Lies, Damn Lies, and Statistics

In looking at user comments on Ting service, I get the impression that there is a high level of customer satisfaction.

Thinking about the number that Ting showcases of the average bill being $23 across it subscriber base of 150,000 has me wondering what type of subscriber is the ideal fit for Ting service?

There is a blog post by a Ting rep reacting to the new Verizon unlimited plan that makes the following statement:

"With one device – 97.6% of our customers paid less than the $80 they would pay on Verizon"

That seems to imply for almost everyone going with Ting is a
"smarter" move than going with Verizon.

While this may be true if one simply looks at the distribution of Ting subscribers it is in fact beyond meaningless as a comparison. If average billing is $23 it has to be the case that most Ting subscribers pay less than the Verizon unlimited rate. It must also be true that Ting subscribers are getting a lot less than Verizon subscribers.

In fact, the claim is by any standards simply wrong.

If one had a single line on Ting, just 2500 minutes, 1000 texts and 4GB would cost $93.6 and that is more than what Verizon would cost for such usage. (I have a family member of Verizon who typically has that type of usage). In addition, Verizon has a better network, a nationwide chain of physical stores, and many device incentive programs.

So there seem to be two separate basic questions:

  1. how does one do an apples to apples comparison of cell services and is it even possible to do so?

Having thought about it quite a bit over the last year, I am inclined to think it simply cannot be done in a meaningful way.

  1. should the government take a position on the issue in either one of two ways:

a) demand that carriers who want to represent their service is cheaper than a competitor's provide a detailed worksheet that shows how the comparison is being made including adjustments for such things as network coverage, physical stores, device incentives, roaming, international use etc with a dollar value associated with each.

b) acknowledge that any such comparisons are purely subjective and warn consumers that claims by carriers are nothing more than marketing hype

It's interesting how these things vary country to country, often as a result of regulation.

In the UK, I always found it easy to do comparisons. I suspect it's a mixture of cultural factors and regulation.

Australia completely bemused me, and I am good at cutting through the marketing and figuring out the actual pricing model. At that time, most operators were allowed to represent their services in some really strange obfuscated way that made it hard to figure out the basics. It's hard to recall and explain, but imagine walking into a grocery store in Continental Europe where you're trying to figure out what bread to buy and they are selling varieties from all different European countries. Not only do you not know the differences between the breads, but they're priced in different currencies and you have no idea what the exchange rates are.

Your bill would be over $23 with 1gb and any voice minutes or texts at all at ting.

Basically they are bragging that their users don't use that much data I guess.

I don't see how it makes much sense at all comparing them. It is like blu comparing its $50 phone to an iphone.

I can't see anyone reading that and thinking they should choose ting over verizon unlimited if they were considering verizon unlimited in the first place.

It's marketing. And it's target market is those who use the big 4 directly paying over $23 a month.

I think the bread and currency analogy is a great way of looking at the cell phone comparison problem.:slight_smile:

A lot seems to depend on whether one is looking a just one line or several lines.

For the "average" $23 bill at Ting, a group of 4 could get the Sprint unlimited deal that is presently available and runs for a year.

So, in that case, it definitely seems more sensible to go with Sprint than for instance have 4 lines on Ting.

  1. how does one do an apples to apples comparison of cell services and is it even possible to do so?

Forums are pretty good at doing so. But most consumers do not have the time and/or inclination to read the forums. There are many phone forums and places that do discuss and compare the various phone plans that are at hand. However, like so many things that are mass consumed, most consumers do little research before purchasing, which may be the reason why there isn't something like a "Consumer Reports" for cell phone plans, which would at least attempt to be objective about such matters. But then again how many people would be willing to subscribe to such a source, or would utilize such a source even if offered for free. To me, it is the enormity of the nexus connecting marketing hype and ill-informed consumers that is the problem, as attested by the billions that are spent each year on totally unproven medical supplements that claim to "help" cure just about every malady known to mankind, which is but one of many examples. "What is to be done?" as Lenin once wrote. Free wheeling forums will have to suffice for now.:huh:

I love the topic. I've always said, as some of you know, that "stats lie". Of course, valid stats do not lie themselves, how they are used is often misleading though. There are so many ways to market a stat. I could say my own company has the cheapest broadband in the US and in some sense, one could argue it is a true statement, but, it's just marketing. But we see this all the time in so many industries, competing commercials (everyone seems to have the fastest network, how can that be?), politics of course, you name it. the trick is always reading through what was said, and, not said. The not said is usually the giveaway.

Of all the facets about living in this crazy world - for me - deceit in its many & varied forms is pretty much the worst of it.
Outright (or even involuntary...) lies, misdirection, mis- and dis-information and any/all other forms of such things just rankle me.

So here's a goofy bit of cooperation between the above and a real life situation many of us are all too familiar with:

While we remain very grateful that we were helped for 8+ months NOT to have a cellular bill to pay...
At the very same time, the quagmire of confusion which had to be endured to get the benefit of it, as well as its blurry ending was quite disturbing and even stressful at times.

How is that for a strange concatenation of things ?!?

Now, I must add - having moved forward into simpler, easier, less time-eating alternatives IS also a HUGE relief and also very appreciated.

So now it is 2/20/2017, and its home page, forum, etc. remain as they have been for a while now - no closure announcement on the main page - but no offerings either.

'The end' is supposed to come tomorrow sometime though - which makes me curious to see how that may look...?

Will go 404 ??
Will any synopsis of what actually, truly occurred ever be found anyplace ??

It would be grand if ultimately somewhere, somehow the actual, truthful telling of events might appear so as to settle the bafflement of those who used RP services and remain unsettled as to what really happened.

Of course learning the truth will not change what happened, nor is it likely to change anyone's response to any reappearance of that service under any guise - but it would still set some minds at ease, I think.

One of the challenges is simply assessing how meaningful a particular type of data set is.

Data on temperature are usually very precise. Snowfall data are not.

Budget deficits can be calculated precisely. GDP and unemployment rates cannot.

People compare yardage totals for Quarterbacks but of course, other variables besides the skill and execution of the Quarterback are involved.

Reports are issued regularly about how effective flu vaccines were in a given year but that is largely conjecture as so many non-measurable variables are involved.

The most challenging are statistics that involve comparisons: at Costo beef prices per lb can range from well under $10/lb for good quality to $100 for Waygu beef. There is no sensible way to say which is a better "value". All that can be said is what the prices actually are. One can compare the prices of Honda and Toyota cars in a given class and trim but there is no objective way to determine which is a better value.

But is the data on temperature actually precise? Maybe today. But maybe not. Now, I am not stating a position here so please don't take this the wrong way readers I am not making a political statement here that is not the point at all. However, I have discovered NOAA has gone back in time and adjusted previous measured temperatures in their data set from decades ago. One reference:

So, is the data really precise and if it is, why does one adjust the actual measurements! One tries to adjust for urban heat island effects today, but they are (somewhat) guesses and they try and re-visit and see if they can make it better over time.

Last winter (2015), a report and map came out that showed our winter here in my region of the country as being much warmer than normal. Yet, on all of the local weather stations, it was recorded as cooler than normal. Now, this can be a case of different measurement periods, based on something else, etc. One just wonders how it can be that is is warmer and colder at the same time! Who is correct, and based on what, why the difference. It's the same thing always about stats. What exactly do they measure, how, etc. Then there is the problem of reporting.

That being said, at least this winter, everyone agrees for my region it was much warmer. For once!

Until you realize there is no warmer or colder, the only option left is to flee the planet. :slight_smile:

I suppose my thinking about temperature readings was a bit naive.

On the assumption that a single person could have lived the entire time since the mercury thermometer was invented in 1714 and had taken a daily reading at exactly the same location and hour over this time it would be possible to have a (fairly) precise record of temperatures for that specific case. How any actual records of temperatures stack up against such a fictional example I have no idea.

Snow is an entirely different thing altogether. First of all, it is not possible to measure snowfall as such so what is measured is accumulation. That, in turn, is affected by the type of snow and the interval between measurements. In January 2016 there was a major snowstorm in the Washington DC area and official totals at Reagan National Airport were generally viewed as being problematic since the totals seemed way out of line with measurements from nearby recording stations.

Blood pressure is another example. Here in the US, the standard for prescribing medications for hypertension is a single or a few readings in a doctor's office. In Europe, it is common to base recommendations on readings from a device that is worn for at least 24 or more hours. Blood pressure levels, unlike things such as body temperature or cholesterol levels, can be significantly affected by the mere act of measuring it. White coat syndrome is a known issue.

What any of this means for "policy" decisions is anyone's guess.

Yes, and like blood pressure. they measure my wife's glucose level in the office and base a decision on that measurement (Yes, I understand there are other markers as well, but still, they speak of glucose). Never mind what it was the other 90 days or so since her last appt.

I am sure their measurement is correct, I mean somewhat sure, but, even if it is, is that the best methodology to use for glucose decisions? As a non medical person, I am skeptical!

In another thread here the question was raised as to what if anything the statistics shown on Social in relation to posting by a member might mean.

This is not an easy question to answer for a variety of reasons but let me go out on a limb and say they have no meaning that people will ever be able to agree upon.

The first thing that needs to be recognized is that what shows up on Social is not necessarily fully indicative of what a member posted or how members reacted to what was posted. That is because posts can be removed both with and without the consent of the poster. With that caveat here is a set of data relating to four members.

Just looking at the summary measures (average, and standard deviation) it seems pretty clear that nothing very useful can be said about these data. They are simply numbers, nothing more.

By contrast, the other image is indeed full of meaning and potentially useful. [HINT: skeingaggle might set you on the right track]

There are many ways to distort those stats as well. For example, say I know oldbooks 1, and, every time he posts, I, the oldbooks1 fanboy might always vote for his posts, and, he might always vote for my posts. Not as a method to distort stats, just because we just might feel inclined to do so. Get a little bit larger group of like minded folks, and, you might have the best stats out there. But they would certainly not state anything useful in that case as posited in the other thread. There were many such groups on the R+ forums.

There are many such scenarios. I would agree that the stats in this case are not a way to definitively say much. One might believe they do, but, this is typical stats.

I needed to look up something today from the US Census Bureau which is among the most reputable sources of official statistics in the world.

The issue I have looking at is household income for 2015.

In a blog, TV show, sound bite it would be correct to describe that number as being well below $70K, roughly $70k or more than $90k.

All three numbers are correct and all measure "average" household income.

One represents the income that the largest number of people have, one represents the number that reflects the dividing point between the lower and upper halves, and the top number represent what most people think of as "average" which is the mean. In this case, that measure gives a very distorted view because billionaire households are given the same weight as ones that are truly impoverished.

These numbers are not subjective but how to interpret them is.

For instance, if you were in charge of a government program to promote a more equal distribution of income which measure would you focus on?

There is no obvious answer but depending on which one you picked, the approach you took and the outcome would be very different than if you selected a different measure.

Statistics have shown that 86% of statistics are made up.

I think a more accurate number is 100% which is the same percentage as the number of made up words.:slight_smile:

If there are 100 people in the US, one of them makes exactly 1 billion dollars, and the other 99 make 0, one could say the average US income is 10 million dollars and promote how rich we are since that's the average income. I would tend to disagree with that assessment!