Dear Statalist:
I have the following question on how to interpret the results of the ## interactions for Cancer and caregiving, in the following model using modified Poisson regression analysis with robust standard errors:
glm k6cat Cancer##caregiving i.sex_cat ib0.age_cat ib0.edu i.job i.setaiexpbyninsu_cat ib1.smoking Diabetes Eye_disease Hypertension Stroke Myocardial_infarction Asthma Gastroduodenal_dis Disease_of_liver_bowel Rheumatoid_arthritis Arthropathy Lower_back_pain Osteoporosis Fracture, fam(Poisson) link(log) nolog eform vce(robust) allbaselevels
---------------------------------------------------------------------------------------------
| Robust
k6cat | IRR Std. Err. z P>|z| [95% Conf. Interval]
----------------------------+----------------------------------------------------------------
Cancer |
not cancer | 1 (base)
cancer | 1.467965 .0191241 29.47 0.000 1.430957 1.50593
|
caregiving |
not caregiving | 1 (base)
caregiving | 1.297198 .0212495 15.88 0.000 1.256211 1.339522
|
Cancer#caregiving |
not cancer#not caregiving | 1 (base)
not cancer#caregiving | 1 (base)
cancer#not caregiving | 1 (base)
cancer#caregiving | .9320186 .0878053 -0.75 0.455 .7748775 1.121027
.
.
.
.
_cons | .2917038 .0021386 -168.05 0.000 .2875423 .2959256
For what I understand is:
1*cons= is the base value in the interaction: not cancer*not caregiving
2* At Cancer, it means the IRR values when only Cancer as separate/not interacted variable.
Cancer
not cancer 1 (base)
cancer: the IRR value when cancer=1
3* Or it means:
Cancer
not cancer 1 (base)
cancer: the IRR value of the term caregiving=0/cancer=1
That is the reason why cancer#not caregiving doesnt appear in the following as base
Cancer#caregiving |
not cancer#not caregiving | 1 (base)
not cancer#caregiving | 1 (base)
cancer#not caregiving | 1 (base)
because its value has been already showned in
Cancer |
not cancer | 1 (base)
cancer | 1.467965 .0191241 29.47 0.000 1.430957 1.50593
Thank you so much for your help.
Related Posts with Interaction intepretation
density plothello. I am not familiar with the density plot and I have a simple question. I have information abou…
Nested loop over observations and varlistHello everyone! I am trying to do something which, upon reading manuals and forum entries, seems per…
Dropping observation under conditionsDear expers, my data: Code: * Example generated by -dataex-. To install: ssc install dataex clear …
Drop variable if varlabel contains certain expressionHey Statalisters, I am trying to drop all variables that have the expression "Domestic Currency" in…
How to treat differing likert scalesHi all, I am using panel data from British Election Survey to perform a fixed effect OLS regression …
Subscribe to:
Post Comments (Atom)
0 Response to Interaction intepretation
Post a Comment