Guncel_Odev_5_Part_II
Onur Aktepe
2025-12-12
load("data/ASGSGPR5.Rdata")
load("data/ASASGPR5.Rdata")
load("data/ASATURR5.Rdata")
load("data/ASGTURR5.Rdata")library(haven)
library(dplyr)
library(tidyverse)
library(GGally)
library(psych)Veri Setlerinin Kendi İçinde Birleşitirilmesi
tur <- bind_rows(ASATURR5, ASGTURR5)
sgpr<- bind_rows(ASASGPR5, ASGSGPR5)Türkiye Veri Seti
pirls_tur<- tur %>%
select(IDCNTRY, IDSCHOOL, IDSTUD, TOTWGT,
ASRREA01:ASRREA05, ASBG01, ASBG04, ASBG05A)Türkiye Veri Seti Değişken Adlarının Türkçeye Çevirilmesi
pirls_tur1 <- pirls_tur%>% mutate(ulke= IDCNTRY, cinsiyet=ASBG01, teknolojik_imkan=ASBG05A, kitap_sayisi= ASBG04)pirls_tur2<-pirls_tur1 %>% mutate(cinsiyet = case_when(
cinsiyet == 1 ~ "Erkek",
cinsiyet == 2 ~ "Kadin",
TRUE ~ NA_character_
), ulke= case_when(
ulke==792~ "Turkiye",
TRUE ~ NA_character_
))Singapur Veri Seti
pirls_sgpr<- sgpr %>%
select(IDCNTRY, IDSCHOOL, IDSTUD, TOTWGT,
ASRREA01:ASRREA05, ASBG01, ASBG04, ASBG05A)Singapur Veri Seti Değişkenklerinin Adlarının Türkçeye Çevirilmesi
pirls_sgpr1 <- pirls_sgpr%>% mutate(ulke= IDCNTRY, cinsiyet=ASBG01, teknolojik_imkan=ASBG05A, kitap_sayisi= ASBG04)pirls_sgpr2<-pirls_sgpr1 %>% mutate(cinsiyet = case_when(
cinsiyet == 1 ~ "Erkek",
cinsiyet == 2 ~ "Kadin",
TRUE ~ NA_character_
), ulke= case_when(
ulke==702~ "Singapur",
TRUE ~ NA_character_
))Boş Verilerin Temizlenmesi
pirls_tur3<- na.omit(pirls_tur2)pirls_sgpr3<- na.omit(pirls_sgpr2)pirls_tur3 <- pirls_tur3 %>%
mutate(across(where(is.labelled), as.numeric))
pirls_sgpr3 <- pirls_sgpr3 %>%
mutate(across(where(is.labelled), as.numeric))Veri Setine Yönelik İstatistiklerin İncelenmesi
describe(pirls_tur3 %>% select(cinsiyet, ulke, kitap_sayisi, teknolojik_imkan, ASRREA01, ASRREA02, ASRREA03, ASRREA04, ASRREA05))## vars n mean sd median trimmed mad min max
## cinsiyet* 1 5820 1.50 0.50 2.00 1.50 0.00 1.00 2.00
## ulke* 2 5820 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## kitap_sayisi 3 5820 2.76 1.11 3.00 2.71 1.48 1.00 5.00
## teknolojik_imkan 4 5820 1.28 0.45 1.00 1.22 0.00 1.00 2.00
## ASRREA01 5 5820 506.42 84.06 513.54 509.67 81.85 187.78 777.70
## ASRREA02 6 5820 505.30 84.64 511.00 508.91 82.47 126.61 779.79
## ASRREA03 7 5820 504.55 85.52 511.50 507.71 83.43 166.11 785.54
## ASRREA04 8 5820 503.76 84.08 509.39 506.98 82.19 191.64 756.25
## ASRREA05 9 5820 504.38 84.96 511.32 507.50 82.82 160.51 816.56
## range skew kurtosis se
## cinsiyet* 1.00 -0.01 -2.00 0.01
## ulke* 0.00 NaN NaN 0.00
## kitap_sayisi 4.00 0.39 -0.40 0.01
## teknolojik_imkan 1.00 0.98 -1.04 0.01
## ASRREA01 589.93 -0.39 0.15 1.10
## ASRREA02 653.18 -0.43 0.31 1.11
## ASRREA03 619.43 -0.35 0.12 1.12
## ASRREA04 564.61 -0.38 0.17 1.10
## ASRREA05 656.05 -0.35 0.17 1.11describe(pirls_sgpr3 %>% select(cinsiyet, ulke, kitap_sayisi, teknolojik_imkan, ASRREA01, ASRREA02, ASRREA03, ASRREA04, ASRREA05))## vars n mean sd median trimmed mad min max
## cinsiyet* 1 6633 1.51 0.50 2.00 1.51 0.00 1.00 2.00
## ulke* 2 6633 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## kitap_sayisi 3 6633 2.97 1.20 3.00 2.96 1.48 1.00 5.00
## teknolojik_imkan 4 6633 1.25 0.43 1.00 1.19 0.00 1.00 2.00
## ASRREA01 5 6633 584.31 85.29 594.60 589.72 77.52 203.34 830.63
## ASRREA02 6 6633 584.48 86.24 595.51 589.96 78.99 220.32 835.43
## ASRREA03 7 6633 582.75 85.77 592.70 588.14 77.52 236.20 829.98
## ASRREA04 8 6633 583.58 86.24 594.10 589.06 79.10 216.78 862.93
## ASRREA05 9 6633 582.75 85.44 594.05 588.23 76.79 182.15 855.14
## range skew kurtosis se
## cinsiyet* 1.00 -0.04 -2.00 0.01
## ulke* 0.00 NaN NaN 0.00
## kitap_sayisi 4.00 0.10 -0.79 0.01
## teknolojik_imkan 1.00 1.15 -0.67 0.01
## ASRREA01 627.29 -0.65 0.60 1.05
## ASRREA02 615.12 -0.63 0.54 1.06
## ASRREA03 593.77 -0.62 0.56 1.05
## ASRREA04 646.15 -0.62 0.57 1.06
## ASRREA05 672.99 -0.65 0.61 1.05library(psych)
ggpairs(pirls_tur3 %>% select(cinsiyet, ulke, kitap_sayisi, teknolojik_imkan, ASRREA01,))
ggpairs(pirls_sgpr3 %>% select(cinsiyet, ulke, kitap_sayisi, teknolojik_imkan, ASRREA01,))
Aykırı Değerler, Kaldıraç Etkisi ve Etkileşim Değerlerinin Hesaplanması
reg_tur <- lm(ASRREA01 ~ kitap_sayisi + teknolojik_imkan + cinsiyet, data = pirls_tur3)
reg_tur##
## Call:
## lm(formula = ASRREA01 ~ kitap_sayisi + teknolojik_imkan + cinsiyet,
## data = pirls_tur3)
##
## Coefficients:
## (Intercept) kitap_sayisi teknolojik_imkan cinsiyetKadin
## 487.37 20.72 -24.50 -13.63pirls_tur3$std_resid <- rstandard(reg_tur)
pirls_tur3$leverage <- hatvalues(reg_tur)
pirls_tur3$cooksD <- cooks.distance(reg_tur)
head(pirls_tur3)## # A tibble: 6 × 19
## IDCNTRY IDSCHOOL IDSTUD TOTWGT ASRREA01 ASRREA02 ASRREA03 ASRREA04 ASRREA05
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 792 5001 50010601 64.1 575. 557. 562. 574. 567.
## 2 792 5001 50010602 64.1 624. 613. 545. 504. 557.
## 3 792 5001 50010603 64.1 567. 569. 533. 530. 575.
## 4 792 5001 50010604 64.1 559. 607. 619. 537. 578.
## 5 792 5001 50010605 64.1 570. 570. 578. 581. 564.
## 6 792 5001 50010607 64.1 599. 595. 536. 546. 568.
## # ℹ 10 more variables: ASBG01 <dbl>, ASBG04 <dbl>, ASBG05A <dbl>, ulke <chr>,
## # cinsiyet <chr>, teknolojik_imkan <dbl>, kitap_sayisi <dbl>,
## # std_resid <dbl>, leverage <dbl>, cooksD <dbl>library(ggfortify)
autoplot(reg_tur)
reg_sgpr<- lm(ASRREA01 ~ kitap_sayisi + teknolojik_imkan + cinsiyet, data = pirls_sgpr3)
reg_sgpr##
## Call:
## lm(formula = ASRREA01 ~ kitap_sayisi + teknolojik_imkan + cinsiyet,
## data = pirls_sgpr3)
##
## Coefficients:
## (Intercept) kitap_sayisi teknolojik_imkan cinsiyetKadin
## 549.85 20.92 -16.76 -13.26pirls_sgpr3$std_resid <- rstandard(reg_sgpr)
pirls_sgpr3$leverage <- hatvalues(reg_sgpr)
pirls_sgpr3$cooksD <- cooks.distance(reg_sgpr)
head(pirls_sgpr3)## # A tibble: 6 × 19
## IDCNTRY IDSCHOOL IDSTUD TOTWGT ASRREA01 ASRREA02 ASRREA03 ASRREA04 ASRREA05
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 702 5001 50010201 4.85 499. 487. 487. 468. 520.
## 2 702 5001 50010202 4.85 579. 651. 631. 644. 625.
## 3 702 5001 50010203 4.85 604. 618. 591. 590. 620.
## 4 702 5001 50010204 4.85 542. 498. 490. 539. 537.
## 5 702 5001 50010205 4.85 531. 540. 549. 557. 540.
## 6 702 5001 50010206 4.85 540. 576. 598. 559. 538.
## # ℹ 10 more variables: ASBG01 <dbl>, ASBG04 <dbl>, ASBG05A <dbl>, ulke <chr>,
## # cinsiyet <chr>, teknolojik_imkan <dbl>, kitap_sayisi <dbl>,
## # std_resid <dbl>, leverage <dbl>, cooksD <dbl>library(ggfortify)
autoplot(reg_sgpr)
Moderasyon Analizinin Yapılması
moderation_tur <- lm(kitap_sayisi~teknolojik_imkan*cinsiyet, data = pirls_tur3)
summary(moderation_tur)##
## Call:
## lm(formula = kitap_sayisi ~ teknolojik_imkan * cinsiyet, data = pirls_tur3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.92429 -0.85953 0.07571 0.60175 2.60175
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.45032 0.06125 56.335 <2e-16 ***
## teknolojik_imkan -0.52603 0.04530 -11.613 <2e-16 ***
## cinsiyetKadin -0.19251 0.08640 -2.228 0.0259 *
## teknolojik_imkan:cinsiyetKadin 0.12776 0.06372 2.005 0.0450 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.091 on 5816 degrees of freedom
## Multiple R-squared: 0.03567, Adjusted R-squared: 0.03517
## F-statistic: 71.7 on 3 and 5816 DF, p-value: < 2.2e-16Türkiye
library(mediation)
med_1 <- lm(kitap_sayisi ~ teknolojik_imkan, data = pirls_tur3)
med_2 <- lm(ASRREA01 ~ teknolojik_imkan + kitap_sayisi, data = pirls_tur3)
med_tur <- mediate(
model.m = med_1,
model.y = med_2,
sims = 1000,
boot = TRUE,
treat = "teknolojik_imkan",
mediator = "kitap_sayisi"
)
summary(med_tur)##
## Causal Mediation Analysis
##
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
##
## Estimate 95% CI Lower 95% CI Upper p-value
## ACME -9.60383 -11.23016 -8.18166 < 2.2e-16 ***
## ADE -24.58231 -28.99407 -19.47280 < 2.2e-16 ***
## Total Effect -34.18615 -38.64404 -28.94849 < 2.2e-16 ***
## Prop. Mediated 0.28093 0.23518 0.34276 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Sample Size Used: 5820
##
##
## Simulations: 1000Singapur
moderation_sgpr <- lm(kitap_sayisi~teknolojik_imkan*cinsiyet, data = pirls_sgpr3)
summary(moderation_sgpr)##
## Call:
## lm(formula = kitap_sayisi ~ teknolojik_imkan * cinsiyet, data = pirls_sgpr3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.07539 -0.96676 0.03324 0.92461 2.19912
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.30811 0.06477 51.071 < 2e-16 ***
## teknolojik_imkan -0.23272 0.04981 -4.672 3.05e-06 ***
## cinsiyetKadin -0.17547 0.08997 -1.950 0.0512 .
## teknolojik_imkan:cinsiyetKadin 0.06684 0.06810 0.981 0.3264
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.197 on 6629 degrees of freedom
## Multiple R-squared: 0.006904, Adjusted R-squared: 0.006455
## F-statistic: 15.36 on 3 and 6629 DF, p-value: 5.866e-10library(mediation)
med_1 <- lm(kitap_sayisi ~ teknolojik_imkan, data = pirls_tur3)
med_2 <- lm(ASRREA01 ~ teknolojik_imkan + kitap_sayisi, data = pirls_tur3)
med_sgpr <- mediate(
model.m = med_1,
model.y = med_2,
sims = 1000,
boot = TRUE,
treat = "teknolojik_imkan",
mediator = "kitap_sayisi"
)
summary(med_sgpr)##
## Causal Mediation Analysis
##
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
##
## Estimate 95% CI Lower 95% CI Upper p-value
## ACME -9.60383 -11.11628 -8.14691 < 2.2e-16 ***
## ADE -24.58231 -29.21187 -19.94505 < 2.2e-16 ***
## Total Effect -34.18615 -38.96404 -29.48468 < 2.2e-16 ***
## Prop. Mediated 0.28093 0.23170 0.33304 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Sample Size Used: 5820
##
##
## Simulations: 1000