Hi Marzena,

You might try a means analysis. For your Likert categories, first group them. For instance, the first three (1, 2 and 3) may represent an unacceptable perception, the middle three (4, 5, and 6) may represent a neutral percetpion, and the last three (7, 8, and 9) may represent an acceptable perception. Next, establish some ranges for your respondents. Ranges may be subjective. For the scenario, the range of 1.00 through 3.99 would represent unacceptable means, 4.00 through 6.99 would represent neutral means, and 7.00-9.00 would represent acceptable means.

I've not seen your Likert scale. But, you can use any categories you desire for your instrument instead of acceptable, neutral, and unacceptable. Another concept may be 1 and 2 for a certain category, 3 and 4 for a different category, 5 for neutral, 6 and 7 for another category in the opposite direction, and 8 and 9 for the latter end of the direction. You can set and specify categories and corresponding ranges just about any way you like for your means analysis.

Just my thoughts ...

Have a great day!

--Adrian

Dear Ajit Kumar Roy, thank you for your answer and the explanation.

I have tried using chi-sq. test as you suggested to check if there is a difference between two given groups and their ratings for accentedness/comprehensibility. However, I am getting a statistically significant difference in the baseline condition. It may be simply because of the sample being small but I was wondering if there is any way of normalizing the Likert scale ratings? Or should I treat those solemnly as nominal data? I would really appreciate your input.

Hi Marzena,

You might try a means analysis. For your Likert categories, first group them. For instance, the first three (1, 2 and 3) may represent an unacceptable perception, the middle three (4, 5, and 6) may represent a neutral percetpion, and the last three (7, 8, and 9) may represent an acceptable perception. Next, establish some ranges for your respondents. Ranges may be subjective. For the scenario, the range of 1.00 through 3.99 would represent unacceptable means, 4.00 through 6.99 would represent neutral means, and 7.00-9.00 would represent acceptable means.

I've not seen your Likert scale. But, you can use any categories you desire for your instrument instead of acceptable, neutral, and unacceptable. Another concept may be 1 and 2 for a certain category, 3 and 4 for a different category, 5 for neutral, 6 and 7 for another category in the opposite direction, and 8 and 9 for the latter end of the direction. You can set and specify categories and corresponding ranges just about any way you like for your means analysis.

Just my thoughts ...

Have a great day!

--Adrian

Hi Ajit and Adrian,

Thank you for explaining everything in details. This is really helpful. I will try to do both and see how it works with my data.