## 準備
setwd("C:/Users/koyo/Dropbox/R/140527_Kaken_FB")
library(xlsx)
## データ作成 もとのデータの*を空白に置換しておく（~*）
sho_a <- read.xlsx("sho_a.xlsx", sheetName = "sho_a")
### 必要なデータだけを取り出す
koku <- sho_a[c("schl", "sc_g_c", "Q1", "Q2", "Q31S1", "Q31S2", "Q32S1", "Q32S2",
"Q33S1", "Q33S2", "Q34S1", "Q34S2", "Q35S1", "Q35S2", "Q37", "Q38")]
### 学級別学級規模データの読み込み
cs <- read.xlsx("cs.xlsx", sheetName = "cs")
## データのマージ
koku_cs <- merge(koku, cs, by = "sc_g_c")
### 3,4年生の教諭の担任だけを取り出す
koku_cs <- subset(koku_cs, Q1 > 2)
koku_cs <- subset(koku_cs, Q37 == 2)
## 01データにする
library(memisc)
koku_cs$Q31S1 <- recode(koku_cs$Q31S1, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q31S2 <- recode(koku_cs$Q31S2, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q32S1 <- recode(koku_cs$Q32S1, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q32S2 <- recode(koku_cs$Q32S2, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q33S1 <- recode(koku_cs$Q33S1, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q33S2 <- recode(koku_cs$Q33S2, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q34S1 <- recode(koku_cs$Q34S1, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q34S2 <- recode(koku_cs$Q34S2, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q35S1 <- recode(koku_cs$Q35S1, 0 <- c(2, 3, 4), 1 <- c(1))
koku_cs$Q35S2 <- recode(koku_cs$Q35S2, 0 <- c(2, 3, 4), 1 <- c(1))
### 10年以下
koku_nv <- subset(koku_cs, Q38 > 0 & Q38 < 11)
nrow(koku_nv) #人数
## [1] 105
### 10年以上
koku_ex <- subset(koku_cs, Q38 > 10 & Q38 < 40)
nrow(koku_ex) #人数
## [1] 397
nrow(table(koku_cs$schl))
## [1] 163
nrow(koku_cs)
## [1] 502
library(MCMCpack)
koku.nv.mc.res.Q31S1 <- MCMClogit(Q31S1 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q31S2 <- MCMClogit(Q31S2 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q32S1 <- MCMClogit(Q32S1 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q32S2 <- MCMClogit(Q32S2 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q33S1 <- MCMClogit(Q33S1 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q33S2 <- MCMClogit(Q33S2 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q34S1 <- MCMClogit(Q34S1 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q34S2 <- MCMClogit(Q34S2 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q35S1 <- MCMClogit(Q35S1 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
koku.nv.mc.res.Q35S2 <- MCMClogit(Q35S2 ~ size, data = koku_nv, burnin = 10000,
mcmc = 50000)
summary(koku.nv.mc.res.Q31S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.9568 1.0697 0.004784 0.014474
## size -0.0521 0.0407 0.000182 0.000552
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.0672 1.2339 1.9287 2.6434 4.1400
## size -0.1343 -0.0784 -0.0514 -0.0246 0.0261
summary(koku.nv.mc.res.Q31S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 3.461 1.2048 0.005388 0.016345
## size -0.115 0.0452 0.000202 0.000623
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 1.232 2.620 3.423 4.2327 5.9684
## size -0.208 -0.144 -0.113 -0.0834 -0.0306
summary(koku.nv.mc.res.Q32S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.04 0.9559 0.004275 0.012683
## size -0.04 0.0368 0.000165 0.000491
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.810 0.3909 1.0250 1.6683 2.9429
## size -0.113 -0.0643 -0.0398 -0.0151 0.0313
summary(koku.nv.mc.res.Q32S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.8173 1.010 0.004519 0.013786
## size -0.0769 0.039 0.000174 0.000533
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.143 1.135 1.8085 2.4742 3.8543
## size -0.156 -0.102 -0.0764 -0.0506 -0.0021
summary(koku.nv.mc.res.Q33S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.30546 1.188 0.005312 0.016125
## size 0.00718 0.046 0.000206 0.000624
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.9387 0.4840 1.27580 2.0789 3.7421
## size -0.0852 -0.0233 0.00775 0.0386 0.0958
summary(koku.nv.mc.res.Q33S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 2.691 1.2523 0.00560 0.017001
## size -0.058 0.0469 0.00021 0.000637
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.383 1.819 2.6501 3.5047 5.2635
## size -0.152 -0.089 -0.0566 -0.0254 0.0308
summary(koku.nv.mc.res.Q34S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 2.4684 1.2494 0.00559 0.017134
## size -0.0452 0.0471 0.00021 0.000644
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.169 1.6052 2.4196 3.2686 5.084
## size -0.141 -0.0757 -0.0442 -0.0129 0.043
summary(koku.nv.mc.res.Q34S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 2.535 1.0556 0.004721 0.014180
## size -0.107 0.0407 0.000182 0.000548
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.550 1.809 2.500 3.2264 4.67
## size -0.189 -0.133 -0.105 -0.0789 -0.03
summary(koku.nv.mc.res.Q35S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.8524 0.9934 0.00444 0.013436
## size -0.0107 0.0381 0.00017 0.000515
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -1.0352 0.1818 0.8316 1.5070 2.8545
## size -0.0868 -0.0359 -0.0102 0.0152 0.0632
summary(koku.nv.mc.res.Q35S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.4978 0.9686 0.004332 0.013119
## size -0.0409 0.0377 0.000169 0.000516
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -1.398 -0.1508 0.4984 1.1377 2.4032
## size -0.115 -0.0661 -0.0406 -0.0153 0.0327
quantile(koku.nv.mc.res.Q31S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.12024 0.01364
quantile(koku.nv.mc.res.Q31S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.19174 -0.04398
quantile(koku.nv.mc.res.Q32S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.10146 0.01966
quantile(koku.nv.mc.res.Q32S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.1419 -0.0135
quantile(koku.nv.mc.res.Q33S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.07014 0.08167
quantile(koku.nv.mc.res.Q33S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.13782 0.01617
quantile(koku.nv.mc.res.Q34S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.12573 0.02981
quantile(koku.nv.mc.res.Q34S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.17510 -0.04161
quantile(koku.nv.mc.res.Q35S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.07422 0.05172
quantile(koku.nv.mc.res.Q35S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.10288 0.02068
geweke.diag(koku.nv.mc.res.Q31S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.1102 0.5105
geweke.diag(koku.nv.mc.res.Q31S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## -0.4235 0.8281
geweke.diag(koku.nv.mc.res.Q32S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## -0.5045 1.1181
geweke.diag(koku.nv.mc.res.Q32S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.8343 -0.4143
geweke.diag(koku.nv.mc.res.Q33S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.7486 -0.2844
geweke.diag(koku.nv.mc.res.Q33S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.35943 0.02435
geweke.diag(koku.nv.mc.res.Q34S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.49882 -0.02645
geweke.diag(koku.nv.mc.res.Q34S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.4276 0.0242
geweke.diag(koku.nv.mc.res.Q35S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.49602 -0.06841
geweke.diag(koku.nv.mc.res.Q35S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.40278 0.03849
koku.ex.mc.res.Q31S1 <- MCMClogit(Q31S1 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q31S2 <- MCMClogit(Q31S2 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q32S1 <- MCMClogit(Q32S1 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q32S2 <- MCMClogit(Q32S2 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q33S1 <- MCMClogit(Q33S1 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q33S2 <- MCMClogit(Q33S2 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q34S1 <- MCMClogit(Q34S1 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q34S2 <- MCMClogit(Q34S2 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q35S1 <- MCMClogit(Q35S1 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
koku.ex.mc.res.Q35S2 <- MCMClogit(Q35S2 ~ size, data = koku_ex, burnin = 10000,
mcmc = 50000)
summary(koku.ex.mc.res.Q31S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.2362 0.4025 1.80e-03 0.005403
## size -0.0204 0.0154 6.89e-05 0.000208
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.4529 0.9632 1.2326 1.50 2.0340
## size -0.0509 -0.0306 -0.0204 -0.01 0.0095
summary(koku.ex.mc.res.Q31S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.6515 0.3697 1.65e-03 0.004955
## size -0.0144 0.0143 6.39e-05 0.000192
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.0745 0.4011 0.6491 0.89876 1.3830
## size -0.0425 -0.0241 -0.0143 -0.00494 0.0137
summary(koku.ex.mc.res.Q32S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.2876 0.3650 1.63e-03 0.004898
## size -0.0056 0.0141 6.32e-05 0.000192
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.4322 0.0426 0.28649 0.53143 1.0091
## size -0.0334 -0.0150 -0.00558 0.00391 0.0222
summary(koku.ex.mc.res.Q32S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) -0.05222 0.3635 1.63e-03 0.004848
## size 0.00274 0.0141 6.29e-05 0.000189
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.7676 -0.2948 -0.05008 0.1918 0.6594
## size -0.0246 -0.0067 0.00262 0.0121 0.0303
summary(koku.ex.mc.res.Q33S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.33087 0.4669 2.09e-03 0.006349
## size 0.00899 0.0183 8.18e-05 0.000251
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.423 1.0180 1.32179 1.6412 2.2601
## size -0.027 -0.0032 0.00891 0.0212 0.0448
summary(koku.ex.mc.res.Q33S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.19101 0.4383 1.96e-03 0.005956
## size 0.00665 0.0171 7.64e-05 0.000233
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.3424 0.89149 1.18605 1.4818 2.0658
## size -0.0271 -0.00487 0.00675 0.0182 0.0401
summary(koku.ex.mc.res.Q34S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.9963 0.4746 0.00212 0.006397
## size -0.0289 0.0179 0.00008 0.000244
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 1.0911 1.6737 1.9849 2.3111 2.94868
## size -0.0645 -0.0407 -0.0287 -0.0169 0.00554
summary(koku.ex.mc.res.Q34S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.040618 0.363 1.63e-03 0.004738
## size -0.000998 0.014 6.28e-05 0.000188
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.6750 -0.2006 0.044937 0.28368 0.7515
## size -0.0283 -0.0105 -0.000937 0.00837 0.0268
summary(koku.ex.mc.res.Q35S1)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 1.6414 0.4596 2.06e-03 0.006142
## size -0.0114 0.0176 7.86e-05 0.000236
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) 0.7640 1.330 1.6355 1.943153 2.5667
## size -0.0462 -0.023 -0.0112 0.000512 0.0228
summary(koku.ex.mc.res.Q35S2)
##
## Iterations = 10001:60000
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 50000
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## (Intercept) 0.03974 0.362 1.62e-03 0.00488
## size -0.00218 0.014 6.26e-05 0.00019
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## (Intercept) -0.6724 -0.2054 0.04349 0.28233 0.7461
## size -0.0295 -0.0115 -0.00236 0.00729 0.0252
quantile(koku.ex.mc.res.Q31S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.045950 0.004905
quantile(koku.ex.mc.res.Q31S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.037962 0.009239
quantile(koku.ex.mc.res.Q32S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02902 0.01788
quantile(koku.ex.mc.res.Q32S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02034 0.02606
quantile(koku.ex.mc.res.Q33S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02109 0.03912
quantile(koku.ex.mc.res.Q33S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02144 0.03481
quantile(koku.ex.mc.res.Q34S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.0586801 0.0005863
quantile(koku.ex.mc.res.Q34S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02412 0.02224
quantile(koku.ex.mc.res.Q35S1[, 2], c(0.05, 0.95))
## 5% 95%
## -0.04052 0.01741
quantile(koku.ex.mc.res.Q35S2[, 2], c(0.05, 0.95))
## 5% 95%
## -0.02514 0.02108
geweke.diag(koku.ex.mc.res.Q31S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.719387 -0.009188
geweke.diag(koku.ex.mc.res.Q31S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.2216 0.3989
geweke.diag(koku.ex.mc.res.Q32S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.3149 0.5270
geweke.diag(koku.ex.mc.res.Q32S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.3852 0.2506
geweke.diag(koku.ex.mc.res.Q33S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.806973 -0.001365
geweke.diag(koku.ex.mc.res.Q33S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.5903 0.1485
geweke.diag(koku.ex.mc.res.Q34S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## -0.1087 0.8898
geweke.diag(koku.ex.mc.res.Q34S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.47798 0.05648
geweke.diag(koku.ex.mc.res.Q35S1)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.5621 0.1723
geweke.diag(koku.ex.mc.res.Q35S2)
##
## Fraction in 1st window = 0.1
## Fraction in 2nd window = 0.5
##
## (Intercept) size
## 0.2064 0.6490
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q31S2, pch = "●", xaxt = "n", yaxt = "n",
xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_ex$size, xlim = c(0, 40), koku_ex$Q31S2, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q31S2 <- mean(koku.nv.mc.res.Q31S2[, 1])
s.koku.nv.res.Q31S2 <- mean(koku.nv.mc.res.Q31S2[, 2])
i.koku.ex.res.Q31S2 <- mean(koku.ex.mc.res.Q31S2[, 1])
s.koku.ex.res.Q31S2 <- mean(koku.ex.mc.res.Q31S2[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q31S2 <- 1/(1 + exp(-i.koku.nv.res.Q31S2 - s.koku.nv.res.Q31S2 *
x))
y.koku.ex.res.Q31S2 <- 1/(1 + exp(-i.koku.ex.res.Q31S2 - s.koku.ex.res.Q31S2 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q31S2, lty = 1)
lines(x, y.koku.ex.res.Q31S2, lty = 2)
# 凡例
legend(3, 0.3, c("10年以下", "", "", "11年以上", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q32S2, pch = "●", xaxt = "n", yaxt = "n",
xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_ex$size, xlim = c(0, 40), koku_ex$Q32S2, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q32S2 <- mean(koku.nv.mc.res.Q32S2[, 1])
s.koku.nv.res.Q32S2 <- mean(koku.nv.mc.res.Q32S2[, 2])
i.koku.ex.res.Q32S2 <- mean(koku.ex.mc.res.Q32S2[, 1])
s.koku.ex.res.Q32S2 <- mean(koku.ex.mc.res.Q32S2[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q32S2 <- 1/(1 + exp(-i.koku.nv.res.Q32S2 - s.koku.nv.res.Q32S2 *
x))
y.koku.ex.res.Q32S2 <- 1/(1 + exp(-i.koku.ex.res.Q32S2 - s.koku.ex.res.Q32S2 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q32S2, lty = 1)
lines(x, y.koku.ex.res.Q32S2, lty = 2)
# 凡例
legend(3, 0.3, c("10年以下", "", "", "11年以上", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q33S2, pch = "●", xaxt = "n", yaxt = "n",
xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_ex$size, xlim = c(0, 40), koku_ex$Q33S2, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q33S2 <- mean(koku.nv.mc.res.Q33S2[, 1])
s.koku.nv.res.Q33S2 <- mean(koku.nv.mc.res.Q33S2[, 2])
i.koku.ex.res.Q33S2 <- mean(koku.ex.mc.res.Q33S2[, 1])
s.koku.ex.res.Q33S2 <- mean(koku.ex.mc.res.Q33S2[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q33S2 <- 1/(1 + exp(-i.koku.nv.res.Q33S2 - s.koku.nv.res.Q33S2 *
x))
y.koku.ex.res.Q33S2 <- 1/(1 + exp(-i.koku.ex.res.Q33S2 - s.koku.ex.res.Q33S2 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q33S2, lty = 1)
lines(x, y.koku.ex.res.Q33S2, lty = 2)
# 凡例
legend(3, 0.3, c("10年以下", "", "", "11年以上", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q34S2, pch = "●", xaxt = "n", yaxt = "n",
xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_ex$size, xlim = c(0, 40), koku_ex$Q34S2, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q34S2 <- mean(koku.nv.mc.res.Q34S2[, 1])
s.koku.nv.res.Q34S2 <- mean(koku.nv.mc.res.Q34S2[, 2])
i.koku.ex.res.Q34S2 <- mean(koku.ex.mc.res.Q34S2[, 1])
s.koku.ex.res.Q34S2 <- mean(koku.ex.mc.res.Q34S2[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q34S2 <- 1/(1 + exp(-i.koku.nv.res.Q34S2 - s.koku.nv.res.Q34S2 *
x))
y.koku.ex.res.Q34S2 <- 1/(1 + exp(-i.koku.ex.res.Q34S2 - s.koku.ex.res.Q34S2 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q34S2, lty = 1)
lines(x, y.koku.ex.res.Q34S2, lty = 2)
# 凡例
legend(3, 0.3, c("10年以下", "", "", "11年以上", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q35S2, pch = "●", xaxt = "n", yaxt = "n",
xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_ex$size, xlim = c(0, 40), koku_ex$Q35S2, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q35S2 <- mean(koku.nv.mc.res.Q35S2[, 1])
s.koku.nv.res.Q35S2 <- mean(koku.nv.mc.res.Q35S2[, 2])
i.koku.ex.res.Q35S2 <- mean(koku.ex.mc.res.Q35S2[, 1])
s.koku.ex.res.Q35S2 <- mean(koku.ex.mc.res.Q35S2[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q35S2 <- 1/(1 + exp(-i.koku.nv.res.Q35S2 - s.koku.nv.res.Q35S2 *
x))
y.koku.ex.res.Q35S2 <- 1/(1 + exp(-i.koku.ex.res.Q35S2 - s.koku.ex.res.Q35S2 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q35S2, lty = 1)
lines(x, y.koku.ex.res.Q35S2, lty = 2)
# 凡例
legend(3, 0.3, c("10年以下", "", "", "11年以上", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q31S2, pch = "●", col = 8, xaxt = "n",
yaxt = "n", xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q31S1, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q31S2 <- mean(koku.nv.mc.res.Q31S2[, 1])
s.koku.nv.res.Q31S2 <- mean(koku.nv.mc.res.Q31S2[, 2])
i.koku.nv.res.Q31S1 <- mean(koku.nv.mc.res.Q31S1[, 1])
s.koku.nv.res.Q31S1 <- mean(koku.nv.mc.res.Q31S1[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q31S2 <- 1/(1 + exp(-i.koku.nv.res.Q31S2 - s.koku.nv.res.Q31S2 *
x))
y.koku.nv.res.Q31S1 <- 1/(1 + exp(-i.koku.nv.res.Q31S1 - s.koku.nv.res.Q31S1 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q31S2, lty = 1)
lines(x, y.koku.nv.res.Q31S1, lty = 2)
# 凡例
legend(3, 0.3, c("理由・考え方", "", "", "正誤", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q32S2, pch = "●", col = 8, xaxt = "n",
yaxt = "n", xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q32S1, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q32S2 <- mean(koku.nv.mc.res.Q32S2[, 1])
s.koku.nv.res.Q32S2 <- mean(koku.nv.mc.res.Q32S2[, 2])
i.koku.nv.res.Q32S1 <- mean(koku.nv.mc.res.Q32S1[, 1])
s.koku.nv.res.Q32S1 <- mean(koku.nv.mc.res.Q32S1[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q32S2 <- 1/(1 + exp(-i.koku.nv.res.Q32S2 - s.koku.nv.res.Q32S2 *
x))
y.koku.nv.res.Q32S1 <- 1/(1 + exp(-i.koku.nv.res.Q32S1 - s.koku.nv.res.Q32S1 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q32S2, lty = 1)
lines(x, y.koku.nv.res.Q32S1, lty = 2)
# 凡例
legend(3, 0.3, c("理由・考え方", "", "", "正誤", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q33S2, pch = "●", col = 8, xaxt = "n",
yaxt = "n", xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q33S1, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q33S2 <- mean(koku.nv.mc.res.Q33S2[, 1])
s.koku.nv.res.Q33S2 <- mean(koku.nv.mc.res.Q33S2[, 2])
i.koku.nv.res.Q33S1 <- mean(koku.nv.mc.res.Q33S1[, 1])
s.koku.nv.res.Q33S1 <- mean(koku.nv.mc.res.Q33S1[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q33S2 <- 1/(1 + exp(-i.koku.nv.res.Q33S2 - s.koku.nv.res.Q33S2 *
x))
y.koku.nv.res.Q33S1 <- 1/(1 + exp(-i.koku.nv.res.Q33S1 - s.koku.nv.res.Q33S1 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q33S2, lty = 1)
lines(x, y.koku.nv.res.Q33S1, lty = 2)
# 凡例
legend(3, 0.3, c("理由・考え方", "", "", "正誤", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q34S2, pch = "●", col = 8, xaxt = "n",
yaxt = "n", xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q34S1, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q34S2 <- mean(koku.nv.mc.res.Q34S2[, 1])
s.koku.nv.res.Q34S2 <- mean(koku.nv.mc.res.Q34S2[, 2])
i.koku.nv.res.Q34S1 <- mean(koku.nv.mc.res.Q34S1[, 1])
s.koku.nv.res.Q34S1 <- mean(koku.nv.mc.res.Q34S1[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q34S2 <- 1/(1 + exp(-i.koku.nv.res.Q34S2 - s.koku.nv.res.Q34S2 *
x))
y.koku.nv.res.Q34S1 <- 1/(1 + exp(-i.koku.nv.res.Q34S1 - s.koku.nv.res.Q34S1 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q34S2, lty = 1)
lines(x, y.koku.nv.res.Q34S1, lty = 2)
# 凡例
legend(3, 0.3, c("理由・考え方", "", "", "正誤", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q35S2, pch = "●", col = 8, xaxt = "n",
yaxt = "n", xlab = "", ylab = "", axes = F)
par(new = T)
plot(koku_nv$size, xlim = c(0, 40), koku_nv$Q35S1, pch = "×", xaxt = "n", yaxt = "n",
xlab = "学級規模", ylab = "実施状況", axes = F)
axis(side = 1, at = seq(0, 40, 5), cex.axis = 0.8)
axis(side = 2, at = seq(0, 1, 1), labels = c("全く～半分くらい", "いつも・ほとんど"),
cex.axis = 0.8)
# ロジスティック曲線のための値取り出し
i.koku.nv.res.Q35S2 <- mean(koku.nv.mc.res.Q35S2[, 1])
s.koku.nv.res.Q35S2 <- mean(koku.nv.mc.res.Q35S2[, 2])
i.koku.nv.res.Q35S1 <- mean(koku.nv.mc.res.Q35S1[, 1])
s.koku.nv.res.Q35S1 <- mean(koku.nv.mc.res.Q35S1[, 2])
# 曲線に流し込むx軸範囲と区切り
x <- seq(0, 40, 1)
# ロジスティック曲線の式
y.koku.nv.res.Q35S2 <- 1/(1 + exp(-i.koku.nv.res.Q35S2 - s.koku.nv.res.Q35S2 *
x))
y.koku.nv.res.Q35S1 <- 1/(1 + exp(-i.koku.nv.res.Q35S1 - s.koku.nv.res.Q35S1 *
x))
# 曲線描画
lines(x, y.koku.nv.res.Q35S2, lty = 1)
lines(x, y.koku.nv.res.Q35S1, lty = 2)
# 凡例
legend(3, 0.35, c("理由・考え方", "", "", "正誤", ""), pch = c("●", "", "",
"×", ""), lty = c(0, 1, 0, 0, 2), bty = "n")
