Why Effects package in Julia is returning an error following the output of MixedModels?

user_14537549 0
Viewed 65

I want to run the average marginal effects between two predictors: education and unemployment rate. While the package Effects in Julia can do the trick, i am having an error which i will describe below. First, i show the code that works perfectly fine and the results i get from running the MixedModelspackage. Second, i show the part of the code that is not working.

using Arrow, Effects, DataFrames, MixedModels, PooledArrays, FreqTables
dffnm = "df4_full"
df = DataFrame(Arrow.Table(dffnm))

df1 = DataFrame( 
    unemploy = disallowmissing(df.macro_unemployment),
    workhours_imputed = df.workhours_imputed,
    education = PooledArray(df.education),
    id = PooledArray(string.(disallowmissing(df.id))),
    age = Int8.(disallowmissing(df.age)) .- Int8(42), # center the age at 42
    sex = PooledArray(df.sex),)

form1 = @formula workhours_imputed ~ 1 + sex + age  + education + education * unemploy + sex * unemploy + (1|id)

contr = Dict(nm => Grouping() for nm in (:country, :country_year, :id))

contr = Dict(:country => Grouping(),
             :country_year => Grouping(),
             :id => Grouping(),
             :sex => DummyCoding(base="Female"),
             :education => DummyCoding(base="Tertiary"))

m1 = @time fit(MixedModel, form1, df1, contrasts=contr)

Here is the output: enter image description here

Everything works perfectly so far. The aim is to run the average marginal effects for education and unemployment on working hours. However, i thought that i would start with a simpler example which is just taking the average marginal effects of education (i follow their example here: https://beacon-biosignals.github.io/Effects.jl/stable/#Interaction-Terms-in-Effects):

design = Dict(:education => unique(df1))
effects(design, m1)

Error:

MethodError: no method matching effects!(::Dict{Symbol, DataFrame}, ::LinearMixedModel{Float64})
    Closest candidates are:
      effects!(::DataFrame, ::StatsBase.RegressionModel; eff_col, err_col, typical) at /Users/jmoawad/.julia/packages/Effects/B3CEy/src/regressionmodel.jl:40
    
    Stacktrace:
     [1] top-level scope
       @ In[28]:1
     [2] eval
       @ ./boot.jl:360 [inlined]
     [3] include_string(mapexpr::typeof(REPL.softscope), mod::Module, code::String, filename::String)
       @ Base ./loading.jl:1116

Could someone help me to fix this? I am also unsure how to run the average marginal effects for the interaction between education and unemployment rate.

mixed-models julia
1 Answers

Based on the suggestions of @sundar & @Peter Deffebach (from another thread).

Here is the solution code:

You have to filter the missing first of education after running the model:

filter!(!ismissing, df1.education)

Then you choose the what categories of interest or what range of interest among both variables:

design = Dict(:education => ["Tertiary","Below upper-secondary"],:unemploy => [5,10,15,20,25])

Then you use the effects package:

effects(design, m1)
user_145375490
Related