Dear all,

I am computing a multinomial logit on two outcomes (Marriage, Dissolution vs a reference outcome Censored).
I perform an interaction between eddyad and save23. As you can see, from the equation of Marriage, the interaction for a specific level of education and save23 results signficant.
This result would be in line with my hypothesis. However, if I do the Wald test for the all interaction (see below),
the whole interaction results not signficant.

I am uncertain on whether it is the case of commenting on the signficant interaction in the equation of Marriage.
If you could give me your advice, it would be heplful.

Thank you.Best regards, Lydia
Multnomial output
Code:
. svy: mlogit uniontype2  i.durationtype i.eddyad##i.save23 i.wave5 ib0.evunionpr   if marital==3&duratyrs3>
> 0&duratyrs3<6&sampst==1&ivfioall==1&homo3==0&(minagentry>18&minagentry<37)&sex!=.&educ2_f!=4&educ2_m!=4&ag
> esm2_f!=.&agesm2_m!=.  , rrr level(95) 
(running mlogit on estimation sample)

Survey: Multinomial logistic regression

Number of strata   =         1                  Number of obs     =      3,430

Linearized
uniontype2         RRR      Std. Err.      t    P>t     [95% Conf. Interval]

Dissolution         
durationtype 
3rd-4th-5th     .5506879   .1072258    -3.06   0.002     .3757766    .8070146

eddyad 
One HE         .7122864   .1922541     -1.26   0.209     .4193515    1.209849
Both HE         1.012125   .3568225     0.03   0.973     .5066604    2.021861

1.save23        1.412005   .3301809     1.48   0.140     .8922956    2.234414

eddyad#save23 
One HE#1       1.166484   .4611182     0.39   0.697     .5369254    2.534217
Both HE#1      .9583136   .5243717    -0.08   0.938     .3274022    2.805006

wave5 
1998-2008      1.552379   .3395749     2.01   0.045       1.0105     2.38484
2009-2019      1.623017   .4363169     1.80   0.072     .9575636    2.750924

evunionpr 
Re-p man     1.274238   .3198818     0.97   0.335     .7785061    2.085638
Re-par       2.150853   .5067261     3.25   0.001     1.354527    3.415339
Missing      8.367679   2.235536     7.95   0.000     4.953031    14.13641

_cons       .0621653   .0142148   -12.15   0.000     .0396856    .0973785


Marriage            
durationtype 
3rd-4th-5th  1.406513   .1967009     2.44   0.015     1.068891    1.850777

eddyad 
One HE      1.217075   .2510323     0.95   0.341      .811894    1.824464
Both HE     1.627588   .3438196     2.31   0.021     1.075168    2.463842

1.save23    .7895424   .1937916    -0.96   0.336     .4876989    1.278201

eddyad#save23 
One HE#1     .5595597    .221171    -1.47   0.142     .2575857    1.215546
Both HE#1    .3399812    .160764    -2.28   0.023     .1343942    .8600612

wave5 
1998-2008     .5290171   .0940735    -3.58   0.000     .3731534    .7499842
2009-2019     .3611195   .0722338    -5.09   0.000     .2438621    .5347582

evunionpr 
Re-partn     1.274623   .2505097     1.23   0.217     .8666648    1.874616
Re-partne    1.412494   .3152844     1.55   0.122     .9114191    2.189048
Missing      .5560639   .2399269    -1.36   0.174      .238413    1.296939

_cons    .2672354   .0470906    -7.49   0.000     .1890979    .3776603



Wald test
Code:
 ( 1)  [Censored]1o.eddyad#1o.save23 = 0
 ( 2)  [Dissolution]1.eddyad#1.save23 = 0
 ( 3)  [Marriage]1.eddyad#1.save23 = 0
 ( 4)  [Censored]2o.eddyad#1o.save23 = 0
 ( 5)  [Dissolution]2.eddyad#1.save23 = 0
 ( 6)  [Marriage]2.eddyad#1.save23 = 0
       Constraint 1 dropped
       Constraint 4 dropped

       F(  4,   847) =    1.53
            Prob > F =    0.1911