I am running a model which has many interactions terms and I am facing difficulties in interpreting my results.
My dependent variable is the log of the hourly wage. The main covariates are female (1 if the respondent is female), vet_yes ( which is equal to one if the highest qualification of the individual is vocational qualification) and age which is a continuous variable (centered).
The main problem is that while I can easily understand where those coefficient come from when I don't include in the model the interaction between age squared and vet, once I include them I struggle to understand the meaning of those coefficients.
To compute the age slope for each combination of vet and gender I used:
margins, dydx(centered_age) over(female vet_yes)
Could someone advise me on how to interpret those results?
The following is my output:
log_trim_earnhrppp_cont_dcl | Coef. | Std. Err. | t | P>t | [95% Conf. | Interval] |
1.female | -.1833839 | .0057791 | -31.73 | 0.000 | -.1947109 | -.172057 |
1.vet_yes | -.0471009 | .0087654 | -5.37 | 0.000 | -.0642811 | -.0299207 |
female#vet_yes | ||||||
1 1 | .0034801 | .0078891 | 0.44 | 0.659 | -.0119825 | .0189428 |
centered_age | .0101485 | .0003907 | 25.98 | 0.000 | .0093827 | .0109142 |
female#c.centered_age | ||||||
1 | -.0032397 | .0004942 | -6.56 | 0.000 | -.0042083 | -.002271 |
vet_yes#c.centered_age | ||||||
1 | -.0035711 | .0005128 | -6.96 | 0.000 | -.0045761 | -.0025661 |
female#vet_yes#c.centered_age | ||||||
1 1 | .0030438 | .0006858 | 4.44 | 0.000 | .0016996 | .004388 |
vet_yes#c.centered_age#c.centered_age | ||||||
0 | -.0005042 | .0000218 | -23.18 | 0.000 | -.0005468 | -.0004616 |
1 | -.0002864 | .0000212 | -13.53 | 0.000 | -.0003279 | -.0002449 |
vetc | .2781538 | .0079748 | 34.88 | 0.000 | .2625232 | .2937844 |
vet_miss | -.0026209 | .0117015 | -0.22 | 0.823 | -.0255558 | .020314 |
centered_numscore1 | .0922938 | .0026553 | 34.76 | 0.000 | .0870893 | .0974983 |
uni_degree | .2847923 | .0065771 | 43.30 | 0.000 | .2719011 | .2976834 |
prof_degree | .1373752 | .0065492 | 20.98 | 0.000 | .1245387 | .1502117 |
post_secondary | .0633137 | .0080403 | 7.87 | 0.000 | .0475547 | .0790727 |
med_upper_postsec | .0336602 | .0051705 | 6.51 | 0.000 | .0235261 | .0437943 |
med_unidegree | .0302929 | .0072227 | 4.19 | 0.000 | .0161364 | .0444494 |
med_miss | -.0231609 | .0147224 | -1.57 | 0.116 | -.0520169 | .0056952 |
ded_upper_postsec | .0255946 | .0051962 | 4.93 | 0.000 | .0154101 | .0357791 |
ded_unidegree | .0266815 | .0068149 | 3.92 | 0.000 | .0133242 | .0400388 |
ded_miss | -.0342802 | .0119437 | -2.87 | 0.004 | -.0576899 | -.0108705 |
books | .0101725 | .0017325 | 5.87 | 0.000 | .0067769 | .0135682 |
children | .0292157 | .0050228 | 5.82 | 0.000 | .0193709 | .0390605 |
daustria | -.3231752 | .0088421 | -36.55 | 0.000 | -.3405058 | -.3058447 |
dczechrepublic | -1.027535 | .0137303 | -74.84 | 0.000 | -1.054446 | -1.000624 |
destonia | -.7824016 | .0107683 | -72.66 | 0.000 | -.8035076 | -.7612956 |
dfinland | -.284851 | .0075514 | -37.72 | 0.000 | -.2996518 | -.2700501 |
dfrance | -.4507601 | .0079694 | -56.56 | 0.000 | -.4663802 | -.4351399 |
dgermany | -.3003515 | .0102186 | -29.39 | 0.000 | -.3203801 | -.280323 |
direland | .1254127 | .0123201 | 10.18 | 0.000 | .1012653 | .14956 |
djapan | -.2922338 | .0104464 | -27.97 | 0.000 | -.3127088 | -.2717588 |
dkorea | -.2558537 | .0139317 | -18.36 | 0.000 | -.28316 | -.2285475 |
dnetherlands | -.1632647 | .0094638 | -17.25 | 0.000 | -.1818137 | -.1447157 |
dnorway | -.012887 | .0075659 | -1.70 | 0.089 | -.0277163 | .0019423 |
dpoland | -1.065299 | .0123996 | -85.91 | 0.000 | -1.089602 | -1.040995 |
dslovakrepublic | -1.141286 | .0119926 | -95.17 | 0.000 | -1.164792 | -1.117781 |
dspain | -.3037618 | .0133448 | -22.76 | 0.000 | -.3299177 | -.2776059 |
dsweden | -.2408123 | .0075793 | -31.77 | 0.000 | -.2556678 | -.2259569 |
duk | -.0637246 | .0112755 | -5.65 | 0.000 | -.0858247 | -.0416245 |
dusa | .0634285 | .0121456 | 5.22 | 0.000 | .0396232 | .0872339 |
_cons | 2.847976 | .0107049 | 266.04 | 0.000 | 2.826994 | 2.868957 |
Maria
0 Response to Interpreting interaction effects
Post a Comment