Skip to contents

Cross-section data on states in 1950.

Usage

data("MurderRates")

Format

A data frame containing 44 observations on 8 variables.

rate

Murder rate per 100,000 (FBI estimate, 1950).

convictions

Number of convictions divided by number of murders in 1950.

executions

Average number of executions during 1946–1950 divided by convictions in 1950.

time

Median time served (in months) of convicted murderers released in 1951.

income

Median family income in 1949 (in 1,000 USD).

lfp

Labor force participation rate in 1950 (in percent).

noncauc

Proportion of population that is non-Caucasian in 1950.

southern

Factor indicating region.

Source

Maddala (2001), Table 8.4, p. 330

References

Maddala, G.S. (2001). Introduction to Econometrics, 3rd ed. New York: John Wiley.

McManus, W.S. (1985). Estimates of the Deterrent Effect of Capital Punishment: The Importance of the Researcher's Prior Beliefs. Journal of Political Economy, 93, 417–425.

Stokes, H. (2004). On the Advantage of Using Two or More Econometric Software Systems to Solve the Same Problem. Journal of Economic and Social Measurement, 29, 307–320.

Examples

data("MurderRates")

## Maddala (2001, pp. 331)
fm_lm <- lm(rate ~ . + I(executions > 0), data = MurderRates)
summary(fm_lm)
#> 
#> Call:
#> lm(formula = rate ~ . + I(executions > 0), data = MurderRates)
#> 
#> Residuals:
#>     Min      1Q  Median      3Q     Max 
#> -4.5306 -1.1247 -0.2501  0.8773  6.3045 
#> 
#> Coefficients:
#>                        Estimate Std. Error t value Pr(>|t|)  
#> (Intercept)           -8.497374  10.421497  -0.815   0.4204  
#> convictions           -3.696766   2.675976  -1.381   0.1759  
#> executions            -3.566994   6.593152  -0.541   0.5919  
#> time                  -0.017940   0.006836  -2.624   0.0128 *
#> income                -4.094888   1.778390  -2.303   0.0274 *
#> lfp                    0.399650   0.219809   1.818   0.0776 .
#> noncauc                6.444205   5.493819   1.173   0.2487  
#> southernyes            2.540775   1.315494   1.931   0.0616 .
#> I(executions > 0)TRUE  2.598009   1.231312   2.110   0.0421 *
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 2.349 on 35 degrees of freedom
#> Multiple R-squared:  0.7746,	Adjusted R-squared:  0.723 
#> F-statistic: 15.03 on 8 and 35 DF,  p-value: 2.859e-09
#> 

model <- I(executions > 0) ~ time + income + noncauc + lfp + southern
fm_lpm <- lm(model, data = MurderRates)
summary(fm_lpm)
#> 
#> Call:
#> lm(formula = model, data = MurderRates)
#> 
#> Residuals:
#>      Min       1Q   Median       3Q      Max 
#> -0.70536 -0.26992  0.09636  0.27047  0.50990 
#> 
#> Coefficients:
#>               Estimate Std. Error t value Pr(>|t|)   
#> (Intercept)  1.9929477  1.3260699   1.503  0.14113   
#> time         0.0014615  0.0009978   1.465  0.15122   
#> income       0.6577319  0.2404256   2.736  0.00941 **
#> noncauc      1.9882407  0.7595481   2.618  0.01264 * 
#> lfp         -0.0545619  0.0282569  -1.931  0.06098 . 
#> southernyes  0.3433565  0.1892451   1.814  0.07753 . 
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 0.3533 on 38 degrees of freedom
#> Multiple R-squared:  0.3376,	Adjusted R-squared:  0.2504 
#> F-statistic: 3.873 on 5 and 38 DF,  p-value: 0.0062
#> 

## Binomial models. Note: southern coefficient
fm_logit <- glm(model, data = MurderRates, family = binomial)
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
summary(fm_logit)
#> 
#> Call:
#> glm(formula = model, family = binomial, data = MurderRates)
#> 
#> Coefficients:
#>               Estimate Std. Error z value Pr(>|z|)  
#> (Intercept)   10.99326   20.77336   0.529   0.5967  
#> time           0.01943    0.01040   1.868   0.0617 .
#> income        10.61013    5.65409   1.877   0.0606 .
#> noncauc       70.98785   36.41181   1.950   0.0512 .
#> lfp           -0.66763    0.47668  -1.401   0.1613  
#> southernyes   17.33126 2872.17069   0.006   0.9952  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> (Dispersion parameter for binomial family taken to be 1)
#> 
#>     Null deviance: 44.584  on 43  degrees of freedom
#> Residual deviance: 17.465  on 38  degrees of freedom
#> AIC: 29.465
#> 
#> Number of Fisher Scoring iterations: 19
#> 

fm_logit2 <- glm(model, data = MurderRates, family = binomial,
  control = list(epsilon = 1e-15, maxit = 50, trace = FALSE))
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
summary(fm_logit2)
#> 
#> Call:
#> glm(formula = model, family = binomial, data = MurderRates, control = list(epsilon = 1e-15, 
#>     maxit = 50, trace = FALSE))
#> 
#> Coefficients:
#>               Estimate Std. Error z value Pr(>|z|)  
#> (Intercept)  1.099e+01  2.077e+01   0.529   0.5967  
#> time         1.943e-02  1.040e-02   1.868   0.0617 .
#> income       1.061e+01  5.654e+00   1.877   0.0606 .
#> noncauc      7.099e+01  3.641e+01   1.950   0.0512 .
#> lfp         -6.676e-01  4.767e-01  -1.401   0.1613  
#> southernyes  3.133e+01  1.733e+07   0.000   1.0000  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> (Dispersion parameter for binomial family taken to be 1)
#> 
#>     Null deviance: 44.584  on 43  degrees of freedom
#> Residual deviance: 17.465  on 38  degrees of freedom
#> AIC: 29.465
#> 
#> Number of Fisher Scoring iterations: 33
#> 

fm_probit <- glm(model, data = MurderRates, family = binomial(link = "probit"))
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
summary(fm_probit)
#> 
#> Call:
#> glm(formula = model, family = binomial(link = "probit"), data = MurderRates)
#> 
#> Coefficients:
#>              Estimate Std. Error z value Pr(>|z|)  
#> (Intercept)   6.91551   11.34891   0.609   0.5423  
#> time          0.01131    0.00567   1.995   0.0460 *
#> income        6.46112    3.10937   2.078   0.0377 *
#> noncauc      42.49827   19.62469   2.166   0.0303 *
#> lfp          -0.40929    0.26523  -1.543   0.1228  
#> southernyes   5.46462  551.39841   0.010   0.9921  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> (Dispersion parameter for binomial family taken to be 1)
#> 
#>     Null deviance: 44.584  on 43  degrees of freedom
#> Residual deviance: 17.281  on 38  degrees of freedom
#> AIC: 29.281
#> 
#> Number of Fisher Scoring iterations: 19
#> 

fm_probit2 <- glm(model, data = MurderRates , family = binomial(link = "probit"),
  control = list(epsilon = 1e-15, maxit = 50, trace = FALSE))
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
summary(fm_probit2)
#> 
#> Call:
#> glm(formula = model, family = binomial(link = "probit"), data = MurderRates, 
#>     control = list(epsilon = 1e-15, maxit = 50, trace = FALSE))
#> 
#> Coefficients:
#>               Estimate Std. Error z value Pr(>|z|)  
#> (Intercept)  6.916e+00  1.135e+01   0.609   0.5423  
#> time         1.131e-02  5.670e-03   1.995   0.0460 *
#> income       6.461e+00  3.109e+00   2.078   0.0377 *
#> noncauc      4.250e+01  1.962e+01   2.166   0.0303 *
#> lfp         -4.093e-01  2.652e-01  -1.543   0.1228  
#> southernyes  7.881e+00  1.253e+06   0.000   1.0000  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> (Dispersion parameter for binomial family taken to be 1)
#> 
#>     Null deviance: 44.584  on 43  degrees of freedom
#> Residual deviance: 17.281  on 38  degrees of freedom
#> AIC: 29.281
#> 
#> Number of Fisher Scoring iterations: 35
#> 

## Explanation: quasi-complete separation
with(MurderRates, table(executions > 0, southern))
#>        southern
#>         no yes
#>   FALSE  9   0
#>   TRUE  20  15