Cronbach's alpha assumes that all the items measure a single underlying concept, but it does not assess that assumption. That is why Peter Samuels recommends Exploratory Factor Analysis, because EFA shows how many separate scales may be necessary.

Because questions on this topic come up often here, I have collected a set of resources on the topic:

https://www.researchgate.net/post/How_can_one_help_people_create_simple_scales_from_Likert-scored_items

It depends slightly on what type of scale you are using. If you are using a scale that has already been validated, there should be instructions in the original study to tell you how to score the scale. If you are creating your own scale, one place to start is to check reliability (e.g., Chronbach's Alpha) and then create a new variable that is the total scale score. Before doing either of these things, it is important to note if any of the items ought to be reverse scored.

If there are items that need to be reverse scored, the first step is to create new variables with those items reverse scored. To do so, go to: Transform->Recode into Different Variables and then input the information to reverse score. I usually call my reverse-scored variables the same thing as the original variable with an "r" in front of it. So Variable1 is recoded into "rVariable1."

Once you have taken care of reverse-scored items, you can check reliability with Chronbach's alpha. If you are satisfied with that result, you can create a total score for the scale. To do so, go to: Transform->Compute variable. Enter the new variable name, then double-click on each variable that is a part of the scale (using reverse-scored variables if there are any and NOT the original variable if it is reverse scored) with a + sign between each. When you click "okay," SPSS will create a new variable with the total scale score. You can then use that total score for analyses.

This can also be done through the use of some simple syntax.

Thanks Amy for your help. The scale is a likert scale and while i am carrying out a Mann-Whitney test , I am not sure if i have to select all the scale items  in the test field or is it appropriate to collect few items that will answer my Hypothesis? I have 14 questions and carrying out a test such as Mann Whitney test , etc. do i have to select questions that are more closely related to the hypothesis?

Dear Sunny,

I recommend that you use an exploratory factor analysis rather than guessing which items to combine together and use as it will handle the process systematically. Have a look at my article. Doing Cronbach's alpha is a much later stage and is more for confirmation.

Thanks Peter. 

 Hi Peter, which of your article are you referring me to? Thanks

Cronbach's alpha assumes that all the items measure a single underlying concept, but it does not assess that assumption. That is why Peter Samuels recommends Exploratory Factor Analysis, because EFA shows how many separate scales may be necessary.

Because questions on this topic come up often here, I have collected a set of resources on the topic:

https://www.researchgate.net/post/How_can_one_help_people_create_simple_scales_from_Likert-scored_items

Thanks David.

The sample size I have to carry out exploratory factor analysis is 60 with 27 likely scale items, but I have read on several literatures that the sample size should be at

Least 300 samples. I am a bit confused on what to do, as I am pretty new to this . The likert scale items I have are 27 items with three sections, do I have to carry out three separate exploratory factor analysis on each section to enable me to have smaller items to carry out tests to answer my hypothesis? Please I do need urgent help, as the deadline for my assignment is only two weeks and I haven't done much. Thanks for your help in advance.

No doubt sample size is important in EFA, however,  in one of the most cited study titled "Best Practices in Exploratory Factor Analysis: Four Recommendations for Getting the Most From Your Analysis",  it was stated that they found studies where items to Subject ratio was 2:1. So I would suggest you to give it a shot. EFA will be conducted on whole 27 item, as your objective is to identify the variables on which these 27 items will going to map.

Did you trying finding videos on EFA in YouTube?  I found this medium very helpful in learning EFA not long ago. For example have a look on this one: https://youtu.be/-6tw7ebr860

Good luck!

Thanks Mr Ali, I have viewed several videos on YouTube, still is still confusing on what to do . I will check out the YouTube video link you attached. Thanks for your help.