Regression PASCHPOL MATHS
# Linear regression analysis of Exam result compared with parental involvement
# Before any linear regression analysis, it is important to make a visualization with plot()
setwd("C:/Users/antho/Downloads/r final project")
library(flexdashboard)
## Warning: package 'flexdashboard' was built under R version 4.3.1
library(shiny)
pisair <- readRDS("PISA_2018_IRL.rds")
# Regression Analysis for PV1MATH
plot(pisair$PASCHPOL, pisair$PV1MATH,
main = "Regression Analysis of Exam Result vs Parental Involvement (PV1MATH)",
xlab = "Parental Involvement",
ylab = "PV1MATH")
lm_math <- lm(pisair$PV1MATH ~ pisair$PASCHPOL)
summary(lm_math)
##
## Call:
## lm(formula = pisair$PV1MATH ~ pisair$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -307.460 -51.715 2.008 54.663 237.424
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 503.516 1.114 451.850 <2e-16 ***
## pisair$PASCHPOL -2.818 1.153 -2.445 0.0145 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 77.14 on 4926 degrees of freedom
## (649 observations deleted due to missingness)
## Multiple R-squared: 0.001212, Adjusted R-squared: 0.001009
## F-statistic: 5.977 on 1 and 4926 DF, p-value: 0.01453
boxplot(pisair$PV1MATH, main = "PV1MATH Results")
# Regression Analysis for PV1SCIE
plot(pisair$PASCHPOL, pisair$PV1SCIE,
main = "Regression Analysis of Exam Result vs Parental Involvement (PV1SCIE)",
xlab = "Parental Involvement",
ylab = "PV1SCIE")
lm_scie <- lm(pisair$PV1SCIE ~ pisair$PASCHPOL)
summary(lm_scie)
##
## Call:
## lm(formula = pisair$PV1SCIE ~ pisair$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -324.28 -60.07 1.16 61.28 280.21
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 499.779 1.271 393.119 < 2e-16 ***
## pisair$PASCHPOL -3.548 1.315 -2.697 0.00701 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 88.01 on 4926 degrees of freedom
## (649 observations deleted due to missingness)
## Multiple R-squared: 0.001475, Adjusted R-squared: 0.001272
## F-statistic: 7.276 on 1 and 4926 DF, p-value: 0.007013
boxplot(pisair$PV1SCIE, main = "PV1SCIE Results")
# Regression Analysis for PV1READ
plot(pisair$PASCHPOL, pisair$PV1READ,
main = "Regression Analysis of Exam Result vs Parental Involvement (PV1READ)",
xlab = "Parental Involvement",
ylab = "PV1READ")
lm_read <- lm(pisair$PV1READ ~ pisair$PASCHPOL)
summary(lm_read)
##
## Call:
## lm(formula = pisair$PV1READ ~ pisair$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -317.18 -59.33 2.46 63.96 279.67
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 523.193 1.290 405.654 < 2e-16 ***
## pisair$PASCHPOL -5.334 1.334 -3.998 6.49e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 89.28 on 4926 degrees of freedom
## (649 observations deleted due to missingness)
## Multiple R-squared: 0.003234, Adjusted R-squared: 0.003032
## F-statistic: 15.98 on 1 and 4926 DF, p-value: 6.489e-05
boxplot(pisair$PV1READ, main = "PV1READ Results")
# Gender Subset Analysis for PV1READ
subset_male <- subset(pisair, ST004D01T == '2')
subset_female <- subset(pisair, ST004D01T == '1')
lm(subset_male$PV1READ ~ subset_male$PASCHPOL)
##
## Call:
## lm(formula = subset_male$PV1READ ~ subset_male$PASCHPOL)
##
## Coefficients:
## (Intercept) subset_male$PASCHPOL
## 512.13 -3.47
lm(subset_female$PV1READ ~ subset_female$PASCHPOL)
##
## Call:
## lm(formula = subset_female$PV1READ ~ subset_female$PASCHPOL)
##
## Coefficients:
## (Intercept) subset_female$PASCHPOL
## 533.57 -6.15
# ... (continue with other analyses)
# Fit linear regression model with PASCHPOL, PV1READ, PV1SCIE, PV1MATH, and ESCS
model_ESCS <- lm(pisair$PASCHPOL ~ pisair$PV1MATH + pisair$ESCS)
# Summary of the model
summary(model_ESCS)
##
## Call:
## lm(formula = pisair$PASCHPOL ~ pisair$PV1MATH + pisair$ESCS)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.3398 -0.6515 -0.0280 0.5693 2.4169
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.2830126 0.0938771 3.015 0.00259 **
## pisair$PV1MATH -0.0002282 0.0001861 -1.226 0.22020
## pisair$ESCS -0.0543856 0.0169881 -3.201 0.00138 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9521 on 4883 degrees of freedom
## (691 observations deleted due to missingness)
## Multiple R-squared: 0.003238, Adjusted R-squared: 0.002829
## F-statistic: 7.93 on 2 and 4883 DF, p-value: 0.0003643
# ... (continue with other analyses)
# ... (previous code)
# Filter PASCHPOL by socio-economic status (e.g., high and low ESCS)
high_ESCS_data <- subset(pisair, ESCS >= 0.5693)
low_ESCS_data <- subset(pisair, ESCS <= -0.6515)
# Fit separate models for high and low socio-economic status
high_ESCS_model <- lm(high_ESCS_data$PV1MATH ~ high_ESCS_data$PASCHPOL)
low_ESCS_model <- lm(low_ESCS_data$PV1MATH ~ low_ESCS_data$PASCHPOL)
# Summary of the models for high and low socio-economic status
summary(high_ESCS_model)
##
## Call:
## lm(formula = high_ESCS_data$PV1MATH ~ high_ESCS_data$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -269.217 -47.782 2.473 50.649 209.354
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 530.945 1.760 301.63 <2e-16 ***
## high_ESCS_data$PASCHPOL -1.365 1.869 -0.73 0.465
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 72.72 on 1735 degrees of freedom
## (189 observations deleted due to missingness)
## Multiple R-squared: 0.0003074, Adjusted R-squared: -0.0002688
## F-statistic: 0.5334 on 1 and 1735 DF, p-value: 0.4653
summary(low_ESCS_model)
##
## Call:
## lm(formula = low_ESCS_data$PV1MATH ~ low_ESCS_data$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -239.021 -53.950 -0.363 51.825 224.521
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 468.757 2.624 178.628 <2e-16 ***
## low_ESCS_data$PASCHPOL -3.181 2.561 -1.242 0.214
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 76.89 on 917 degrees of freedom
## (179 observations deleted due to missingness)
## Multiple R-squared: 0.00168, Adjusted R-squared: 0.0005917
## F-statistic: 1.543 on 1 and 917 DF, p-value: 0.2144
# Fit linear regression model with PASCHPOL, PV1SCIE, and ESCS
modelScience_ESCS <- lm(pisair$PASCHPOL ~ pisair$PV1SCIE + pisair$ESCS)
# Summary of the model
summary(modelScience_ESCS)
##
## Call:
## lm(formula = pisair$PASCHPOL ~ pisair$PV1SCIE + pisair$ESCS)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.3569 -0.6503 -0.0313 0.5765 2.4152
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.2866259 0.0817200 3.507 0.000457 ***
## pisair$PV1SCIE -0.0002374 0.0001628 -1.458 0.144780
## pisair$ESCS -0.0534523 0.0169178 -3.160 0.001590 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9521 on 4883 degrees of freedom
## (691 observations deleted due to missingness)
## Multiple R-squared: 0.003365, Adjusted R-squared: 0.002957
## F-statistic: 8.243 on 2 and 4883 DF, p-value: 0.0002667
# Filter PASCHPOL by socio-economic status for science
high_ESCS_dataScience <- subset(pisair, ESCS >= 0.5765)
low_ESCS_dataScience <- subset(pisair, ESCS <= -0.6503)
# Fit separate models for high and low socio-economic status for science
high_ESCS_modelScience <- lm(high_ESCS_dataScience$PV1SCIE ~ high_ESCS_dataScience$PASCHPOL)
low_ESCS_modelScience <- lm(low_ESCS_dataScience$PV1SCIE ~ low_ESCS_dataScience$PASCHPOL)
# Summary of the models for science
summary(high_ESCS_modelScience)
##
## Call:
## lm(formula = high_ESCS_dataScience$PV1SCIE ~ high_ESCS_dataScience$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -270.923 -54.507 5.691 60.044 250.174
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 530.037 2.067 256.381 <2e-16 ***
## high_ESCS_dataScience$PASCHPOL -3.006 2.199 -1.367 0.172
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 85.08 on 1720 degrees of freedom
## (186 observations deleted due to missingness)
## Multiple R-squared: 0.001086, Adjusted R-squared: 0.0005049
## F-statistic: 1.869 on 1 and 1720 DF, p-value: 0.1717
summary(low_ESCS_modelScience)
##
## Call:
## lm(formula = low_ESCS_dataScience$PV1SCIE ~ low_ESCS_dataScience$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -286.070 -62.200 -0.061 57.947 225.951
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 461.623 2.962 155.862 <2e-16 ***
## low_ESCS_dataScience$PASCHPOL -3.803 2.883 -1.319 0.187
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 86.99 on 919 degrees of freedom
## (179 observations deleted due to missingness)
## Multiple R-squared: 0.00189, Adjusted R-squared: 0.000804
## F-statistic: 1.74 on 1 and 919 DF, p-value: 0.1874
# Fit linear regression model with PASCHPOL, PV1READ, and ESCS
model_ESCSReading <- lm(pisair$PASCHPOL ~ pisair$PV1READ + pisair$ESCS)
# Summary of the model
summary(model_ESCSReading)
##
## Call:
## lm(formula = pisair$PASCHPOL ~ pisair$PV1READ + pisair$ESCS)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.4044 -0.6559 -0.0373 0.5679 2.4279
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.4053733 0.0840988 4.820 1.48e-06 ***
## pisair$PV1READ -0.0004559 0.0001601 -2.847 0.00442 **
## pisair$ESCS -0.0465304 0.0168781 -2.757 0.00586 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9515 on 4883 degrees of freedom
## (691 observations deleted due to missingness)
## Multiple R-squared: 0.004584, Adjusted R-squared: 0.004176
## F-statistic: 11.24 on 2 and 4883 DF, p-value: 1.345e-05
# Filter PASCHPOL by socio-economic status for reading
high_ESCS_dataReading <- subset(pisair, ESCS >= 0.5679)
low_ESCS_dataReading <- subset(pisair, ESCS <= -0.6559)
# Fit separate models for high and low socio-economic status for reading
high_ESCS_modelReading <- lm(high_ESCS_dataReading$PV1READ ~ high_ESCS_dataReading$PASCHPOL)
low_ESCS_modelReading <- lm(low_ESCS_dataReading$PV1READ ~ low_ESCS_dataReading$PASCHPOL)
# Summary of the models for reading
summary(high_ESCS_modelReading)
##
## Call:
## lm(formula = high_ESCS_dataReading$PV1READ ~ high_ESCS_dataReading$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -310.647 -55.068 3.861 61.095 224.312
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 552.924 2.073 266.704 <2e-16 ***
## high_ESCS_dataReading$PASCHPOL -4.246 2.202 -1.928 0.054 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 85.7 on 1737 degrees of freedom
## (189 observations deleted due to missingness)
## Multiple R-squared: 0.002137, Adjusted R-squared: 0.001562
## F-statistic: 3.719 on 1 and 1737 DF, p-value: 0.05396
summary(low_ESCS_modelReading)
##
## Call:
## lm(formula = low_ESCS_dataReading$PV1READ ~ low_ESCS_dataReading$PASCHPOL)
##
## Residuals:
## Min 1Q Median 3Q Max
## -262.114 -63.109 1.368 61.386 253.648
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 484.974 3.009 161.186 <2e-16 ***
## low_ESCS_dataReading$PASCHPOL -6.335 2.933 -2.159 0.0311 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 87.97 on 912 degrees of freedom
## (179 observations deleted due to missingness)
## Multiple R-squared: 0.005087, Adjusted R-squared: 0.003996
## F-statistic: 4.663 on 1 and 912 DF, p-value: 0.03108
# ... (continue with other analyses)
# Construction of new index 1 (ENGAGE), measuring parents' initiative to engage in schools
# $PA008Q01TA: Discussed my child’s behavior with a teacher on my initiative.
# $PA008Q03TA: Discussed my child’s progress with a teacher on my initiative.
# $PA008Q09NA: Talked about how to support learning at home and homework with my child’s teachers.
# $PA008Q10NA: Exchanged ideas on parenting, family support, or the child’s development with my child’s teachers.
# Performance of linear regressions
pisair$ENGAGE <- pisair$PA008Q01TA + pisair$PA008Q03TA + pisair$PA008Q09NA + pisair$PA008Q10NA
lm(pisair$PV1MATH ~ pisair$ENGAGE)
##
## Call:
## lm(formula = pisair$PV1MATH ~ pisair$ENGAGE)
##
## Coefficients:
## (Intercept) pisair$ENGAGE
## 452.700 7.848
summary(lm(pisair$PV1MATH ~ pisair$ENGAGE))
##
## Call:
## lm(formula = pisair$PV1MATH ~ pisair$ENGAGE)
##
## Residuals:
## Min 1Q Median 3Q Max
## -307.912 -51.444 1.891 53.784 224.216
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 452.700 4.936 91.71 <2e-16 ***
## pisair$ENGAGE 7.848 0.739 10.62 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 76.26 on 4815 degrees of freedom
## (760 observations deleted due to missingness)
## Multiple R-squared: 0.02288, Adjusted R-squared: 0.02268
## F-statistic: 112.8 on 1 and 4815 DF, p-value: < 2.2e-16
plot(pisair$ENGAGE, pisair$PV1MATH,
main = "Linear Regression of PV1MATH vs Parental Engagement",
xlab = "Parental Engagement",
ylab = "PV1MATH")
abline(lm(pisair$PV1MATH ~ pisair$ENGAGE), col = "red")
lm(pisair$PV1READ ~ pisair$ENGAGE)
##
## Call:
## lm(formula = pisair$PV1READ ~ pisair$ENGAGE)
##
## Coefficients:
## (Intercept) pisair$ENGAGE
## 465.610 8.848
lm(pisair$PV1SCIE ~ pisair$ENGAGE)
##
## Call:
## lm(formula = pisair$PV1SCIE ~ pisair$ENGAGE)
##
## Coefficients:
## (Intercept) pisair$ENGAGE
## 449.449 7.765
# ... (continue with other analyses)
# $PA009Q06NA: I felt unwelcome at my child’s school.
# $PA009Q07NA: I feel generally awkward in a school.
# $PA009Q09NA: I think participation is not relevant for my child’s development.
# $PA009Q10NA: I do not know how I could participate in school activities.
# $PA009Q11NA: My child does not want me to participate.
# Performance of linear regressions
pisair$HINDER <- pisair$PA009Q06NA + pisair$PA009Q07NA + pisair$PA009Q09NA + pisair$PA009Q10NA + pisair$PA009Q11NA
lm(pisair$PV1MATH ~ pisair$HINDER)
##
## Call:
## lm(formula = pisair$PV1MATH ~ pisair$HINDER)
##
## Coefficients:
## (Intercept) pisair$HINDER
## 428.013 7.873
summary(lm(pisair$PV1MATH ~ pisair$HINDER))
##
## Call:
## lm(formula = pisair$PV1MATH ~ pisair$HINDER)
##
## Residuals:
## Min 1Q Median 3Q Max
## -314.87 -52.62 2.40 53.65 232.95
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 428.013 14.171 30.204 < 2e-16 ***
## pisair$HINDER 7.873 1.469 5.359 8.78e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 77 on 4807 degrees of freedom
## (768 observations deleted due to missingness)
## Multiple R-squared: 0.005938, Adjusted R-squared: 0.005731
## F-statistic: 28.71 on 1 and 4807 DF, p-value: 8.779e-08
plot(pisair$HINDER, pisair$PV1MATH,
main = "Linear Regression of PV1MATH vs Parental Hindrance",
xlab = "Parental Hindrance",
ylab = "PV1MATH")
abline(lm(pisair$PV1MATH ~ pisair$HINDER), col = "red")
lm(pisair$PV1READ ~ pisair$HINDER)
##
## Call:
## lm(formula = pisair$PV1READ ~ pisair$HINDER)
##
## Coefficients:
## (Intercept) pisair$HINDER
## 441.812 8.488
lm(pisair$PV1SCIE ~ pisair$HINDER)
##
## Call:
## lm(formula = pisair$PV1SCIE ~ pisair$HINDER)
##
## Coefficients:
## (Intercept) pisair$HINDER
## 424.590 7.862
# ... (continue with other analyses)