# Significant direct effect and non significant TOTAL effect

I have a cross-lagged model with 3 waves and two variables. Here the regressions paths.

  CEE3 ~ a * CEE1
CEE5 ~ b * CEE3
TM3  ~ c * TM1
TM3  ~ d * CEE1
CEE3 ~ e * TM1
TM5  ~ f * TM3
TM5  ~ g * CEE3
CEE5 ~ h * TM3
`

The path "g" is significant (p<.01) BUT the TOTAL effect of CEE3 to TM5 is not! As you can see there is no way of having any indirect effects. How can this happen. The TOTAL effect is the direct effect plus any indirect effect, BUT if there is no indirect effect the TOTAL effect should be the direct effect, am I right? How to interpret that? Thank you for your support. Regards,

• It's kind of hard to interpret the model. Do you have a path diagram? (The only way I can work out what is happening is to draw one). – Jeremy Miles May 18 '18 at 19:12
• Also, is this the complete model? Was it saturated? – Jeremy Miles May 18 '18 at 19:12
• (I reformatted the text to make it easier to read). – Jeremy Miles May 18 '18 at 19:14
• Sorry @JeremyMiles I didn´t come back earlier...I just couldn´t add any picture to my post. This is the complete model. if you let me know how can I add a picture to mey post I can sen one. Many thanks once again – Renata Sarmento Sep 17 '18 at 15:47
• I was doing some research and maybe the problem here is talking about total/direct/indirect effects doesn´t make much sense in this kind of models?? We should only take into account the paths coefficients?? – Renata Sarmento Sep 17 '18 at 15:56