| summ.lm | R Documentation |
Linear regression summaries with options
Description
summ() prints output for a regression model in a fashion similar tosummary(), but formatted differently with more options.
Usage
## S3 method for class 'lm'summ( model, scale = FALSE, confint = getOption("summ-confint", FALSE), ci.width = getOption("summ-ci.width", 0.95), robust = getOption("summ-robust", FALSE), cluster = NULL, vifs = getOption("summ-vifs", FALSE), digits = getOption("jtools-digits", 2), pvals = getOption("summ-pvals", TRUE), n.sd = 1, center = FALSE, transform.response = FALSE, scale.only = FALSE, data = NULL, part.corr = FALSE, model.info = getOption("summ-model.info", TRUE), model.fit = getOption("summ-model.fit", TRUE), model.coefs = getOption("summ-model.coefs", TRUE), which.cols = NULL, vcov = NULL, ...)Arguments
model | A |
scale | If |
confint | Show confidence intervals instead of standard errors? Defaultis |
ci.width | A number between 0 and 1 that signifies the width of thedesired confidence interval. Default is |
robust | If not Default is This requires the |
cluster | For clustered standard errors, provide the column name ofthe cluster variable in the input data frame (as a string). Alternately,provide a vector of clusters. Note that you must set |
vifs | If |
digits | An integer specifying the number of digits past the decimal toreport in the output. Default is 2. You can change the default number ofdigits for all jtools functions with |
pvals | Show p values? If |
n.sd | If |
center | If you want coefficients for mean-centered variables but don'twant to standardize, set this to |
transform.response | Should scaling/centering apply to responsevariable? Default is |
scale.only | If you want to scale but not center, set this to |
data | If you provide the data used to fit the model here, that dataframe is used to re-fit the model (if |
part.corr | Print partial (labeled "partial.r") andsemipartial (labeled "part.r") correlations with the table?Default is |
model.info | Toggles printing of basic information on sample size,name of DV, and number of predictors. |
model.fit | Toggles printing of model fit statistics. |
model.coefs | Toggles printing of model coefficents. |
which.cols | Developmental feature. By providing columns by name,you can add/remove/reorder requested columns in the output. Not fullysupported, for now. |
vcov | You may provide your own variance-covariance matrix for theregression coefficients if you want to calculate standard errors insome way not accommodated by the |
... | Among other things, arguments are passed to |
Details
By default, this function will print the following items to theconsole:
The sample size
The name of the outcome variable
The R-squared value plus adjusted R-squared
A table with regression coefficients, standard errors, t-values, andp values.
There are several options available for robust. The heavylifting is done by sandwich::vcovHC(), where those are betterdescribed.Put simply, you may choose from "HC0" to "HC5". Based on therecommendation of the developers of sandwich, the default is set to"HC3". Stata's default is "HC1", so that choice may be betterif the goal is to replicate Stata's output. Any option that is understoodby vcovHC() will be accepted. Cluster-robust standard errors arecomputed if cluster is set to the name of the input data's clustervariable or is a vector of clusters.
The scale and center options are performed viarefittingthe model with scale_mod() and center_mod(),respectively. Each of those in turn uses gscale() for themean-centering and scaling.
If using part.corr = TRUE, then you will get these two commoneffect size metrics on the far right two columns of the output table.However, it should be noted that these do not go hand in hand withrobust standard error estimators. The standard error of the coefficientdoesn't change the point estimate, just the uncertainty. However,this function uses t-statistics in its calculation of thepartial and semipartial correlation. This provides what amounts to aheteroskedasticity-adjusted set of estimates, but I am unaware of anystatistical publication that validates this type of use. Pleaseuse these as a heuristic when used alongside robust standard errors; donot report the "robust" partial and semipartial correlations inpublications.
Value
If saved, users can access most of the items that are returned inthe output (and without rounding).
coeftable | The outputted table of variables and coefficients |
model | The model for which statistics are displayed. This would bemost useful in cases in which |
Much other information can be accessed as attributes.
Author(s)
Jacob Long jacob.long@sc.edu
References
King, G., & Roberts, M. E. (2015). How robust standard errors exposemethodologicalproblems they do not fix, and what to do about it. PoliticalAnalysis,23(2), 159–179. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1093/pan/mpu015")}
Lumley, T., Diehr, P., Emerson, S., & Chen, L. (2002). The Importance of theNormality Assumption in Large Public Health Data Sets.Annual Review ofPublic Health, 23, 151–169.\Sexpr[results=rd]{tools:::Rd_expr_doi("10.1146/annurev.publhealth.23.100901.140546")}
See Also
scale_mod() can simply perform the standardization ifpreferred.
gscale() does the heavy lifting for mean-centering and scalingbehind the scenes.
Other summ: summ.glm(),summ.merMod(),summ.rq(),summ.svyglm()
Examples
# Create lm objectfit <- lm(Income ~ Frost + Illiteracy + Murder, data = as.data.frame(state.x77))# Print the output with standardized coefficients and 3 digitssumm(fit, scale = TRUE, digits = 3)