• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Misleading graphs

quadraginta said:
But his wasn't

Just taking an opportunity for a teaching moment?

That usually makes more sense when someone has made a mistake. Correcting someone who didn't doesn't have the same sort of effect.
Click to expand...

This is getting meta.
 
quadraginta said:
But his wasn't

Just taking an opportunity for a teaching moment?

That usually makes more sense when someone has made a mistake. Correcting someone who didn't doesn't have the same sort of effect.
Click to expand...
The highlighted line was perfectly explanatory. I was merely pointing out an omission in jt512's lesson and possible consequences if somebody else followed his instructions to the letter.
 
jt512 said:
however, there is a rule that you can only post such a "hot link" if the hosting website allows it.
Click to expand...


And how would you find out if they do? Try to look for that site's policies or whatever? Sounds a bit ... involved?
 
Minoosh said:
(...)This is the part I'm slightly conflicted about. Some students thrive on ambiguity, but for others these discussions might amount to cognitive overload.

They aren't all at the same level(...)
Click to expand...


Sure, that’s a tough call to take. Especially where children are concerned.

My only experience with teaching is delivering the odd training module at work, but I suppose that is far easier, since the “students” there are all full-grown adults who (mostly) know their minds. Sure, not everyone is of the same level there either, and this conflict that you talk of arises there also, but that isn’t really difficult to address, in my experience. You simply bring out issues extraneous to the core material (if and when such issues come up), and then leave it to the group to take it further or not. And if, as sometimes happens, an off-topic discussion veers of to unexpected lengths, you simply cut it short after a while citing time constraints. Sometimes if they’re interested some people come across and discuss specific areas separately with you as well, later on. I don’t know, this is probably not a problem when you’re dealing with adults, especially when they comprise a fairly homogeneous group.

But of course you’re teaching children. I wouldn’t presume to guess at what might be the best way to deal with this sort of thing with kids. But then, do you necessarily have to treat them differently, in this respect, just because they’re children? Do children behave so very differently from how adults do? (I don’t know the answer, no clue at all! Just thinking aloud. If you don’t know the answer either, perhaps you could just try it once and see if it works, taking a more open approach I mean? You can always stop doing it if doesn’t seem to be working.)
 
psionl0 said:
1. You used imgw tags but your instruction was for img tags only. 2. Somebody using img tags instead of imgw tags could easily end up posting an image that is disruptively large. 3. Size matters. :p
Click to expand...


You're trying to have us believe that smaller is better? Hmmm! :)
 
Chanakya said:
jt512 said:
however, there is a rule that you can only post such a "hot link" if the hosting website allows it.
Click to expand...


And how would you find out if they do? Try to look for that site's policies or whatever? Sounds a bit ... involved?
Click to expand...


Not so much.

To start with, if you're not sure, don't hotlink. You can store the image to a hosting site such as Imgur, etc, which definitely does allow hotlinking. Or store it on your own system and upload it to the board from there. (Note: any copyright issues are on you.)

Some site are quite straightforward about their sharing and hotlinking policies.

If you check out the "Help" pull-down menu at the lower right corner of every page banner you will find a link to the "Forum Help & Member Support" page. The first thread listed in the stickies there is "List of sites that allow hotlinking". This is not intended to be a final or comprehensive list, but is a list of sites which this board has so far verified as being hotlink approved.

If you find it difficult to determine whether some site allows hotlinking you can always ask them. Send them an email. Then get their site added to the list.

Just refrain from hotlinking if you don't know for sure. It doesn't prevent you from getting an image posted. It just adds a step or two. If that's too much trouble then the image probably wasn't all that important anyway.

You can always provide a link to an image. If posting the image itself is too much of a bother.
 
Last edited:
GodMark2 said:
I see a missed teachable moment here. Instead of trying to tell them why it's wrong or misleading, show them. Make some of your own graphs of the same data that completely mislead: (x,log(y)), (x, exp(y)), (x, -y), y-axis starts at -200, etc. Then have them describe the graphs they see, and then you can discuss why some are misleading, and in what ways.
Click to expand...

This.

For some ideas, I'd recommend which should be required reading for everyone at some point I their lives.
https://www.amazon.ca/How-Lie-Statistics-Darrell-Huff/dp/0393310728
 
Last edited:
quadraginta said:
Not so much.

To start with, if you're not sure, don't hotlink. You can store the image to a hosting site such as Imgur, etc, which definitely does allow hotlinking. Or store it on your own system and upload it to the board from there. (Note: any copyright issues are on you.)

Some site are quite straightforward about their sharing and hotlinking policies.

If you check out the "Help" pull-down menu at the lower right corner of every page banner you will find a link to the "Forum Help & Member Support" page. The first thread listed in the stickies there is "List of sites that allow hotlinking". This is not intended to be a final or comprehensive list, but is a list of sites which this board has so far verified as being hotlink approved.

If you find it difficult to determine whether some site allows hotlinking you can always ask them. Send them an email. Then get their site added to the list.

Just refrain from hotlinking if you don't know for sure. It doesn't prevent you from getting an image posted. It just adds a step or two. If that's too much trouble then the image probably wasn't all that important anyway.

You can always provide a link to an image. If posting the image itself is too much of a bother.
Click to expand...


Thanks for letting me know about that list of sites which permit “hotlinking”, quadraginta. I hadn’t known about this. Although your link to those sites does not seem to be working. I found that list in this thread, the first thread in the FH&MS section that you'd pointed me to.

Incidentally, I was just asking about all this because I was generally curious, that’s all. The graph I had originally wanted to copy was directly from my PC (where I'd copied it from my mail in order to edit it slightly before sticking it here). It was part of a current report that my team had prepared recently, that seemed apt here. No big deal, easy enough to conceive of any number of such examples. Which is why I never attached it at all, even after figuring out how to. (Given the realization, since I had time to think about it, that I oughtn’t perhaps be posting confidential material here at all, not even a graph with the legends and reference points removed.) You’re right, in my case the image wasn’t all that important anyway.
 
I will never not love this graph
HZe4vKym.png
 
lylfyl said:
I will never not love this graph
[qimg]http://i.imgur.com/HZe4vKym.png[/qimg]
Click to expand...

**giggle** :thumbsup:

Also, where the "time" is on the x-axis it can be fun to set 0 to the time of the Big Bang. It can make for a very short area of actual data up and over to the right.
 
My biggest issue with that problem is that it shouldn't be illustrated with a linear plot. How do you define a tenth of a topping? Such a scenario should use a bar chart.
 
I regularly use offset crossing points on graphs. For instance I put 6 dataloggers in an incubator to measure temperature and humidity, logging every 10 seconds over a weekend. The specs require the incubator to operate at 37 +/- 1 deg C temperature and 85% +/- 5% relative humidity.

The plot I produce uses scales that allow a minimum clear space to the left and right of the readings and below the minimum reading and above the maximum reading with sensible major units (6hrs in time, 0.2 deg C temperature and 2% RH typically). That way everybody can see when the measurements started, when they ended and what the variation was.

If I started the scales at 0 deg C and 0% RH the detail of the variation would be lost in an almost straight line at the top of the graph for both traces.

From setting up graphs involving time plots in Excel the number 0.0416666666 is indelibly burned into my brain!
 
I often work with graphs that start at 220 and finish at 260 over time periods from seconds to weeks- starting at 0 would actually be misleading in these graphs (of mains voltages and periods of over or under voltage outside of the allowed standards) as the scale would be so compressed that the actual over/under volts wouldn't even be visible (unless you blow up the graph to the stage they are- ie visually starting it at 220v instead of simply truncating all voltages under 220....)

So starting my graphs @ 0,0 would actually be misleading!!!
Giving rise to the impression there hadnt been over/under voltage situations when there had been (but had simply not been shown due to the small variation at the top of the scale not being readily visible)
 
I vote that all graphs that include time need to start with the Big bang. All temperatures must be in Kelvin, starting with zero.

Anything else would be cheating.
 
Lukraak_Sisser said:
As others have said, as long as you label it correctly a graph can start at any coordinates you like. One of the main problems with students is that they don't label anything at all. It's often worth a few lessons to focus on that so that they can learn to differentiate between graphs ment to explain and graphs ment to mislead. At least we might see less of them here on the forum in a few years defending things like the electric universe of Mill's nonsense which all work with misleading graphs.
Click to expand...

This. A thousand times this.

It doesn't help that the most-used (and often only known) software to make graphs (*cough*Excel*cough*) makes it ridiculously easy to get mis-, non-labeled graphs.
 
Minoosh said:
I'm working as a long-term substitute for 3 math classes. Yesterday I gave students a very basic graphing problem. A cheese pizza costs $5 ...
Click to expand...

I'm really surprised (not being facetious) ... from what little I've seen on line I thought there was a set curriculum, with all the lessons planned out ... so you just taught them to pass the standardized tests they got at the end of the year.

How much leeway do you have in the curriculum?
 
jt512 said:
By definition of what? Can you cite a source?
Click to expand...

Definition of a Cartesian coordinate system. Depending on how you define it, the x-axis is either the set of points with y coordinate equal to zero (and likewise for the y-axis), or coordinates get defined as signed distances from the x-axis and y-axis. In both cases the x-axis and y-axis intersect at coordinates (0,0).

That doesn't mean anything about how you draw a graph, of course, you can start your graph at, say, y = 500 if you want. But that horizontal line you draw at y = 500 is not the x-axis, strictly speaking.
 
Last edited:
You must log in or register to reply here.

Back
Top Bottom