Package 'summclust'

Plotting method for objects of type summclust

Description

Plots residual leverage, partial leverage and the leave-one-cluster-out regression coefficients

Usage

## S3 method for class 'summclust'
plot(x, ...)

Arguments

x	An object of type summclust
...	other optional function arguments

Details

Note that the function requires ggplot2 to be installed.

Value

A list containing

residual_leverage	A ggplot of the residual leverages
coef_leverage	A ggplot of the coefficient leverages
coef_beta	A ggplot of the leave-one-out cluster jackknife regression coefficients

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = lm_fit,
   params = c("msp", "union"),
   cluster = ~ind_code,
 )

 plot(res)
}

---

A summary() method for objects of type summclust

Description

A summary() method for objects of type summclust

Usage

## S3 method for class 'summclust'
summary(object, ...)

Arguments

object	An object of type summclust
...	misc arguments

Value

The function summary.summclust returns a range of cluster leverage statistics based on an object of type summclust

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){
library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = lm_fit,
   params = c("msp", "union"),
   cluster = ~ind_code,
 )

 summary(res)
}

---

Compute Influence and Leverage Metrics

Description

Compute influence and leverage metrics for clustered inference based on the Cluster Jackknife described in MacKinnon, Nielsen & Webb (2022).

Usage

summclust(obj, ...)

Arguments

obj	An object of class lm or fixest
...	Other arguments

Value

An object of type summclust, including a CRV3 variance-covariance estimate as described in MacKinnon, Nielsen & Webb (2022)

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

See Also

summclust.lm, summclust.fixest

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = lm_fit,
   params = c("msp", "union"),
   cluster = ~ind_code,
 )

 summary(res)
 tidy(res)
 plot(res)
}

---

Compute Influence and Leverage Metrics for objects of type fixest

Description

Compute influence and leverage metrics for clustered inference based on the Cluster Jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type fixest.

Usage

## S3 method for class 'fixest'
summclust(
  obj,
  cluster,
  params,
  absorb_cluster_fixef = TRUE,
  type = "CRV3",
  ...
)

Arguments

obj	An object of type fixest
cluster	A clustering vector
params	A character vector of variables for which leverage statistics should be computed. If NULL, leverage statistics will be computed for all k model covariates
absorb_cluster_fixef	TRUE by default. Should the cluster fixed effects be projected out? This increases numerical stability and decreases computational costs
type	"CRV3" or "CRV3J" following MacKinnon, Nielsen & Webb
...	other function arguments passed to 'vcov'

Value

An object of type summclust, including a CRV3 variance-covariance estimate as described in MacKinnon, Nielsen & Webb (2022)

coef_estimates	The coefficient estimates of the linear model.
vcov	A CRV3 or CRV3J variance-covariance matrix estimate as described in MacKinnon, Nielsen & Webb (2022)
leverage_g	A vector of leverages.
leverage_avg	The cluster leverage.
partial_leverage	The partial leverages.
coef_var_leverage_avg	Coefficient of Variation for the leverage statistic
coef_var_leverage_g	Coefficient of Variation for the Partial Leverage Statistics
coef_var_N_G	Coefficient of Variation for the Cluster Sizes.
beta_jack	The jackknifed' leave-on-cluster-out regression coefficients.
params	The input parameter vector 'params'.
N_G	The number of clusters-
call	The summclust() function call.
cluster	The names of the clusters.

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust")
 && requireNamespace("haven")
 && requireNamespace("fixest")){

library(summclust)
library(haven)
library(fixest)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

feols_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = feols_fit,
   params = c("msp", "union"),
   cluster = ~ind_code,
 )

 summary(res)
 tidy(res)
 plot(res)
}

---

Compute Influence and Leverage Metrics for objects of type lm

Description

Compute influence and leverage metrics for clustered inference based on the Cluster Jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type lm.

Usage

## S3 method for class 'lm'
summclust(obj, cluster, params, type = "CRV3", ...)

Arguments

obj	An object of type lm
cluster	A clustering vector
params	A character vector of variables for which leverage statistics should be computed.
type	"CRV3" or "CRV3J" following MacKinnon, Nielsen & Webb. CRV3 by default
...	other function arguments passed to 'vcov'

Value

An object of type summclust, including a CRV3 variance-covariance estimate as described in MacKinnon, Nielsen & Webb (2022)

coef_estimates	The coefficient estimates of the linear model.
vcov	A CRV3 or CRV3J variance-covariance matrix estimate as described in MacKinnon, Nielsen & Webb (2022)
leverage_g	A vector of leverages.
leverage_avg	The cluster leverage.
partial_leverage	The partial leverages.
beta_jack	The jackknifed' leave-on-cluster-out regression coefficients.
params	The input parameter vector 'params'.
N_G	The number of clusters-
call	The summclust() function call.
cluster	The names of the clusters.

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = lm_fit,
   cluster = ~ind_code,
   params = c("msp", "union")
 )

 summary(res)
 tidy(res)
 plot(res)
}

---

S3 method to summarize objects of class boottest into tidy data.frame

Description

Obtain results from a summclust object in a tidy data frame.

Usage

## S3 method for class 'summclust'
tidy(x, ...)

Arguments

x	An object of class 'summclust'
...	Other arguments

Value

A data.frame containing coefficient estimates, t-statistics, standard errors, p-value, and confidence intervals based on CRV3 variance-covariance matrix and t(G-1) distribution

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

res <- summclust(
   obj = lm_fit,
   params = c("msp", "union"),
   cluster = ~ind_code,
 )

 tidy(res)
}

---

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022)

Description

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022)

Usage

vcov_CR3J(obj, ...)

Arguments

obj	An object of class lm or fixest computed?
...	misc function argument

Value

An object of type 'vcov_CR3J'

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

See Also

vcov_CR3J.lm, vcov_CR3J.fixest

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

# CRV3 standard errors
vcov <- vcov_CR3J(
   obj = lm_fit,
   cluster = ~ind_code,
   type = "CRV3"
)

# CRV3 standard errors
vcovJN <- vcov_CR3J(
   obj = lm_fit,
   cluster = ~ind_code,
   type = "CRV3J",
)
}

---

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type fixest

Description

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type fixest

Usage

## S3 method for class 'fixest'
vcov_CR3J(
  obj,
  cluster,
  type = "CRV3",
  return_all = FALSE,
  absorb_cluster_fixef = TRUE,
  ...
)

Arguments

obj	An object of type fixest
cluster	A clustering vector
type	"CRV3" or "CRV3J" following MacKinnon, Nielsen & Webb. CRV3 by default
return_all	Logical scalar, FALSE by default. Should only the vcov be returned (FALSE) or additional results (TRUE)
absorb_cluster_fixef	TRUE by default. Should the cluster fixed effects be projected out? This increases numerical stability.
...	other function arguments passed to 'vcov'

Value

An object of class vcov_CR3J

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust")
&& requireNamespace("haven")
&& requireNamespace("fixest")){

library(summclust)
library(haven)
library(fixest)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

feols_fit <- feols(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

# CRV3 standard errors
vcov <- vcov_CR3J(
   obj = feols_fit,
   cluster = ~ind_code,
   type = "CRV3"
)

# CRV3 standard errors
vcovJN <- vcov_CR3J(
   obj = feols_fit,
   cluster = ~ind_code,
   type = "CRV3J",
)
}

---

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type lm

Description

Compute CRV3 covariance matrices via a cluster jackknife as described in MacKinnon, Nielsen & Webb (2022) for objects of type lm

Usage

## S3 method for class 'lm'
vcov_CR3J(obj, cluster, type = "CRV3", return_all = FALSE, ...)

Arguments

obj	An object of type lm
cluster	A clustering vector
type	"CRV3" or "CRV3J" following MacKinnon, Nielsen & Webb. CRV3 by default
return_all	Logical scalar, FALSE by default. Should only the vcov be returned (FALSE) or additional results (TRUE)
...	other function arguments passed to 'vcov'

Value

An object of class vcov_CR3J

References

MacKinnon, James G., Morten Ørregaard Nielsen, and Matthew D. Webb. "Leverage, influence, and the jackknife in clustered regression models: Reliable inference using summclust." arXiv preprint arXiv:2205.03288 (2022).

Examples

if(requireNamespace("summclust") && requireNamespace("haven")){

library(summclust)
library(haven)

nlswork <- read_dta("http://www.stata-press.com/data/r9/nlswork.dta")
# drop NAs at the moment
nlswork <- nlswork[, c("ln_wage", "grade", "age", "birth_yr", "union", "race", "msp", "ind_code")]
nlswork <- na.omit(nlswork)

lm_fit <- lm(
  ln_wage ~ union +  race + msp + as.factor(birth_yr) + as.factor(age) + as.factor(grade),
  data = nlswork)

# CRV3 standard errors
vcov <- vcov_CR3J(
   obj = lm_fit,
   cluster = ~ind_code,
   type = "CRV3"
)

# CRV3 standard errors
vcovJN <- vcov_CR3J(
   obj = lm_fit,
   cluster = ~ind_code,
   type = "CRV3J",
)
}

---