URGENT Help with Likert and Parameter Estimates

[#1]

I am running a Fit Model with where I'm testing a dependent variable (Likert responses) to several independent x variables (also Likert ). both sets are 1 - 7 responses from strongly disagree to strongly agree.  My issue is in the reporting.  How do I assess the overall estimate instead of just the individual matches looking at image below.

I'm trying to achieve output like shown in spss 

SPSS output

where variables like bargaining power over key suppliers are rolled up.  Or is it that I only need to use the one that matches what I was testing for (the one most influential as in those that marked 7?)???  Please help.  Or if I need to organize my data differently or do a different test any direction would be great.

[#2]

You set up your regression in JMP as a nominal logistic, so you are regressing the generalized logits against the linear predictor. That model means that if there are 7 levels in your response, then there will be 6 logits and 6 sets of parameters, one set for each logit. Your linear predictor is a combination of categorical factors except for Firm Size. That parameterization further expands the number of estimates within each set.

Can you use the Effect Likelihood Ratio Tests to assess the significance of the individual terms in the model? This table appears below the Parameter Estimates table.

[#3]

1. Thank you so much for responding I really appreciate your insight

Next, yes I can use the results below but am unsure what I would list as the estimate or if I need to list each one (or the most important ones).

I’m trying to use the same fields as a different study that is depicted in the second image. Do you know how I would find those results?

[#4 · Accepted solution]

I am not familiar with this study or the way that SPSS might parameterize the nominal logistic regression model. There are two possibilities that you might consider:

1. The published study that you refer to might have treated the categorical predictors as if they were continuous variables. There would then be a single estimate for each predictor.
2. Your response uses a Likert scale so you could the ordinal modeling type for Y. Ordinal logistic regression uses cumulative logits and you would have a single estimate for each predictor.

You might also include multiple correspondence analysis in your study to explore the associations among these categorical Y and X.