Marcan
Senior Member
I have made already a few critical remarks earlier about these polls/surveys on Disqus. They are often poorly thought through, and the results are therefore not really valid.
When you ask your readers to rate certain services or products and ask them which brand do you value the most, most obviously most people will vote for the most popular services or brands. This might just reflect their popularity, not actually the quality of the products or services. When a certain bank A has 45% of market penetration, account holders, mediocre service levels and quite high rates, most of the users will still be happy with their services, otherwise they would have left. Maybe 60% of them will rate them the most valued bank in SA.
When a competitor B has extremely good services and lower banking rates, but has only 20% users, 90% of that 20% might rate them the most valued bank in the country.
60% of 45= 27% will vote bank A as the most valued bank.
90% of 20= 18% will vote bank B as the most valued bank.
Such a survey design only confirms existing biases. Many other errors and biases might creep in poorly designed and executed polls and surveys.
No doubt there is plenty of literature in the field of statistics, sociology, economics on all this.
Have been reading Freakonomics, and started on the sequel SuperFreakonomics. Written by a lateral thinking economist and a journo. An alternative way of looking at things, and economic data. Unconventional wisdom, things are often not what they seem to be, when only having a superficial glance at matters.
Yesterday's MyBB poll even contained obvious misinformation. One was asked to rate a number of mobile. wireless data offers;
Which data package do you think offers the best value?
I made a Disqus comment below the latest MyBB article : " I thought Cell C had much better deals and only could find a Mediaplay SIM only deal with 20Gb per month for R 239/pm and free streaming and subscription of the Cell C Black streaming service on their website. Another Cell C deal is a Fixed 4G deal which includes a home router and 15Gb per month for R 219. "
And added that this was poor journalism, well I used a bit stronger lingo. As expected the Disqus post was immediately removed.
Many visit MyBB to get info about telecoms and internet services, of which a large part is not so well informed to start with. They might easily conclude that Cell C data offers are extremely poor value for money.
Grossly misleading.
The biggest poll in SA took place on 8 may, and 57% of the the ones that actually bothered to vote still voted for the same misruling party. They all have their specific reason for that. I have been discussing this with many supposedly typical ANC or EFF voters. the last years, especially the last half year. Very interesting indeed. Will discuss that maybe another time.
When you ask your readers to rate certain services or products and ask them which brand do you value the most, most obviously most people will vote for the most popular services or brands. This might just reflect their popularity, not actually the quality of the products or services. When a certain bank A has 45% of market penetration, account holders, mediocre service levels and quite high rates, most of the users will still be happy with their services, otherwise they would have left. Maybe 60% of them will rate them the most valued bank in SA.
When a competitor B has extremely good services and lower banking rates, but has only 20% users, 90% of that 20% might rate them the most valued bank in the country.
60% of 45= 27% will vote bank A as the most valued bank.
90% of 20= 18% will vote bank B as the most valued bank.
Such a survey design only confirms existing biases. Many other errors and biases might creep in poorly designed and executed polls and surveys.
No doubt there is plenty of literature in the field of statistics, sociology, economics on all this.
Have been reading Freakonomics, and started on the sequel SuperFreakonomics. Written by a lateral thinking economist and a journo. An alternative way of looking at things, and economic data. Unconventional wisdom, things are often not what they seem to be, when only having a superficial glance at matters.
Yesterday's MyBB poll even contained obvious misinformation. One was asked to rate a number of mobile. wireless data offers;
Which data package do you think offers the best value?
- Rain 1GB for R50 with unlimited off-peak data (43%, 1,126 Votes)
- Vodacom 20GB for R199 (16%, 427 Votes)
- Telkom 70GB for R459 (15%, 405 Votes)
- MTN 20GB for R199 (10%, 267 Votes)
- Vodacom 10GB for R149 (6%, 149 Votes)
- Telkom 40GB for R359 (5%, 132 Votes)
- Telkom 20GB for R249 (4%, 117 Votes)
- Cell C 10GB for R301 (1%, 22 Votes)
I made a Disqus comment below the latest MyBB article : " I thought Cell C had much better deals and only could find a Mediaplay SIM only deal with 20Gb per month for R 239/pm and free streaming and subscription of the Cell C Black streaming service on their website. Another Cell C deal is a Fixed 4G deal which includes a home router and 15Gb per month for R 219. "
And added that this was poor journalism, well I used a bit stronger lingo. As expected the Disqus post was immediately removed.
Many visit MyBB to get info about telecoms and internet services, of which a large part is not so well informed to start with. They might easily conclude that Cell C data offers are extremely poor value for money.
Grossly misleading.
The biggest poll in SA took place on 8 may, and 57% of the the ones that actually bothered to vote still voted for the same misruling party. They all have their specific reason for that. I have been discussing this with many supposedly typical ANC or EFF voters. the last years, especially the last half year. Very interesting indeed. Will discuss that maybe another time.
