Celebrate Excellence in Education: Nominate Outstanding Educators by April 15!
Found this content helpful? Log in or sign up to leave a like!
I am transitioning from another LMS to Canvas. First of all, I LOVE that canvas has built-in quiz question formulas for mean, median, combinations, permutations, and more. Can you also include an option for standard deviation? That would make my life so much easier when I write quiz questions for students to calculate the standard deviation of a set of numbers or probability distribution table values. Could this request be something the canvas programmer gurus could look into?
Standard deviation = √[(Σx2)/N] , where Σx2 is the summation of the square of deviation of all observations from Mean. ∴ The standard deviation of given data set of weights is √163.
Good morning, @GinaLeffers ...
Welcome to the Canvas Community! Your question sounds more like a Feature Idea submission to enhance Canvas. Full disclosure, I am a former Canvas administrator, and I didn't teach courses in Canvas, so I don't have lots of experience with the different formula options in quizzes. However, I did a quick search of existing Feature Ideas, and I came across this one (there may be others, too):
standard deviation in test scores
I'm not sure if this existing Feature Idea is exactly what you are asking for or not...but I thought I'd share. If you want to search for other existing Feature Ideas, you can go here, and type in keywords to search for:
Idea Conversations - Instructure Community (canvaslms.com)
However, just so you are aware, the Feature Idea process is temporary on hold until the beginning of February while the process undergoes some clean-up and re-vamping. At this time, nobody can submit new Feature Ideas, comment on existing Feature Ideas, or give star ratings to those ideas. Here are some links that you might want to look through:
And...since you are a New Member here in the Community, I'll provide you with this link as well. Keep in mind that you need to "rank up" (similar to "leveling up" in video games) in order to submit new Feature Ideas.
I hope this information will be of some help to you. Please let Community members know if you have any questions. Thanks!
Any chance I've leveled up yet in canvas? I would love to see a standard deviation option in the formula quiz editor!
@GinaLeffers ...
Yes...you start out at a Community Novice...and then the next level is Community Member (what you're currently at). You can read more about the ranking levels here: How do I rank up in the Instructure Community? - Instructure Community - 522020 (canvaslms.com).
Leveling up in the Community doesn't have anything to do with whether a standard deviation function will be added or not.
I've been using Canvas since 2012 and I really don't see standard deviation being added. Support for mathematics and the sciences is fairly lacking. There isn't even a function for entering a list of numbers -- other than to reverse() or sort() it.
The list of statistical functions is basic. Functions for finding quartiles or percentiles are missing (granted there are about 7 ways to find a quartile, so maybe it's better that it's not there), although you could sort() a list and then find the number at() a specific position to create your own function.
You're also limited with the functionality of the random number generator in formula questions in Canvas. There is no way to say, for example, "give me a list of 5 random integers between -20 and 20 without repetition."
If you want to use the formula question type to find the standard deviation, you will have to program it yourself.
Thankfully, both Classic Quizzes and New Quizzes support multiple steps in formulas. This is documented for Classic Quizzes, but not so clearly for New Quizzes. The documentation says "write the formula or formulas". If someone was starting with New Quizzes without using Classic Quizzes, they might not understand how to do that.
Here's how you could create a question with new quizzes.
The Question could look like this:
Find the sample standard deviation of these random numbers with no context:
`x1`, `x2`, `x3`, `x4`, `x5`
The Formula Definition would look like this (all in the same box).
m=mean(x1,x2,x3,x4,x5)
v=(x1-m)^2+(x2-m)^2+(x3-m)^2+(x4-m)^2+(x5-m)^2
s=sqrt(v/4)
If you want the population standard deviation, then change v/4 to v/5.
Extending that to larger sample sizes is just a matter of adding more variables.
If you want to do the whole thing on a single line without storing intermediate results to other variables, then you have to use some of those shortcuts they taught me 35+ years ago in stats class that they then decided we shouldn't teach students how to do by hand anymore because the don't help them understand the concepts.
You know, those things like \(\displaystyle s=\sqrt{\frac{\sum{x^2}-n{\bar x}^2}{n-1}}\)
sqrt((x1^2+x2^2+x3^2+x4^2+x5^2-5*mean(x1,x2,x3,x4,x5)^2)/4)
The single line formula is easy to get a parentheses in the wrong spot, while the multiline formula takes more typing.
Another approach is to harness the power of Excel to write numeric questions instead of formula questions. Create the text and answer in Excel and then transfer them to Canvas.
For example, in cells A2:E2, I could have =RANDBETWEEN(-20,20)
to generate five random numbers between -20 and 20, inclusive.
In cell F2, I have =STDEV.S(A2:E2)
which is the answer to the question. I will copy/paste it into the answer box in Canvas.
In cell G2, I have ="Find the sample standard deviation of these numbers: "&TEXTJOIN(", ",TRUE,A2:E2)
. I will copy/paste this into the question box in Canvas.
Every time I press F9, it generates a new question.
In Canvas, I create a bunch of questions and then put them into a question group (Classic Quizzes) or question bank (New Quizzes) that holds several of these questions.
If you have software such as Respondus 4.0, you can import an Excel file and create questions from that.
The Excel approach allows a lot of flexibility. Basically, anything that you can calculate in Excel can be converted into a question. You could even find the p-value for an Anderson-Darling test for normality, but there is no way I would want to try that with just Canvas.
To participate in the Instructure Community, you need to sign up or log in:
Sign In