All of the models in this package are binomial logistic regressions. They can be summarized using the base R function summary().
summary(lr_caused)
#>
#> Call:
#> glm(formula = jur_caused ~ jur_race_w, family = "binomial", data = bullpen)
#>
#> Deviance Residuals:
#> Min 1Q Median 3Q Max
#> -1.275 -1.032 -1.032 1.140 1.330
#>
#> Coefficients:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.35118 0.05024 -6.990 2.75e-12 ***
#> jur_race_wBlack 0.46857 0.09616 4.873 1.10e-06 ***
#> jur_race_wHispanic 0.57681 0.09992 5.772 7.81e-09 ***
#> jur_race_wAsian 0.44041 0.16749 2.629 0.00855 **
#> jur_race_wOther 0.23939 0.15130 1.582 0.11361
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> (Dispersion parameter for binomial family taken to be 1)
#>
#> Null deviance: 4324.5 on 3127 degrees of freedom
#> Residual deviance: 4276.9 on 3123 degrees of freedom
#> AIC: 4286.9
#>
#> Number of Fisher Scoring iterations: 4If a more presentation-ready summary is desired, the summ function from the jtools package works well.
jtools::summ(lr_caused)| Observations | 3128 |
| Dependent variable | jur_caused |
| Type | Generalized linear model |
| Family | binomial |
| Link | logit |
| 𝛘²(4) | 47.67 |
| Pseudo-R² (Cragg-Uhler) | 0.02 |
| Pseudo-R² (McFadden) | 0.01 |
| AIC | 4286.87 |
| BIC | 4317.11 |
| Est. | S.E. | z val. | p | |
|---|---|---|---|---|
| (Intercept) | -0.35 | 0.05 | -6.99 | 0.00 |
| jur_race_wBlack | 0.47 | 0.10 | 4.87 | 0.00 |
| jur_race_wHispanic | 0.58 | 0.10 | 5.77 | 0.00 |
| jur_race_wAsian | 0.44 | 0.17 | 2.63 | 0.01 |
| jur_race_wOther | 0.24 | 0.15 | 1.58 | 0.11 |
| Standard errors: MLE |
The odds ratios of each model, along with a 95% confidence interval, can be calculated using the following:
calculate_odds(lr_caused)
#> Waiting for profiling to be done...
#> OR 2.5 % 97.5 %
#> (Intercept) 0.7038582 0.6376588 0.7764931
#> jur_race_wBlack 1.5977013 1.3235160 1.9296937
#> jur_race_wHispanic 1.7803473 1.4642817 2.1666846
#> jur_race_wAsian 1.5533432 1.1188127 2.1596235
#> jur_race_wOther 1.2704701 0.9434932 1.7086446To reproduce the figures as they are seen in the paper, the following code can be used:
calculate_odds(lr_caused) |>
kableExtra::kable(digits = 3) |>
kableExtra::kable_classic(full_width = FALSE, html_font = "Cambria")
#> Waiting for profiling to be done...| OR | 2.5 % | 97.5 % | |
|---|---|---|---|
| (Intercept) | 0.704 | 0.638 | 0.776 |
| jur_race_wBlack | 1.598 | 1.324 | 1.930 |
| jur_race_wHispanic | 1.780 | 1.464 | 2.167 |
| jur_race_wAsian | 1.553 | 1.119 | 2.160 |
| jur_race_wOther | 1.270 | 0.943 | 1.709 |