Page 1 of 1

Issues with transformed variables

Posted: Sun May 16, 2021 2:30 pm
by Ceecy
I'm currently running Jamovi to analyze the questionnaire I sent out. The responses in my research paper are based on the Likert scale with level 1(strongly disagree) to level 5 (Strongly agree). I have a group of 4 questions for each topic set. For example, there are 4 questions to "Brand image" and 4 questions to "Brand loyalty". And I'm intending to test the relationship between these 2 variables.

The first step I believe I should do is to add the values (Strongly agree = 5, etc) of these 4 questions, and then divide them by 4 to get a mean value. And then I use regression.

Based on the data I exported from Qualtrics, the Likert scale is identified as nominal, therefore, I was trying to create a transformed ordinal variable for each question. But my coding did not work. I attached a screenshot of it, would you please help me take a look to see what is the issue?

Also, are there any good tutorial videos or materials for Likert scale regression analysis in Jamovi? I tried to search but did not find any.
Your help would be much appreciated.

Re: Issues with transformed variables

Posted: Sun May 16, 2021 4:18 pm
by MAgojam
Hi, @Ceecy.

Before making suggestions or anything else, I draw your attention to the jamovi sheet.
Looking at your screenshot, it seems to me that your data in jamovi got dirty.
The first two rows should not be present (imported) in jamovi, possibly resulting in header lines or something else in the data file you used.
Make sure you have a sheet with correct data before continuing.

Cheers,
Maurizio

Re: Issues with transformed variables

Posted: Mon May 17, 2021 7:10 am
by jonathon
maurizio is right. take a look at wake's presentation on importing data from qualtrics into jamovi:

https://docs.google.com/presentation/d/e/2PACX-1vSl5QgAZLndsN6XM0iPZJo4PItti92b8cmjpLOuyjmMDbqPb35Xf9zvX5Sf2Nspv2MWTd0rC1EFpWRJ/pub?start=false&loop=false&delayms=10000

for your recode rule, you want quotation marks, like this:

if $source == 'Somewhat agree'

cheers

jonathon

Re: Issues with transformed variables

Posted: Sun May 30, 2021 7:01 am
by Ceecy
OMG THANK YOU SO MUCH FOR POINTING THAT OUT. I just simply exported as csv and didn't even bother... But i will do it the right way now. Thank you guys!!

Re: Issues with transformed variables

Posted: Sun May 30, 2021 10:27 am
by Ceecy
I have to thank you guys again. I re-imported the data and cleaned the data, now the results look totally different! Like completely different ... I have interesting results now, significant effects and stuff. Thank you Maurizio and Jonathon.

Have a great day!
Ceecy

Re: Issues with transformed variables

Posted: Sat May 14, 2022 12:10 am
by janmarielle.pepito
Hello! May I know where you found some tutorial videos or materials for Likert scale regression analysis in Jamovi? I need it for the data analysis of our research. Thank you!