data belong to liu
wxy
2019/10/31
library(tidyverse)## -- Attaching packages --------------------------------------- tidyverse 1.2.1 --## √ ggplot2 3.2.1 √ purrr 0.3.2
## √ tibble 2.1.2 √ dplyr 0.8.3
## √ tidyr 1.0.0 √ stringr 1.4.0
## √ readr 1.3.1 √ forcats 0.4.0## Warning: package 'ggplot2' was built under R version 3.6.1## Warning: package 'tidyr' was built under R version 3.6.1## Warning: package 'dplyr' was built under R version 3.6.1## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()library(lavaan)## Warning: package 'lavaan' was built under R version 3.6.1## This is lavaan 0.6-5## lavaan is BETA software! Please report any bugs.library(semPlot)## Warning: package 'semPlot' was built under R version 3.6.1## Registered S3 methods overwritten by 'huge':
## method from
## plot.sim BDgraph
## print.sim BDgraph读数据，并把男女分开，分别作了模型
ddd <- read_csv("./child20164.csv")## Parsed with column specification:
## cols(
## age = col_double(),
## gender = col_double(),
## edu_expect = col_double(),
## care_edu = col_double(),
## communicate = col_double(),
## tutor_fees = col_double(),
## pd501b = col_double(),
## pd577r = col_double(),
## parents_edu = col_double(),
## finc = col_double(),
## urban = col_double(),
## total_care = col_double(),
## tutor_fees_log = col_double(),
## school_fees_log = col_double(),
## other_log = col_double()
## )ddd %>% colnames()## [1] "age" "gender" "edu_expect"
## [4] "care_edu" "communicate" "tutor_fees"
## [7] "pd501b" "pd577r" "parents_edu"
## [10] "finc" "urban" "total_care"
## [13] "tutor_fees_log" "school_fees_log" "other_log"ddd %>% count(urban)## # A tibble: 2 x 2
## urban n
## <dbl> <int>
## 1 0 1006
## 2 1 136boy <- ddd %>%
filter(gender == 1 )
girl <- ddd %>%
filter(gender == 0 )
city <- ddd %>%
filter(urban == 1 )
countryside <- ddd %>%
filter(urban == 0 )#df <- ddd %>%
#select(gender, edu_expect, age)
#da <- as.data.frame(df)
#da
#cor.test(ddd[,2],as.double(ddd[,3]))#总的数据的模型
model <- '
# latent variables
f1 =~ a*tutor_fees_log + s*school_fees_log + d*other_log
f2 =~ f*total_care + g*care_edu + h*communicate
# regressions
f1 ~ j*parents_edu
f2 ~ k*parents_edu
f1 ~ q*finc
f2 ~ w*finc
edu_expect~ e*parents_edu
edu_expect~ r*finc
#f1 ~ t*edu_expect
edu_expect ~ y*f2
edu_expect ~ u*f1
# residual covariances
# max_edu ~~ finc_per
#total
total_max_edu := e + (j*u) + (k*y)
total_finc := r + (w*y) + (q*u)
edu_pay := u
edu_care := y
'
fit <- sem(model,data=ddd )
summary(fit,standardized = TRUE)## lavaan 0.6-5 ended normally after 93 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 21
##
## Number of observations 1142
##
## Model Test User Model:
##
## Test statistic 82.848
## Degrees of freedom 21
## P-value (Chi-square) 0.000
##
## Parameter Estimates:
##
## Information Expected
## Information saturated (h1) model Structured
## Standard errors Standard
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## f1 =~
## tutr_fs_lg (a) 1.000 1.440 0.562
## schl_fs_lg (s) 0.166 0.068 2.435 0.015 0.240 0.166
## other_log (d) 0.353 0.139 2.550 0.011 0.509 0.187
## f2 =~
## total_care (f) 1.000 0.153 0.204
## care_edu (g) 4.750 0.810 5.868 0.000 0.727 0.865
## communicat (h) 4.211 0.686 6.139 0.000 0.644 0.791
##
## Regressions:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## f1 ~
## parents_ed (j) 0.076 0.019 3.963 0.000 0.053 0.201
## f2 ~
## parents_ed (k) 0.004 0.001 3.000 0.003 0.029 0.111
## f1 ~
## finc (q) 0.206 0.056 3.717 0.000 0.143 0.188
## f2 ~
## finc (w) 0.007 0.004 1.832 0.067 0.048 0.062
## edu_expect ~
## parents_ed (e) 0.013 0.012 1.114 0.265 0.013 0.038
## finc (r) -0.024 0.034 -0.728 0.467 -0.024 -0.024
## f2 (y) 1.191 0.334 3.569 0.000 0.182 0.139
## f1 (u) 0.164 0.078 2.106 0.035 0.236 0.180
##
## Variances:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## .tutor_fees_log 4.485 0.808 5.548 0.000 4.485 0.684
## .school_fees_lg 2.017 0.089 22.657 0.000 2.017 0.972
## .other_log 7.169 0.322 22.257 0.000 7.169 0.965
## .total_care 0.540 0.023 23.666 0.000 0.540 0.958
## .care_edu 0.178 0.054 3.305 0.001 0.178 0.252
## .communicate 0.247 0.043 5.731 0.000 0.247 0.374
## .edu_expect 1.617 0.074 21.763 0.000 1.617 0.942
## .f1 1.900 0.797 2.385 0.017 0.917 0.917
## .f2 0.023 0.007 3.151 0.002 0.982 0.982
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## total_max_edu 0.031 0.010 3.030 0.002 0.031 0.090
## total_finc 0.018 0.030 0.612 0.541 0.018 0.018
## edu_pay 0.164 0.078 2.106 0.035 0.236 0.180
## edu_care 1.191 0.334 3.569 0.000 0.182 0.139aa <- parameterEstimates(fit)#这里的est是没标注准化的系数，还有置信区间
Est <- parameterEstimates(fit, ci = FALSE, standardized = TRUE)#est为未标准化的，std.all为标准化后的
subset(Est, lhs == "f2")## lhs op rhs label est se z pvalue std.lv std.all
## 4 f2 =~ total_care f 1.000 0.000 NA NA 0.153 0.204
## 5 f2 =~ care_edu g 4.750 0.810 5.868 0.000 0.727 0.865
## 6 f2 =~ communicate h 4.211 0.686 6.139 0.000 0.644 0.791
## 8 f2 ~ parents_edu k 0.004 0.001 3.000 0.003 0.029 0.111
## 10 f2 ~ finc w 0.007 0.004 1.832 0.067 0.048 0.062
## 23 f2 ~~ f2 0.023 0.007 3.151 0.002 0.982 0.982
## std.nox
## 4 0.204
## 5 0.865
## 6 0.791
## 8 0.029
## 10 0.048
## 23 0.982fitMeasures(fit, c("cfi", "rmsea","nfi","srmr"))## cfi rmsea nfi srmr
## 0.933 0.051 0.914 0.047semPaths(fit,whatLabels="est",style="lisrel",layout = "tree2",
residuals=TRUE)
只有男生的模型
model1 <- '
# latent variables
f1 =~ a*tutor_fees_log + s*school_fees_log + d*other_log
f2 =~ f*total_care + g*care_edu + h*communicate
# regressions
f1 ~ j*parents_edu
f2 ~ k*parents_edu
f1 ~ q*finc
f2 ~ w*finc
edu_expect~ e*parents_edu
edu_expect~ r*finc
#f1 ~ t*edu_expect
edu_expect ~ y*f2
edu_expect ~ u*f1
# residual covariances
# max_edu ~~ finc_per
#total
total_max_edu := e + (j*u) + (k*y)
total_finc := r + (w*y) + (q*u)
edu_pay := u
edu_care := y
'
fit1 <- sem(model1,data=boy )
summary(fit1)## lavaan 0.6-5 ended normally after 90 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 21
##
## Number of observations 627
##
## Model Test User Model:
##
## Test statistic 46.264
## Degrees of freedom 21
## P-value (Chi-square) 0.001
##
## Parameter Estimates:
##
## Information Expected
## Information saturated (h1) model Structured
## Standard errors Standard
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|)
## f1 =~
## tutr_fs_lg (a) 1.000
## schl_fs_lg (s) 0.190 0.086 2.206 0.027
## other_log (d) 0.413 0.176 2.350 0.019
## f2 =~
## total_care (f) 1.000
## care_edu (g) 5.889 1.658 3.552 0.000
## communicat (h) 5.457 1.519 3.591 0.000
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## f1 ~
## parents_ed (j) 0.081 0.026 3.167 0.002
## f2 ~
## parents_ed (k) 0.004 0.002 2.335 0.020
## f1 ~
## finc (q) 0.177 0.071 2.472 0.013
## f2 ~
## finc (w) 0.008 0.004 1.731 0.084
## edu_expect ~
## parents_ed (e) -0.006 0.016 -0.381 0.703
## finc (r) -0.054 0.044 -1.230 0.219
## f2 (y) 1.983 0.740 2.680 0.007
## f1 (u) 0.185 0.093 1.987 0.047
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .tutor_fees_log 4.174 1.003 4.161 0.000
## .school_fees_lg 2.176 0.131 16.627 0.000
## .other_log 7.130 0.445 16.015 0.000
## .total_care 0.523 0.030 17.589 0.000
## .care_edu 0.219 0.069 3.190 0.001
## .communicate 0.234 0.060 3.927 0.000
## .edu_expect 1.770 0.111 16.014 0.000
## .f1 2.199 0.997 2.206 0.027
## .f2 0.014 0.008 1.844 0.065
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## total_max_edu 0.017 0.014 1.180 0.238
## total_finc -0.006 0.041 -0.145 0.885
## edu_pay 0.185 0.093 1.987 0.047
## edu_care 1.983 0.740 2.680 0.007parameterEstimates(fit1)#这里的est是没标注准化的系数，还有置信区间## lhs op rhs label est se z
## 1 f1 =~ tutor_fees_log a 1.000 0.000 NA
## 2 f1 =~ school_fees_log s 0.190 0.086 2.206
## 3 f1 =~ other_log d 0.413 0.176 2.350
## 4 f2 =~ total_care f 1.000 0.000 NA
## 5 f2 =~ care_edu g 5.889 1.658 3.552
## 6 f2 =~ communicate h 5.457 1.519 3.591
## 7 f1 ~ parents_edu j 0.081 0.026 3.167
## 8 f2 ~ parents_edu k 0.004 0.002 2.335
## 9 f1 ~ finc q 0.177 0.071 2.472
## 10 f2 ~ finc w 0.008 0.004 1.731
## 11 edu_expect ~ parents_edu e -0.006 0.016 -0.381
## 12 edu_expect ~ finc r -0.054 0.044 -1.230
## 13 edu_expect ~ f2 y 1.983 0.740 2.680
## 14 edu_expect ~ f1 u 0.185 0.093 1.987
## 15 tutor_fees_log ~~ tutor_fees_log 4.174 1.003 4.161
## 16 school_fees_log ~~ school_fees_log 2.176 0.131 16.627
## 17 other_log ~~ other_log 7.130 0.445 16.015
## 18 total_care ~~ total_care 0.523 0.030 17.589
## 19 care_edu ~~ care_edu 0.219 0.069 3.190
## 20 communicate ~~ communicate 0.234 0.060 3.927
## 21 edu_expect ~~ edu_expect 1.770 0.111 16.014
## 22 f1 ~~ f1 2.199 0.997 2.206
## 23 f2 ~~ f2 0.014 0.008 1.844
## 24 parents_edu ~~ parents_edu 14.736 0.000 NA
## 25 parents_edu ~~ finc 0.556 0.000 NA
## 26 finc ~~ finc 1.858 0.000 NA
## 27 total_max_edu := e+(j*u)+(k*y) total_max_edu 0.017 0.014 1.180
## 28 total_finc := r+(w*y)+(q*u) total_finc -0.006 0.041 -0.145
## 29 edu_pay := u edu_pay 0.185 0.093 1.987
## 30 edu_care := y edu_care 1.983 0.740 2.680
## pvalue ci.lower ci.upper
## 1 NA 1.000 1.000
## 2 0.027 0.021 0.358
## 3 0.019 0.068 0.757
## 4 NA 1.000 1.000
## 5 0.000 2.639 9.139
## 6 0.000 2.479 8.435
## 7 0.002 0.031 0.131
## 8 0.020 0.001 0.008
## 9 0.013 0.037 0.317
## 10 0.084 -0.001 0.016
## 11 0.703 -0.038 0.026
## 12 0.219 -0.139 0.032
## 13 0.007 0.533 3.434
## 14 0.047 0.002 0.368
## 15 0.000 2.208 6.141
## 16 0.000 1.919 2.432
## 17 0.000 6.257 8.002
## 18 0.000 0.464 0.581
## 19 0.001 0.085 0.354
## 20 0.000 0.117 0.350
## 21 0.000 1.553 1.986
## 22 0.027 0.245 4.152
## 23 0.065 -0.001 0.029
## 24 NA 14.736 14.736
## 25 NA 0.556 0.556
## 26 NA 1.858 1.858
## 27 0.238 -0.011 0.045
## 28 0.885 -0.086 0.074
## 29 0.047 0.002 0.368
## 30 0.007 0.533 3.434Est <- parameterEstimates(fit1, ci = FALSE, standardized = TRUE)#est为未标准化的，std.all为标准化后的
subset(Est, lhs == "f2")## lhs op rhs label est se z pvalue std.lv std.all
## 4 f2 =~ total_care f 1.000 0.000 NA NA 0.119 0.163
## 5 f2 =~ care_edu g 5.889 1.658 3.552 0.000 0.703 0.832
## 6 f2 =~ communicate h 5.457 1.519 3.591 0.000 0.652 0.803
## 8 f2 ~ parents_edu k 0.004 0.002 2.335 0.020 0.035 0.133
## 10 f2 ~ finc w 0.008 0.004 1.731 0.084 0.064 0.087
## 23 f2 ~~ f2 0.014 0.008 1.844 0.065 0.972 0.972
## std.nox
## 4 0.163
## 5 0.832
## 6 0.803
## 8 0.035
## 10 0.064
## 23 0.972fitMeasures(fit1, c("cfi", "rmsea","nfi","srmr"))## cfi rmsea nfi srmr
## 0.947 0.044 0.910 0.048semPaths(fit1,whatLabels="std.all",style="lisrel",layout = "tree2",
residuals=TRUE)
只有女生的模型
model2 <- '
# latent variables
f1 =~ a*tutor_fees_log + s*school_fees_log + d*other_log
f2 =~ f*total_care + g*care_edu + h*communicate
# regressions
f1 ~ j*parents_edu
f2 ~ k*parents_edu
f1 ~ q*finc
f2 ~ w*finc
edu_expect~ e*parents_edu
edu_expect~ r*finc
#f1 ~ t*edu_expect
edu_expect ~ y*f2
edu_expect ~ u*f1
# residual covariances
# max_edu ~~ finc_per
#total
total_max_edu := e + (j*u) + (k*y)
total_finc := r + (w*y) + (q*u)
edu_pay := u
edu_care := y
'
fit2 <- sem(model2,data=girl )
summary(fit2)## lavaan 0.6-5 ended normally after 77 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 21
##
## Number of observations 515
##
## Model Test User Model:
##
## Test statistic 49.422
## Degrees of freedom 21
## P-value (Chi-square) 0.000
##
## Parameter Estimates:
##
## Information Expected
## Information saturated (h1) model Structured
## Standard errors Standard
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|)
## f1 =~
## tutr_fs_lg (a) 1.000
## schl_fs_lg (s) 0.176 0.123 1.427 0.154
## other_log (d) 0.289 0.227 1.272 0.203
## f2 =~
## total_care (f) 1.000
## care_edu (g) 3.941 0.845 4.663 0.000
## communicat (h) 3.322 0.641 5.180 0.000
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## f1 ~
## parents_ed (j) 0.068 0.029 2.368 0.018
## f2 ~
## parents_ed (k) 0.004 0.003 1.526 0.127
## f1 ~
## finc (q) 0.256 0.087 2.925 0.003
## f2 ~
## finc (w) 0.006 0.007 0.788 0.431
## edu_expect ~
## parents_ed (e) 0.035 0.018 1.978 0.048
## finc (r) 0.016 0.059 0.274 0.784
## f2 (y) 0.676 0.322 2.100 0.036
## f1 (u) 0.149 0.161 0.925 0.355
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .tutor_fees_log 5.272 1.006 5.238 0.000
## .school_fees_lg 1.807 0.118 15.259 0.000
## .other_log 7.182 0.463 15.517 0.000
## .total_care 0.558 0.035 15.833 0.000
## .care_edu 0.127 0.081 1.572 0.116
## .communicate 0.262 0.060 4.405 0.000
## .edu_expect 1.393 0.095 14.691 0.000
## .f1 1.118 0.957 1.168 0.243
## .f2 0.036 0.014 2.614 0.009
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## total_max_edu 0.048 0.014 3.410 0.001
## total_finc 0.058 0.043 1.365 0.172
## edu_pay 0.149 0.161 0.925 0.355
## edu_care 0.676 0.322 2.100 0.036parameterEstimates(fit2)#这里的est是没标注准化的系数，还有置信区间## lhs op rhs label est se z
## 1 f1 =~ tutor_fees_log a 1.000 0.000 NA
## 2 f1 =~ school_fees_log s 0.176 0.123 1.427
## 3 f1 =~ other_log d 0.289 0.227 1.272
## 4 f2 =~ total_care f 1.000 0.000 NA
## 5 f2 =~ care_edu g 3.941 0.845 4.663
## 6 f2 =~ communicate h 3.322 0.641 5.180
## 7 f1 ~ parents_edu j 0.068 0.029 2.368
## 8 f2 ~ parents_edu k 0.004 0.003 1.526
## 9 f1 ~ finc q 0.256 0.087 2.925
## 10 f2 ~ finc w 0.006 0.007 0.788
## 11 edu_expect ~ parents_edu e 0.035 0.018 1.978
## 12 edu_expect ~ finc r 0.016 0.059 0.274
## 13 edu_expect ~ f2 y 0.676 0.322 2.100
## 14 edu_expect ~ f1 u 0.149 0.161 0.925
## 15 tutor_fees_log ~~ tutor_fees_log 5.272 1.006 5.238
## 16 school_fees_log ~~ school_fees_log 1.807 0.118 15.259
## 17 other_log ~~ other_log 7.182 0.463 15.517
## 18 total_care ~~ total_care 0.558 0.035 15.833
## 19 care_edu ~~ care_edu 0.127 0.081 1.572
## 20 communicate ~~ communicate 0.262 0.060 4.405
## 21 edu_expect ~~ edu_expect 1.393 0.095 14.691
## 22 f1 ~~ f1 1.118 0.957 1.168
## 23 f2 ~~ f2 0.036 0.014 2.614
## 24 parents_edu ~~ parents_edu 14.171 0.000 NA
## 25 parents_edu ~~ finc 0.457 0.000 NA
## 26 finc ~~ finc 1.544 0.000 NA
## 27 total_max_edu := e+(j*u)+(k*y) total_max_edu 0.048 0.014 3.410
## 28 total_finc := r+(w*y)+(q*u) total_finc 0.058 0.043 1.365
## 29 edu_pay := u edu_pay 0.149 0.161 0.925
## 30 edu_care := y edu_care 0.676 0.322 2.100
## pvalue ci.lower ci.upper
## 1 NA 1.000 1.000
## 2 0.154 -0.066 0.417
## 3 0.203 -0.156 0.734
## 4 NA 1.000 1.000
## 5 0.000 2.285 5.598
## 6 0.000 2.065 4.579
## 7 0.018 0.012 0.124
## 8 0.127 -0.001 0.009
## 9 0.003 0.084 0.427
## 10 0.431 -0.009 0.020
## 11 0.048 0.000 0.070
## 12 0.784 -0.100 0.132
## 13 0.036 0.045 1.307
## 14 0.355 -0.167 0.465
## 15 0.000 3.299 7.244
## 16 0.000 1.575 2.039
## 17 0.000 6.275 8.089
## 18 0.000 0.489 0.627
## 19 0.116 -0.031 0.286
## 20 0.000 0.146 0.379
## 21 0.000 1.207 1.579
## 22 0.243 -0.757 2.993
## 23 0.009 0.009 0.064
## 24 NA 14.171 14.171
## 25 NA 0.457 0.457
## 26 NA 1.544 1.544
## 27 0.001 0.020 0.076
## 28 0.172 -0.025 0.142
## 29 0.355 -0.167 0.465
## 30 0.036 0.045 1.307Est <- parameterEstimates(fit2, ci = FALSE, standardized = TRUE)#est为未标准化的，std.all为标准化后的
subset(Est, lhs == "f2")## lhs op rhs label est se z pvalue std.lv std.all
## 4 f2 =~ total_care f 1.000 0.000 NA NA 0.192 0.248
## 5 f2 =~ care_edu g 3.941 0.845 4.663 0.000 0.755 0.904
## 6 f2 =~ communicate h 3.322 0.641 5.180 0.000 0.636 0.779
## 8 f2 ~ parents_edu k 0.004 0.003 1.526 0.127 0.020 0.076
## 10 f2 ~ finc w 0.006 0.007 0.788 0.431 0.031 0.038
## 23 f2 ~~ f2 0.036 0.014 2.614 0.009 0.992 0.992
## std.nox
## 4 0.248
## 5 0.904
## 6 0.779
## 8 0.020
## 10 0.031
## 23 0.992fitMeasures(fit2, c("cfi", "rmsea","nfi","srmr"))## cfi rmsea nfi srmr
## 0.936 0.051 0.896 0.051semPaths(fit2,whatLabels="std.all",style="lisrel",layout = "tree2",
residuals=TRUE)
城市的模型
model2 <- '
# latent variables
f1 =~ a*tutor_fees_log + s*school_fees_log + d*other_log
f2 =~ f*total_care + g*care_edu + h*communicate
# regressions
f1 ~ j*parents_edu
f2 ~ k*parents_edu
f1 ~ q*finc
f2 ~ w*finc
edu_expect~ e*parents_edu
edu_expect~ r*finc
#f1 ~ t*edu_expect
edu_expect ~ y*f2
edu_expect ~ u*f1
# residual covariances
# max_edu ~~ finc_per
#total
total_max_edu := e + (j*u) + (k*y)
total_finc := r + (w*y) + (q*u)
edu_pay := u
edu_care := y
'
fit2 <- sem(model2,data=city )## Warning in lav_object_post_check(object): lavaan WARNING: some estimated lv
## variances are negativesummary(fit2)## lavaan 0.6-5 ended normally after 114 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 21
##
## Number of observations 136
##
## Model Test User Model:
##
## Test statistic 17.262
## Degrees of freedom 21
## P-value (Chi-square) 0.695
##
## Parameter Estimates:
##
## Information Expected
## Information saturated (h1) model Structured
## Standard errors Standard
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|)
## f1 =~
## tutr_fs_lg (a) 1.000
## schl_fs_lg (s) 0.169 0.094 1.801 0.072
## other_log (d) -0.049 0.124 -0.397 0.691
## f2 =~
## total_care (f) 1.000
## care_edu (g) 6.929 4.809 1.441 0.150
## communicat (h) 6.328 4.369 1.448 0.147
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## f1 ~
## parents_ed (j) 0.249 0.062 3.993 0.000
## f2 ~
## parents_ed (k) 0.008 0.006 1.353 0.176
## f1 ~
## finc (q) 0.808 0.268 3.010 0.003
## f2 ~
## finc (w) 0.001 0.010 0.116 0.908
## edu_expect ~
## parents_ed (e) 0.061 0.033 1.860 0.063
## finc (r) 0.171 0.128 1.336 0.182
## f2 (y) 0.245 1.209 0.203 0.839
## f1 (u) -0.047 0.077 -0.612 0.541
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .tutor_fees_log 15.389 3.869 3.978 0.000
## .school_fees_lg 2.122 0.275 7.711 0.000
## .other_log 9.213 1.118 8.238 0.000
## .total_care 0.564 0.069 8.216 0.000
## .care_edu 0.177 0.111 1.589 0.112
## .communicate 0.196 0.094 2.078 0.038
## .edu_expect 1.548 0.191 8.095 0.000
## .f1 -4.463 3.430 -1.301 0.193
## .f2 0.010 0.013 0.731 0.465
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## total_max_edu 0.051 0.025 2.050 0.040
## total_finc 0.133 0.111 1.200 0.230
## edu_pay -0.047 0.077 -0.612 0.541
## edu_care 0.245 1.209 0.203 0.839parameterEstimates(fit2)#这里的est是没标注准化的系数，还有置信区间## lhs op rhs label est se z
## 1 f1 =~ tutor_fees_log a 1.000 0.000 NA
## 2 f1 =~ school_fees_log s 0.169 0.094 1.801
## 3 f1 =~ other_log d -0.049 0.124 -0.397
## 4 f2 =~ total_care f 1.000 0.000 NA
## 5 f2 =~ care_edu g 6.929 4.809 1.441
## 6 f2 =~ communicate h 6.328 4.369 1.448
## 7 f1 ~ parents_edu j 0.249 0.062 3.993
## 8 f2 ~ parents_edu k 0.008 0.006 1.353
## 9 f1 ~ finc q 0.808 0.268 3.010
## 10 f2 ~ finc w 0.001 0.010 0.116
## 11 edu_expect ~ parents_edu e 0.061 0.033 1.860
## 12 edu_expect ~ finc r 0.171 0.128 1.336
## 13 edu_expect ~ f2 y 0.245 1.209 0.203
## 14 edu_expect ~ f1 u -0.047 0.077 -0.612
## 15 tutor_fees_log ~~ tutor_fees_log 15.389 3.869 3.978
## 16 school_fees_log ~~ school_fees_log 2.122 0.275 7.711
## 17 other_log ~~ other_log 9.213 1.118 8.238
## 18 total_care ~~ total_care 0.564 0.069 8.216
## 19 care_edu ~~ care_edu 0.177 0.111 1.589
## 20 communicate ~~ communicate 0.196 0.094 2.078
## 21 edu_expect ~~ edu_expect 1.548 0.191 8.095
## 22 f1 ~~ f1 -4.463 3.430 -1.301
## 23 f2 ~~ f2 0.010 0.013 0.731
## 24 parents_edu ~~ parents_edu 18.887 0.000 NA
## 25 parents_edu ~~ finc 0.752 0.000 NA
## 26 finc ~~ finc 0.949 0.000 NA
## 27 total_max_edu := e+(j*u)+(k*y) total_max_edu 0.051 0.025 2.050
## 28 total_finc := r+(w*y)+(q*u) total_finc 0.133 0.111 1.200
## 29 edu_pay := u edu_pay -0.047 0.077 -0.612
## 30 edu_care := y edu_care 0.245 1.209 0.203
## pvalue ci.lower ci.upper
## 1 NA 1.000 1.000
## 2 0.072 -0.015 0.353
## 3 0.691 -0.291 0.193
## 4 NA 1.000 1.000
## 5 0.150 -2.497 16.355
## 6 0.147 -2.234 14.890
## 7 0.000 0.127 0.371
## 8 0.176 -0.003 0.018
## 9 0.003 0.282 1.334
## 10 0.908 -0.018 0.020
## 11 0.063 -0.003 0.125
## 12 0.182 -0.080 0.421
## 13 0.839 -2.125 2.616
## 14 0.541 -0.198 0.104
## 15 0.000 7.806 22.972
## 16 0.000 1.583 2.662
## 17 0.000 7.021 11.405
## 18 0.000 0.429 0.698
## 19 0.112 -0.041 0.395
## 20 0.038 0.011 0.380
## 21 0.000 1.173 1.922
## 22 0.193 -11.186 2.259
## 23 0.465 -0.016 0.035
## 24 NA 18.887 18.887
## 25 NA 0.752 0.752
## 26 NA 0.949 0.949
## 27 0.040 0.002 0.100
## 28 0.230 -0.084 0.351
## 29 0.541 -0.198 0.104
## 30 0.839 -2.125 2.616Est <- parameterEstimates(fit2, ci = FALSE, standardized = TRUE)#est为未标准化的，std.all为标准化后的## Warning in sqrt(ETA2): 产生了NaNs## Warning in sqrt(ETA2): 产生了NaNs
## Warning in sqrt(ETA2): 产生了NaNssubset(Est, lhs == "f2")## lhs op rhs label est se z pvalue std.lv std.all
## 4 f2 =~ total_care f 1.000 0.000 NA NA 0.103 0.136
## 5 f2 =~ care_edu g 6.929 4.809 1.441 0.150 0.715 0.862
## 6 f2 =~ communicate h 6.328 4.369 1.448 0.147 0.653 0.828
## 8 f2 ~ parents_edu k 0.008 0.006 1.353 0.176 0.073 0.318
## 10 f2 ~ finc w 0.001 0.010 0.116 0.908 0.011 0.011
## 23 f2 ~~ f2 0.010 0.013 0.731 0.465 0.898 0.898
## std.nox
## 4 0.136
## 5 0.862
## 6 0.828
## 8 0.073
## 10 0.011
## 23 0.898fitMeasures(fit2, c("cfi", "rmsea","nfi","srmr"))## cfi rmsea nfi srmr
## 1.000 0.000 0.897 0.057semPaths(fit2,whatLabels="std.all",style="lisrel",layout = "tree2",
residuals=TRUE)## Warning in sqrt(ETA2): 产生了NaNs
## Warning in sqrt(ETA2): 产生了NaNs
## Warning in sqrt(ETA2): 产生了NaNs
农村的模型
model2 <- '
# latent variables
f1 =~ a*tutor_fees_log + s*school_fees_log + d*other_log
f2 =~ f*total_care + g*care_edu + h*communicate
# regressions
f1 ~ j*parents_edu
f2 ~ k*parents_edu
f1 ~ q*finc
f2 ~ w*finc
edu_expect~ e*parents_edu
edu_expect~ r*finc
#f1 ~ t*edu_expect
edu_expect ~ y*f2
edu_expect ~ u*f1
# residual covariances
# max_edu ~~ finc_per
#total
total_max_edu := e + (j*u) + (k*y)
total_finc := r + (w*y) + (q*u)
edu_pay := u
edu_care := y
'
fit2 <- sem(model2,data=countryside )
summary(fit2)## lavaan 0.6-5 ended normally after 80 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 21
##
## Number of observations 1006
##
## Model Test User Model:
##
## Test statistic 66.298
## Degrees of freedom 21
## P-value (Chi-square) 0.000
##
## Parameter Estimates:
##
## Information Expected
## Information saturated (h1) model Structured
## Standard errors Standard
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|)
## f1 =~
## tutr_fs_lg (a) 1.000
## schl_fs_lg (s) 0.250 0.117 2.138 0.033
## other_log (d) 0.484 0.225 2.151 0.031
## f2 =~
## total_care (f) 1.000
## care_edu (g) 4.792 0.881 5.437 0.000
## communicat (h) 3.978 0.667 5.966 0.000
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## f1 ~
## parents_ed (j) 0.019 0.019 1.006 0.314
## f2 ~
## parents_ed (k) 0.002 0.002 1.585 0.113
## f1 ~
## finc (q) 0.123 0.052 2.376 0.018
## f2 ~
## finc (w) 0.006 0.004 1.404 0.160
## edu_expect ~
## parents_ed (e) 0.015 0.011 1.329 0.184
## finc (r) -0.031 0.033 -0.936 0.349
## f2 (y) 1.233 0.349 3.534 0.000
## f1 (u) 0.162 0.088 1.845 0.065
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .tutor_fees_log 3.946 0.712 5.543 0.000
## .school_fees_lg 1.983 0.101 19.641 0.000
## .other_log 6.805 0.352 19.315 0.000
## .total_care 0.536 0.024 22.202 0.000
## .care_edu 0.146 0.065 2.243 0.025
## .communicate 0.278 0.046 6.017 0.000
## .edu_expect 1.631 0.078 21.024 0.000
## .f1 1.447 0.704 2.056 0.040
## .f2 0.024 0.008 3.016 0.003
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## total_max_edu 0.021 0.011 1.880 0.060
## total_finc -0.003 0.031 -0.111 0.912
## edu_pay 0.162 0.088 1.845 0.065
## edu_care 1.233 0.349 3.534 0.000parameterEstimates(fit2)#这里的est是没标注准化的系数，还有置信区间## lhs op rhs label est se z
## 1 f1 =~ tutor_fees_log a 1.000 0.000 NA
## 2 f1 =~ school_fees_log s 0.250 0.117 2.138
## 3 f1 =~ other_log d 0.484 0.225 2.151
## 4 f2 =~ total_care f 1.000 0.000 NA
## 5 f2 =~ care_edu g 4.792 0.881 5.437
## 6 f2 =~ communicate h 3.978 0.667 5.966
## 7 f1 ~ parents_edu j 0.019 0.019 1.006
## 8 f2 ~ parents_edu k 0.002 0.002 1.585
## 9 f1 ~ finc q 0.123 0.052 2.376
## 10 f2 ~ finc w 0.006 0.004 1.404
## 11 edu_expect ~ parents_edu e 0.015 0.011 1.329
## 12 edu_expect ~ finc r -0.031 0.033 -0.936
## 13 edu_expect ~ f2 y 1.233 0.349 3.534
## 14 edu_expect ~ f1 u 0.162 0.088 1.845
## 15 tutor_fees_log ~~ tutor_fees_log 3.946 0.712 5.543
## 16 school_fees_log ~~ school_fees_log 1.983 0.101 19.641
## 17 other_log ~~ other_log 6.805 0.352 19.315
## 18 total_care ~~ total_care 0.536 0.024 22.202
## 19 care_edu ~~ care_edu 0.146 0.065 2.243
## 20 communicate ~~ communicate 0.278 0.046 6.017
## 21 edu_expect ~~ edu_expect 1.631 0.078 21.024
## 22 f1 ~~ f1 1.447 0.704 2.056
## 23 f2 ~~ f2 0.024 0.008 3.016
## 24 parents_edu ~~ parents_edu 13.456 0.000 NA
## 25 parents_edu ~~ finc 0.336 0.000 NA
## 26 finc ~~ finc 1.775 0.000 NA
## 27 total_max_edu := e+(j*u)+(k*y) total_max_edu 0.021 0.011 1.880
## 28 total_finc := r+(w*y)+(q*u) total_finc -0.003 0.031 -0.111
## 29 edu_pay := u edu_pay 0.162 0.088 1.845
## 30 edu_care := y edu_care 1.233 0.349 3.534
## pvalue ci.lower ci.upper
## 1 NA 1.000 1.000
## 2 0.033 0.021 0.479
## 3 0.031 0.043 0.924
## 4 NA 1.000 1.000
## 5 0.000 3.065 6.519
## 6 0.000 2.671 5.286
## 7 0.314 -0.018 0.055
## 8 0.113 -0.001 0.005
## 9 0.018 0.022 0.225
## 10 0.160 -0.002 0.014
## 11 0.184 -0.007 0.038
## 12 0.349 -0.095 0.033
## 13 0.000 0.549 1.916
## 14 0.065 -0.010 0.334
## 15 0.000 2.550 5.341
## 16 0.000 1.785 2.181
## 17 0.000 6.114 7.496
## 18 0.000 0.488 0.583
## 19 0.025 0.018 0.273
## 20 0.000 0.188 0.369
## 21 0.000 1.479 1.784
## 22 0.040 0.068 2.827
## 23 0.003 0.008 0.040
## 24 NA 13.456 13.456
## 25 NA 0.336 0.336
## 26 NA 1.775 1.775
## 27 0.060 -0.001 0.043
## 28 0.912 -0.064 0.057
## 29 0.065 -0.010 0.334
## 30 0.000 0.549 1.916Est <- parameterEstimates(fit2, ci = FALSE, standardized = TRUE)#est为未标准化的，std.all为标准化后的
subset(Est, lhs == "f2")## lhs op rhs label est se z pvalue std.lv std.all
## 4 f2 =~ total_care f 1.000 0.000 NA NA 0.156 0.208
## 5 f2 =~ care_edu g 4.792 0.881 5.437 0.000 0.745 0.890
## 6 f2 =~ communicate h 3.978 0.667 5.966 0.000 0.619 0.761
## 8 f2 ~ parents_edu k 0.002 0.002 1.585 0.113 0.015 0.057
## 10 f2 ~ finc w 0.006 0.004 1.404 0.160 0.037 0.050
## 23 f2 ~~ f2 0.024 0.008 3.016 0.003 0.994 0.994
## std.nox
## 4 0.208
## 5 0.890
## 6 0.761
## 8 0.015
## 10 0.037
## 23 0.994fitMeasures(fit2, c("cfi", "rmsea","nfi","srmr"))## cfi rmsea nfi srmr
## 0.940 0.046 0.916 0.044semPaths(fit2,whatLabels="std.all",style="lisrel",layout = "tree2",
residuals=TRUE)