Hi sorry for lack of clarity the components from the likert scale are cost, time, quality, client satisfaction, sustainability, scope and risk which were all ranked for iimportance within a project out of 1 strongly disagree to 6 strongly agree. What I want to find out ideally is to determine the most important components needed to evaluate a project and whether experience is correlated to old iron triangle success criteria i.e. Cost, time and quality. so more experienced project managers prefer the "iron triangle model compared to younger less experienced who want newer success criteria which is not in a model.
With the recommendations to formulate a more current updated model e.g. Cost, risk and client satisfaction.
There are some useful videos in YouTube that explain how to analyse Likert scale through SPSS. So, you just need to search in Youtube if you prefer a visual aid!
For example:
https://www.youtube.com/watch?v=nyzmHke8DiE
https://www.youtube.com/watch?v=x684kPgiay0
https://www.youtube.com/watch?v=QuEmL2mGlGI
Also, Please check the attached file which is a brief handout.
You can use any SPSS version to perform your corelational analysis. Just average out your likert scale items into single composite variable and perform corelational analysis. Any simple video of SPSS regarding correlation can help you out
As Likert Scale provides data on ordinal scale, you can go for non-parametric test as mentioned by Dr. Frimpong.The basic assumptions in Parametric tests like correlation, regression, factor analysis etc are that data set should follow normal distribution and data should be in Interval or Ratio scale. But, in social sciences, we do sometimes comprise with these basic assumptions and even with ordinal scale we do parametric test. But, once again the results are comprised and may not be true to the fact.
Using Likert scale means your data are labeled scale on the definition of your variables in SPSS. You may use corss-tabulations with descriptive statistics to characterize the relations and study Pearson R. You can also use Bivariate analysis to study each pair of the variables you need to correlate. Perform this analysis between what will be your dependent variable, so that you may later on choose your Regression analysis to do multivariable analysis, and son on. Your can perform reliability testing later on.
so potentially I can use 'experience' as my dependent variable with cost, time, quality and client satisfaction as my independent variable components that are being rated 1-6. Which are being measured in importance for measures of project success e.g one respondent could say cost (4), time 2, client satisfaction 6 and quality 4 with the aim of building a formulated model which is relevant to new projects
From then surely ordinal regression is preferable for a study of this kind?
Recall that you have a large sample of respondents, when you crosstab first to study relations, you may choose the statistics to visually assess relationships first, the use Pearson R.
Your respondents' responses have a variety of numbers, SPSS will work on the averages when using regression, and to formulate your final equation you have to analyze the Beta coefficients which are statistically valid. You may use forward step regression or other.
All the answers were helpful, but still the fact that with «social sciences, we do sometimes comprise with these basic assumptions and even with ordinal scale we do parametric test» is confusing.