Different model types
Example 6.5 of Mize, Doan, and Long (2019): ordinal versus nominal
Marginal effects are on the scale of the outcome – here, probabilities of each party-identification category – so they can be compared across different types of models if both predict the same quantity. This example fits an ordered logit and a multinomial logit to the same five-category party identification variable with the same predictors, and asks whether the two models imply different effects of age. If they do, it is evidence the ordering imposed by the ordinal model is incorrect.
use "https://tdmize.github.io/data/data/gss_cme", clear(gss_cme.dta | GSS 1972 - 2016 Weighted | 2018-07-10)
drop if year < 2010(53,043 observations deleted)
drop if missing(partyid5, woman, edyrs, age, parent, married, faminc, employed, region4, year)(1,244 observations deleted)
.
quietly ologit partyid5 c.age##c.age i.woman c.edyrs i.parent i.married i.race ///
c.faminc i.employed i.region4 i.year, vce(robust)
estimates store ordinal.
quietly mlogit partyid5 c.age##c.age i.woman c.edyrs i.parent i.married i.race ///
c.faminc i.employed i.region4 i.year, vce(robust)
estimates store nominalAge enters with a squared term, so the marginal effect depends on where the change starts. start(age=20) and amount(10) define the effect as the change in each probability from age 20 to age 30, with the other variables held at their means:
mecompare age, models(ordinal nominal) start(age=20) amount(10) covariates(atmeans)Predicting: Pr(partyid5)
Marginal effects and cross-model differences (N_ordinal=8179) (N_nominal=8179)
| ME # Estimate Robust SE P>|z|
---------------------------------+---------------------------------------
age + 10 (centered) |
Strong Dem - ordinal | 1 0.021 0.004 0.000
Strong Dem - nominal | 2 0.029 0.003 0.000
Strong Dem - Difference | 3 -0.008 0.003 0.011
Democrat - ordinal | 4 0.028 0.007 0.000
Democrat - nominal | 5 -0.008 0.013 0.556
Democrat - Difference | 6 0.035 0.011 0.001
Independent - ordinal | 7 -0.002 0.001 0.033
Independent - nominal | 8 0.008 0.012 0.485
Independent - Difference | 9 -0.010 0.012 0.400
Republican - ordinal | 10 -0.028 0.006 0.000
Republican - nominal | 11 -0.039 0.013 0.004
Republican - Difference | 12 0.011 0.012 0.346
Strong Repub - ordinal | 13 -0.020 0.005 0.000
Strong Repub - nominal | 14 0.009 0.003 0.005
Strong Repub - Difference | 15 -0.029 0.005 0.000
Each outcome category gets its own block. A significant Difference means the ordinal model’s proportional-odds assumption changes the conclusion for that category.
Back to top