How to Convert a 5-Point Likert Scale into a 7-Point Likert Scale?

rd75

New Member
#1
Good afternoon,

First, if I have missed this being posted before, I apologize, but I could not find it.

I am having a bit of a nightmare due to a mistake I made on my survey for my dissertation. I used various questions involving a Likert scale, but used both 5-point scales and 7-point scales. How can I convert the 5-point scale into a 7-point scale using SPSS (or excel if its easier!) so I can analyse them?

Second question, The questions on the survey relate to risk perception, using the variables of perceived likelihood and training/knowledge. What would the best form of analysis be on SPSS? I am thinking Pearson's correlation, but would appreciate any advise available from an experienced user.

Many thanks in advance for your help!!
 

Lazar

Phineas Packard
#2
If you want them on a common metric just z-standadize them.

As for your second question more info is needed. What question do you want to answer?
 

rd75

New Member
#3
If you want them on a common metric just z-standadize them.

As for your second question more info is needed. What question do you want to answer?
Hi Lazar, thanks for the reply. Unfortunately I'm not the best when it comes to statistics and new to SPSS, what is z-standardizing and how do I go about doing that?

The second question: I'm looking to prove that when the level in one variable goes up (training or knowledge), the other goes down (perceived risk). I will do this by comparing the two likert scale responses (once I've converted the 5 point into a 7 point scale!).
 

Lazar

Phineas Packard
#4
http://www.ats.ucla.edu/stat/spss/faq/standardize.htm Standadize changes the metric of a variable such that it has a mean of zero and a standard deviation of 1.

It seems that you want to explore the predictive effect of training and knowledge (as two different variables) on percieved risk. As such I think you might as well use regression.

In passing you do not need to change the likert scale responses. There is not a problem exploring the association of two variables that are on a different metric.
 

rd75

New Member
#5
Ok that will make it a lot easier if I do not need to convert the scales. I need to do two different analysis, first as per the above, predict the effect of one variable at a time (training, then knowledge) on perceived risk. Secondly I need to predict the effect of both together on perceived risk.

Would regression still be the best method? If so, what would be the correct tool? Analyze - Regression - Ordinal?

I apologise again if these are obvious questions, I'm new to statistics and SPSS and have limited understanding of the software!

Thanks again,