My reply yesterday might not have gotten through, so I'm posting it here.
There is a case study for ordinal regression available via Help > Topics. I think you would find it helpful if you haven't already read it. Here are a few quotes from it (emphasis added)
The first thing you see in the output is a warning about cells with zero frequencies. The reason this warning comes up is that the model includes continuous covariates. Certain fit statistics for the model depend on aggregating the data based on unique predictor and outcome value patterns. For instance, all cases where the applicants have current payments on debt, one other credit at the bank, own their home, have no other installment debts, are 49 years old and are seeking a 12-month loan are combined to form a cell. However, because Duration in months and Age in years are both continuous, most cases have unique values for those variables. This results in a very large table with many empty cells, which makes it difficult to interpret some of the fit statistics. You have to be careful in evaluating this model, particularly when looking at chi-square-based fit statistics.
[goodness of fit] statistics can be very useful for models with a small number of categorical predictors. Unfortunately, these statistics are both sensitive to empty cells. When estimating models with continuous covariates, there are often many empty cells, as in this example. Therefore, you shouldn't rely on either of these test statistics with such models. Because of the empty cells, you can't be sure that these statistics will really follow the chi-square distribution, and the significance values won't be accurate.
--------
You can still use the confusion matrix to assess how well the model does. Also, if you are unsure about the empty cells, you can display the cell information to see exactly where they are.
------------------------------
Jon Peck
------------------------------
Original Message:
Sent: Mon September 11, 2023 10:54 AM
From: Jon Peck
Subject: #SPSS Statistics - Ordinal Logistic Regression
Perhaps you specified your independent variables as factors when they should be covariates, That could produce a large number of empty cells. If you are still not sure why you have empty cells, run Analyze > Reports > Case Summaries on the factors to get a better picture of the data.
As for "the output doesn't seem helpful", that is too vague to respond to.
--
Original Message:
Sent: 9/11/2023 9:13:00 AM
From: James schafer
Subject: #SPSS Statistics - Ordinal Logistic Regression
I am at a loss as to how to correctly run an ordinal logistic regression using the function analyze/regression/ordinal in the student version of SPSS. My raw data set has 228 rows representing the number of responses received from a questionnaire. Responses to eight questions were coded on a five-point Likert type scale. The respondents answered four questions on attitude, three questions on intention, and one question on their highest level of education received. I have summed the attitudes responses to obtain a composite attitude score for each participant from 4 to 20. I repeated the operation for intention to obtain a composite intention score from 3 to 15. In SPSS, I selected analyze/regression/ordinal with the dependent variable being the composite intention score and the independent variables being the composite attitude and educational level scores. The data are not normally distributed. My output shows a large number of empty cells which I don't understand, and the output doesn't seem helpful. Ultimately, I am trying to reject or fail to accept the null hypothesis: to what extent does educational level moderate attitude thus impacting intention?
I would appreciate any help the community can provide to solve this problem.
------------------------------
James schafer
------------------------------
