R？中负二项式的序数和伪变量之间的相互作用项

如何解决R？中负二项式的序数和伪变量之间的相互作用项

我使用以下输出运行了负二项式回归模型。

Call:
glm.nb(formula = n_Finance ~ gender + year + party.Conservative + 
    party.LibDem + leadership + years.in.parliament,data = Finance_UK,init.theta = 0.2537048556,link = log)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.5126  -1.2119  -0.6583  -0.0299   4.3157  

Coefficients:
                      Estimate Std. Error z value Pr(>|z|)    
(Intercept)         147.295781  10.129801  14.541  < 2e-16 ***
gender               -0.344608   0.052599  -6.552 5.69e-11 ***
year                 -0.072873   0.005060 -14.401  < 2e-16 ***
party.Conservative    0.413795   0.045723   9.050  < 2e-16 ***
party.LibDem         -0.020742   0.073979  -0.280    0.779    
leadership            0.950248   0.043269  21.962  < 2e-16 ***
years.in.parliament  -0.001248   0.002733  -0.457    0.648    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(dispersion parameter for Negative Binomial(0.2537) family taken to be 1)

    Null deviance: 11014  on 10674  degrees of freedom
Residual deviance: 10169  on 10668  degrees of freedom
AIC: 50414

Number of Fisher Scoring iterations: 1


              Theta:  0.25370 
          Std. Err.:  0.00426 

 2 x log-likelihood:  -50398.09200

我的解释变量如下： gender-女性编码为1，男性编码为0

year-分类变量，范围为1997-2010

party.Conservative-如果该成员是保守党成员则为1

party.LibDem-如果该成员参加了自由民主党，则为1

leadership-如果该成员担任领导职务，则为1；否则，则为0

years.in.parliament-范围从0到31

我的因变量是计数变量n_Finance，是成员在给定年份参加与金融相关的辩论的次数。

我已经解释了这些结果，但是我想在性别和年份之间进行互动，以了解男人或女人在一段时间内（1997年至2010年之间）是否越来越多地参与与金融相关的辩论。但是，当我添加gender*year交互时，我的结果在统计上就变得微不足道了，我不确定如何阅读这些内容。我以前只在OLS中使用带有二进制和连续变量的交互，所以我不确定如何使用它。

这是我添加互动时的输出：

Call:
glm.nb(formula = n_Finance ~ party.Conservative + party.LibDem + 
    leadership + years.in.parliament + gender * year,init.theta = 0.2537585135,link = log)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.5162  -1.2150  -0.6609  -0.0267   4.3501  

Coefficients:
                      Estimate Std. Error z value Pr(>|z|)    
(Intercept)         153.835095  11.172623  13.769   <2e-16 ***
party.Conservative    0.410673   0.045722   8.982   <2e-16 ***
party.LibDem         -0.023913   0.073980  -0.323    0.747    
leadership            0.952143   0.043264  22.008   <2e-16 ***
years.in.parliament  -0.001223   0.002733  -0.447    0.655    
gender              -32.394298  25.114414  -1.290    0.197    
year                 -0.076136   0.005580 -13.644   <2e-16 ***
gender:year           0.015993   0.012533   1.276    0.202    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(dispersion parameter for Negative Binomial(0.2538) family taken to be 1)

    Null deviance: 11016  on 10674  degrees of freedom
Residual deviance: 10169  on 10667  degrees of freedom
AIC: 50414

Number of Fisher Scoring iterations: 1


              Theta:  0.25376 
          Std. Err.:  0.00426 

 2 x log-likelihood:  -50396.39100