How to convert ordinal into interval?

#1
Hi all, i'm a new member here..

I have a problem, right now i'm on my undergraduate thesis, my thesis is about the relation between internal audit (X) and internal control report (Y). For X, i need to go on questionnaires, and for Y, i just need to compare the report with the guidelines

Since my questionnaires is based on likert scale, so the data will be ordinal data, anda for Y, the data will be on interval scale. Since the scale is different, i can't use them to test the spearman correlation between them

My question is, are there any method to change my ordinal data into interval? I've heard about methods of successive interval (MSI), but i can't find any guides about that on google. Could you help me explainning what do i have to do in order to change the scale from ordinal to be interval?

Thank you very much for your help..
 
E

elnaz

Guest
#2
Hello

Dont mistake
you should not change my ordinal data into interval while you should transform interval data
into ordinal data and then you should do spearman correlation between them.
know that spearman correlation is for ordianl data but pearson correlation is for interval data then you must convert interval data into ordinal data
and then you obtain spearman correlation between them.
i hope you be succeed
Best regards
Elnaz
 
#3
elnaz said:
Hello

Dont mistake
you should not change my ordinal data into interval while you should transform interval data
into ordinal data and then you should do spearman correlation between them.
know that spearman correlation is for ordianl data but pearson correlation is for interval data then you must convert interval data into ordinal data
and then you obtain spearman correlation between them.
i hope you be succeed
Best regards
Elnaz
oops.. my bad.. thanks for remind me.. what i mean is pearson correlation, not spearman correlation, the spearman correlation is only used to test the validity of my X variable.. since i'm not majoring on statistic, i don't have many resource about statistic, i even couldn't find any clue about how to change ordinal into interval on google.. could you help me explaining how to convert my ordinal scale into interval so i can do pearson correlation between them? thank you very much
 
E

elnaz

Guest
#4
Hello

i said you , this work is wrong, why do you insist that to change ordinal data into interval data , i said ,you should not do that ,you must change interval data into ordinal data , and then use spearman correlation , if i was instead of you , i dont do that at all .i advice you dont follow it , please accept my advice but as you wish.
i hope you will be succeed.
 
#5
I think you might be using the wrong terms.

Do you want to change:
1,2,3,4,5,6,7,8,9 -> A,B,C (A=1-3,B=4-6,C=7-9)?

I would assume that 1-3 was something like "Bad News indeed"
4-6 would be "who cares"
7-9 would be "better than Cut Bread"

I have no idea of correlations, but you should be able to do the above with no loss of meaning, just granularity. It would probably be done in cases where the dataset is somewhat limited in numbers and the power to detect differences is just not there.

I would actually call that a transform, but that's probably a wrong term.