution of the figures, I took the MyBroadband estimate, and the minus the reality figure, giving a "Reality Gap". By ignoring the decimal places, it's possible to group results into 1Mbps slower, 2Mbps slower etc. The following chart chart is the result.
You may notice the from this group the largest "group" of people in the survey, 78 results, matched or were slightly faster than the MyBroadband estimate, their Reality Gap was zero. 72.8% of participants were 1Mbps of more slower than the MyBroadband estimate.
Another thing, the MyBroadband site used "median". There was (I think) a good disucssion about this use on the original YathinkN post (check out the comments, yours truely made a few). I noted while median was used, there was no mean (average) or mode (most frequent). So, in the spirit of fairness, below is the Median, Mean, and Mode for the survey results.
The table above is for all 758 with workable figures. If we reduce that to only the ADSL figures (N=688) we get the following table.
So the Mode being the same between the first and second table tells us the ADSL figures are providing the Mode. Having the Median increase while decreasing the number of observations (N from 758 to 688) doesn't tell us much, except that the middle figure in the observations is higher. The Mean (average) dropping between the two is the impact of removing the speeds higher than ADSL. You may see from this why some of us thought it was important to include these other measures if a site is going to talk about Median, as the MyBroadband site has.
Thought I'd start with a few words about the NSW chart. A few astute readers have picked up on the 30Mbps outlier for the speed reading. As this graph is about ADSL that speed reading is "a little on the high side" :) This is because the current maximum speeds for ADSL2+ in Australia is around 24Mbps. I think it is likely the participant entering in that value simply didn't select the correct item in the drop-down box. If we re-do the NSW graph without that entry it looks like this.
Spreading the word
- Please show your appreciation by following and supporting @lynlinking who works very hard to disseminate both blog posts & relevant media reports on a daily basis :)
Thinkyness came about from the charming @Perorationer who often tweets Jan & Noely in the morning with a random article he has found that would make you think, could be anything from Women in the 1800's to a potential world wide wine shortage which we would then discuss, obviously this led on to us tweeting each other #Thinkyness articles [...] more
All articles displayed on this website are the personal opinion of the writers and do not reflect the official position of any businesses or organisations any writers may be associated with.