Nominal logistic regression examines the relationship between one or
more predictors and a nominal response.
The nominal logistic equation treats each nominal outcome separately.
The logistic regression equation is comprised of multiple logit functions,
one for each value of the response minus one. Each equation has a unique
slope for the predictors. These equations evaluate how the probability
of observing a particular nominal outcome relative to the probability
of observing the reference outcome changes as the predictor variables
change.
Example Output |

|
Logistic Regression Table
Odds 95% CI
Predictor Coef SE Coef Z P Ratio Lower Upper
Logit 1: (8/1)
Constant -0.919125 0.446453 -2.06 0.040
RaceOdds 0.143745 0.0549665 2.62 0.009 1.15 1.04 1.29
Logit 2: (7/1)
Constant -2.11912 0.523139 -4.05 0.000
RaceOdds 0.184382 0.0548107 3.36 0.001 1.20 1.08 1.34
Logit 3: (6/1)
Constant -1.14562 0.451970 -2.53 0.011
RaceOdds 0.159653 0.0546516 2.92 0.003 1.17 1.05 1.31
Logit 4: (5/1)
Constant -0.839914 0.444873 -1.89 0.059
RaceOdds 0.137946 0.0551381 2.50 0.012 1.15 1.03 1.28
Logit 5: (4/1)
Constant -1.11708 0.463681 -2.41 0.016
RaceOdds 0.143264 0.0553128 2.59 0.010 1.15 1.04 1.29
Logit 6: (3/1)
Constant -0.571955 0.439702 -1.30 0.193
RaceOdds 0.117747 0.0559315 2.11 0.035 1.12 1.01 1.26
Logit 7: (2/1)
Constant -0.243669 0.453674 -0.54 0.591
RaceOdds 0.0635537 0.0612533 1.04 0.299 1.07 0.95 1.20
Log-likelihood = -389.629
Test that all slopes are zero: G = 45.535, DF = 7, P-Value = 0.000 |
Interpretation |

|
For the horse racing data, because there are 8 values in the response
information table, there are 7 logit equations.
· Logit
1 is formed using the coefficients for Constant (-0.919125)
and Odds (0.143745) in the logit equation. This
logit equation compares the probability of the horse finishing last to
the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing last instead of first.
· Logit
2 is formed using the coefficients for Constant (-2.11912)
and Odds (0.184382) in the logit equation. This
logit equation compares the probability of the horse finishing seventh
to the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing seventh instead of first.
· Logit
3 is formed using the coefficients for Constant (-1.14562)
and Odds (0.159653) in the logit equation. This
logit equation compares the probability of the horse finishing sixth to
the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing sixth instead of first.
· Logit
4 is formed using the coefficients for Constant (-0.839914)
and Odds (0.137946) in the logit equation. This
logit equation compares the probability of the horse finishing fifth to
the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing fifth instead of first.
· Logit
5 is formed using the coefficients for Constant (-1.11708)
and Odds (0.143264) in the logit equation. This
logit equation compares the probability of the horse finishing fourth
to the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the high the probability of finishing fourth instead of first.
· Logit
6 is formed using the coefficients for Constant (-0.571955)
and Odds (0.117747) in the logit equation. This
logit equation compares the probability of the horse finishing third to
the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing third instead of first.
· Logit
7 is formed using the coefficients for Constant (-0.243669
and Odds (0.0635537) in the logit equation.
This logit equation compares the probability of the horse finishing second
to the probability of a horse finishing first for a given value of RaceOdds.
The positive coefficient for RaceOdds implies that the higher the odds,
the higher the probability of finishing second instead of first.