I did a small user study with 20 participants. The participants were asked to rate two scenarios, how often they use their smartphones to listen to songs and how often they use their smartphones to use social media on a 5 point scale from not at all to always. Now after the study, for the analysis of Likert scale data, Is it okay to use the following:

  • FIRST CASE: Points 1 to 3 are regarded as not at all or near to not at all.
  • SECOND CASE: Points 4 to 5 are regarded as always or near to always.

Am I going in the right direction? Then, after this, I may sum up the frequencies found in the first case and the second case to find that there was more "not at all" or there was more "always".


Frequency of smartphone usage:

1   3   
2   10  
3   3   
4   2   
5   2   

Total 20

The above table shows a frequency table of responses for 20 users on a 5 point Likert scale. The Likert scale is of the form: 1 = not at all 5 = always

Now, according to the two cases mentioned above, if I conclude that 16 (10+3+3) users answered not at all or near to not at all, whereas 9 (2+2) users answered always or near to always. Will this conclusion be correct?

Any help or ideas are appreciated as I am very new to this.


  • 1
    $\begingroup$ By transforming your scale to binary format you loose a lot of information (a vote of 7 would be judged the same as a vote of 4). What questions do you want to answer with your study? $\endgroup$ – TinglTanglBob Jul 11 '18 at 0:13
  • $\begingroup$ @TinglTanglBob I updated the question with a sample query. Am I going in the correct direction? $\endgroup$ – Jishan Jul 11 '18 at 0:42
  • 1
    $\begingroup$ As long as you give a definition what "not at all or near to not at all" and "always or near to always" mean, there would be nothing "wrong" with doing so. But, even with giving a definition, results might be missunderstood. Imagine, in your example, if it were 10 to vote 3 and 3 to vote 2. Your statement would still be the same: 16 people fall in the group of "not at all or near to not at all" (which means they have voted something between 1 and 3 on a 5 point scale). This might transform into 16 people do not use a smartphone at all, while the trueth is, that 10 voted 3. $\endgroup$ – TinglTanglBob Jul 11 '18 at 11:50
  • 1
    $\begingroup$ But in this case i do not see any benefit of using a binary scale over the 5/7-point Likertscale. Is there any need, restriction or something (regarding general conditions) to do so? $\endgroup$ – TinglTanglBob Jul 11 '18 at 11:56
  • $\begingroup$ @TinglTanglBob yes, I understand that. I am still confused. What do you think will be the most prudent way to represent the percentage of participants answers in my question above. Any ideas $\endgroup$ – Jishan Jul 11 '18 at 16:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.