Print the content of an 'est_change'-class object.
Arguments
- x
An 'est_change'-class object.
- digits
The number of digits after the decimal. Default is 3.
- first
Numeric. If not
NULL, it prints only the first k cases, k equal tofirst. Default is 10.- sort_by
String. Should be
"est","gcd", orNULL. If the output was generated byest_change_raw()orest_change_raw_approx()andsort_byis notNULL, then each column is sorted individually, with case IDs inserted before each column. If the output was generated byest_change()orest_change_approx()andsort_byis notNULL, thensort_bydetermines how the cases are sorted. Ifbyis"est", the cases are sorted as for the output ofest_change_raw(). Ifbyis"gcd", the default for the output ofest_change()orest_change_approx(), then cases are sorted by generalized Cook's distance or approximate generalized Cook's distance, depending on which column is available.- ...
Other arguments. They will be ignored.
Details
All the functions on case influence
on parameter estimates, est_change(),
est_change_approx(), est_change_raw(),
and est_change_raw_approx(), return
an est_change-class object. This method will print
the output based on the type of changes and method
used.
Examples
library(lavaan)
# A path model
dat <- pa_dat
mod <-
"
m1 ~ a1 * iv1 + a2 * iv2
dv ~ b * m1
a1b := a1 * b
a2b := a2 * b
"
# Fit the model
fit <- lavaan::sem(mod, dat)
summary(fit)
#> lavaan 0.6.17 ended normally after 1 iteration
#>
#> Estimator ML
#> Optimization method NLMINB
#> Number of model parameters 5
#>
#> Number of observations 100
#>
#> Model Test User Model:
#>
#> Test statistic 6.711
#> Degrees of freedom 2
#> P-value (Chi-square) 0.035
#>
#> Parameter Estimates:
#>
#> Standard errors Standard
#> Information Expected
#> Information saturated (h1) model Structured
#>
#> Regressions:
#> Estimate Std.Err z-value P(>|z|)
#> m1 ~
#> iv1 (a1) 0.215 0.106 2.036 0.042
#> iv2 (a2) 0.522 0.099 5.253 0.000
#> dv ~
#> m1 (b) 0.517 0.106 4.895 0.000
#>
#> Variances:
#> Estimate Std.Err z-value P(>|z|)
#> .m1 0.903 0.128 7.071 0.000
#> .dv 1.321 0.187 7.071 0.000
#>
#> Defined Parameters:
#> Estimate Std.Err z-value P(>|z|)
#> a1b 0.111 0.059 1.880 0.060
#> a2b 0.270 0.075 3.581 0.000
#>
# Approximate case influence
out <- est_change_approx(fit)
out
#>
#> -- Approximate Standardized Case Influence on Parameter Estimates --
#>
#> m1~iv1 m1~iv2 dv~m1 m1~~m1 dv~~dv gcd_approx
#> 16 0.052 -0.038 -0.228 -0.006 0.572 0.372
#> 43 -0.387 -0.249 -0.135 0.201 0.116 0.270
#> 65 0.150 0.189 0.355 0.071 0.148 0.203
#> 85 -0.170 0.211 -0.118 0.315 -0.054 0.187
#> 51 0.405 -0.052 0.094 0.075 -0.046 0.179
#> 34 -0.306 -0.186 -0.110 0.176 0.028 0.163
#> 32 -0.241 0.190 -0.189 0.181 -0.002 0.161
#> 20 -0.234 0.199 -0.140 0.172 -0.034 0.144
#> 18 -0.269 0.035 0.101 0.246 -0.048 0.143
#> 100 -0.001 -0.221 -0.069 0.290 -0.058 0.137
#>
#> Note:
#> - Changes are approximate standardized raw changes if a case is included.
#> - Only the first 10 case(s) is/are displayed. Set ‘first’ to NULL to display all cases.
#> - Cases sorted by approximate generalized Cook's distance.
print(out, sort_by = "est")
#>
#> -- Approximate Standardized Case Influence on Parameter Estimates --
#>
#> id m1~iv1 id m1~iv2 id dv~m1 id m1~~m1 id dv~~dv id gcd_approx
#> 1 51 0.405 43 -0.249 65 0.355 61 0.335 16 0.572 16 0.372
#> 2 43 -0.387 94 0.230 11 -0.254 85 0.315 9 0.272 43 0.270
#> 3 34 -0.306 100 -0.221 16 -0.228 100 0.290 76 0.267 65 0.203
#> 4 18 -0.269 85 0.211 32 -0.189 18 0.246 25 0.264 85 0.187
#> 5 13 0.267 20 0.199 99 0.187 42 0.225 91 0.230 51 0.179
#> 6 32 -0.241 32 0.190 79 0.176 43 0.201 17 0.209 34 0.163
#> 7 20 -0.234 65 0.189 93 0.169 32 0.181 26 0.151 32 0.161
#> 8 75 0.200 34 -0.186 22 0.161 34 0.176 65 0.148 20 0.144
#> 9 42 -0.194 64 -0.165 61 -0.151 40 0.175 62 0.145 18 0.143
#> 10 68 0.174 52 0.161 25 -0.147 20 0.172 90 0.127 100 0.137
#>
#> Note:
#> - Changes are approximate standardized raw changes if a case is included.
#> - Only the first 10 case(s) is/are displayed. Set ‘first’ to NULL to display all cases.
#> - Cases sorted by the absolute values of change or approximate generalized Cook's distance.
out <- est_change_raw_approx(fit)
print(out, first = 3)
#>
#> -- Approximate Case Influence on Parameter Estimates --
#>
#> id m1~iv1 id m1~iv2 id dv~m1 id m1~~m1 id dv~~dv
#> 1 51 0.042 43 -0.025 65 0.037 61 0.042 16 0.106
#> 2 43 -0.040 94 0.023 11 -0.027 85 0.040 9 0.050
#> 3 34 -0.032 100 -0.022 16 -0.024 100 0.037 76 0.049
#>
#> Note:
#> - Changes are approximate raw changes if a case is included.
#> - Only the first 3 case(s) is/are displayed. Set ‘first’ to NULL to display all cases.
#> - Cases sorted by the absolute changes for each variable.
# Examine four selected cases
fit_rerun <- lavaan_rerun(fit, parallel = FALSE,
to_rerun = c(2, 3, 5, 7))
#> The expected CPU time is 0.15 second(s).
#> Could be faster if run in parallel.
est_change(fit_rerun)
#>
#> -- Standardized Case Influence on Parameter Estimates --
#>
#> a1 a2 b m1~~m1 dv~~dv gcd
#> 7 -0.119 0.073 0.065 -0.002 -0.040 0.026
#> 5 0.067 0.028 0.033 0.050 -0.066 0.013
#> 3 -0.038 -0.040 -0.030 -0.063 -0.045 0.010
#> 2 0.007 0.003 -0.013 -0.067 -0.058 0.008
#>
#> Note:
#> - Changes are standardized raw changes if a case is included.
#> - All stored cases are displayed.
#> - Cases sorted by generalized Cook's distance.
est_change_raw(fit_rerun)
#>
#> -- Case Influence on Parameter Estimates --
#>
#> id m1~iv1 id m1~iv2 id dv~m1 id m1~~m1 id dv~~dv id a1b id a2b
#> 1 7 -0.013 7 0.007 7 0.007 2 -0.009 5 -0.013 7 -0.005 7 0.007
#> 2 5 0.007 3 -0.004 5 0.004 3 -0.008 2 -0.011 5 0.004 3 -0.004
#> 3 3 -0.004 5 0.003 3 -0.003 5 0.006 3 -0.008 3 -0.003 5 0.003
#> 4 2 0.001 2 0.000 2 -0.001 7 0.000 7 -0.008 2 0.000 2 -0.001
#>
#> Note:
#> - Changes are raw changes if a case is included.
#> - All stored cases are displayed.
#> - Cases sorted by the absolute changes for each variable.