Comparing the academic performance between two classes

lcy

New Member
#1
Compare the academic performance between two classes

Dear all,
I kindly need your help. Here is my background.
I am a secondary school teacher.
I taught two different classes in the same form last year. Two different teaching method were used on these classes. In order to compare their academic performance under the various teaching methods, what should I do? Here are some questions which I am thinking about.

1.
If the scores of final exam of two classes are used as the only
dependent variable, the two teaching methods are treated as the independent variables.
(a) As all scores of two classes are well collected, could I treat I got the population mean, population variance….
(b) As I got the population information, could I simply compare their mean scores to justify their performance?

2. I wonder, could I need other statistics such as t-test or ANOVA to do the comparison?
(a) If the answer is “yes”, what are the rationale?
(b) It seems t-test is not suitable, as we got the population information instead of sample information. Does it make sense?
(c) How about ANOVA? Is it better than t-test?

3. In most of secondary assessments, their distributions are skewed (negative or positive. For skewed distributions, comparing means, t-test, ANOVA are still valid? IF no, what should I do?

Thanks in advance. Best Regards,
LCY
 

Karabiner

TS Contributor
#2
Two different teaching method were used on these classes.
How large were the classes? Did each class receive only 1 of these teaching methods?
In order to compare their academic performance under the various teaching methods, what should I do?
Is your intention to compare the 2 teaching methods, so that you could determine which is perhaps better?
a) As all scores of two classes are well collected, could I treat I got the population mean, population variance….
No, AFAICS.

With kind regards

K.
 

lcy

New Member
#3
"How large were the classes? Did each class receive only 1 of these teaching methods?"
The size of the classes are similar. Yes, each class received one teaching method only.

"Is your intention to compare the 2 teaching methods, so that you could determine which is perhaps better?"
In fact, comparing the teaching methods is one of my intentions but not the only one.
Could I say that if comparing teaching methods is my goal, then the two classes are only the samples instead of population under these teaching methods.

Could I ask a follow up question? If I wanna compare the performance between the two classes (actually, teachers always compare performance among classes). Then the classes will be the population. Does it make sense?
 

lcy

New Member
#4
a) As all scores of two classes are well collected, could I treat I got the population mean, population variance….
No, AFAICS.


Could you tell me why the answer is No?
 

Karabiner

TS Contributor
#5
I cannot see why the samples should be treated as a populations, or what the use of this would be. If you want to find out whether group differences are beyond what would be expected by mere chance, then you'll have to treat the samples as samples.

With kind regards

Karabiner
 
#6
a) As all scores of two classes are well collected, could I treat I got the population mean, population variance….
No, AFAICS.


Could you tell me why the answer is No?
If you have no interest in extending these teaching methods, or one of them, to different groups of students who were not already taught, then you might be able to consider your current group of students as the population since that is the entirety of the pertinent group.

I think if you are looking to apply the "better" teaching method to future cohorts of students you should treat the two classes as a sample from the population of all students you will teach. This could also be said if the teaching method will be applied to other students who you won't teach personally but will be taught by a colleague.

Simply, if the students who already participated are not the entire group of interest-- that is, they are a subset of a population of students who may receive teaching methods, then you should treat them as a sample as Karabiner is suggesting.
 

lcy

New Member
#7
"If you have no interest in extending these teaching methods, or one of them, to different groups of students who were not already taught, then you might be able to consider your current group of students as the population since that is the entirety of the pertinent group."

Dear Ondansetron ,
Thanks for your reply.
If I consider the group of students as the population, could I simply use their mean scores to compare the performance? Let say the mean scores of Class A is higher the Class B, then the performance of class A is better then class B. Actually, using comparing mean is the general practice in the education field, but I wonder there is any better statistical tools.