1 predictor, 2 different values for baseline statistic & change statistic?

Hi, I am conducting a Cox Regression for my first time. I created a simple model with only 1 predictor (categorical), one dependent variable (continuous survival time), and 1 event that defines whether the hazard criterion was met (dichotomous). Since there is only one step in this model, I expected the Chi-Square coeffecient to be equal to/redundant with the "change in Chi-Square" value given by SPSS... but it is not. The change statistic is slightly larger than the baseline.
1) Given that I have only one predictor and one step, how could this be/what does it mean?
2) Which value is more appropriate to report in an APA research manuscript?
3) Finally, does this issue also affect regular linear regression models which use R-squared?
Many thanks!