Type: Package
Title: DEComposition of Indirect and Direct Effects
Version: 1.3
Date: 2022-06-06
Author: Christiana Kartsonaki
Maintainer: Christiana Kartsonaki <christiana.kartsonaki@gmail.com>
Description: Calculates various estimates for measures of educational differentials, the relative importance of primary and secondary effects in the creation of such differentials and compares the estimates obtained from two datasets.
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
LazyLoad: yes
NeedsCompilation: no
Packaged: 2022-06-06 23:02:42 UTC; christianak
Repository: CRAN
Date/Publication: 2022-06-06 23:20:05 UTC

Decomposition of Indirect and Direct Effects

Description

Calculates various estimates for measures of educational differentials, the relative importance of primary and secondary effects in the creation of such differentials and compares the estimates obtained from two datasets.

Details

Package: DECIDE
Type: Package
Version: 1.3
Date: 2022-06-06
License: GPL (>= 2)
LazyLoad: yes

See relative.importance.

Author(s)

Christiana Kartsonaki

Maintainer: Christiana Kartsonaki <christiana.kartsonaki@gmail.com>

References

Kartsonaki, C., Jackson, M. and Cox, D. R. (2013). Primary and secondary effects: Some methodological issues, in Jackson, M. (ed.) Determined to succeed?, Stanford: Stanford University Press.

Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M. and Cox, D. R. (2005) On Class Differentials in Educational Attainment. Proceedings of the National Academy of Sciences, 102: 9730–9733

Jackson, M., Erikson, R., Goldthorpe, J. H. and Yaish, M. (2007) Primary and secondary effects in class differentials in educational attainment: The transition to A-level courses in England and Wales. Acta Sociologica, 50 (3): 211–229

Compare estimates of log odds, log odds ratios and relative importance obtained by two datasets

Description

Computes 95% confidence intervals for the differences in log odds of transition, log odds ratios and relative importance estimates between the two datasets. Also calculates chi-squared test statistics and p-values for testing whether the differences are different from zero.

Usage

compare.relimp(dataset1, dataset2)

Arguments

dataset1

is the first dataset; a data frame with 4 columns, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

dataset2

is the second dataset; a data frame with 4 columns, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

Value

ci.diff.lo

95% confidence intervals for differences in log odds of transition

test.diff.lo

Test statistic for differences in log odds

test.diff.lo.pvalue

p-value for testing for differences in log odds

ci.diff.lor

95% confidence intervals for differences in log odds ratios

test.diff.lo

Test statistic for differences in log odds ratios

test.diff.lo.pvalue

p-value for testing for differences in log odds ratios

ci.diff.ri.1

95% confidence intervals for relative importance estimates - 1

ci.diff.ri.2

95% confidence intervals for relative importance estimates - 2

ci.diff.ri.avg

95% confidence intervals for relative importance estimates - average

Author(s)

Christiana Kartsonaki

References

Kartsonaki, C., Jackson, M. and Cox, D. R. (2013). Primary and secondary effects: Some methodological issues, in Jackson, M. (ed.) Determined to succeed?, Stanford: Stanford University Press.

Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M. and Cox, D. R. (2005) On Class Differentials in Educational Attainment. Proceedings of the National Academy of Sciences, 102: 9730–9733

Jackson, M., Erikson, R., Goldthorpe, J. H. and Yaish, M. (2007) Primary and secondary effects in class differentials in educational attainment: The transition to A-level courses in England and Wales. Acta Sociologica, 50 (3): 211–229

Examples

# generate two datasets
set.seed(1)
data1 <- data.frame(seq(1:10), rep(c(1, 2), length.out = 10), 
c(rep(0, times = 3), rep(1, times = 7)), 
c(rnorm(4, 0, 1), rnorm(4, 0.5, 1), NA, NA))

data2 <- data.frame(seq(1:10), rep(c(1, 2), length.out = 10), 
c(rep(0, times = 5), rep(1, times = 5)), 
c(rnorm(5, 1, 1), rnorm(5, 0.5, 1)))

# run function
compare.relimp(data1, data2)

Create data frames for each class

Description

Takes a data frame and creates a list of data frames by splitting the data by the factor "class".

Usage

create.classdata(dataset)

Arguments

dataset

A data frame produced by prepare.dataset.

Value

data_class

A list with number of elements equal to the number of classes and each element a data frame for each class.

Author(s)

Christiana Kartsonaki

Examples

# generate a dataset
data <- data.frame(seq(1:10), rep(c(1, 2, 3), length.out = 10), 
rbinom(1, n = 10, p = 0.7), c(rnorm(8, 0, 1), NA, NA))

data_clean <- prepare.data(data)

create.classdata(data_clean)

Plot distributions of performance and transition propensities

Description

Plots distribution of academic performance and probabilities of transition for each class.

Usage

plot_transition(dataset)

Arguments

dataset

A data frame with 4 columns only, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

Value

A plot of the distributions of performance and transition propensities for each class.

Author(s)

Christiana Kartsonaki

References

Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M. and Cox, D. R. (2005) On Class Differentials in Educational Attainment. Proceedings of the National Academy of Sciences, 102: 9730–9733

Kartsonaki, C., Jackson, M. and Cox, D. R. (2013). Primary and secondary effects: Some methodological issues, in Jackson, M. (ed.) Determined to succeed?, Stanford: Stanford University Press.

Jackson, M., Erikson, R., Goldthorpe, J. H. and Yaish, M. (2007) Primary and secondary effects in class differentials in educational attainment: The transition to A-level courses in England and Wales. Acta Sociologica, 50 (3): 211–229

Examples

# generate a dataset
set.seed(1)
data <- data.frame(seq(1:10), rep(c(1, 2), length.out = 10), 
c(rep(0, times = 3), rep(1, times = 7)), 
c(rnorm(4, 0, 1), rnorm(4, 0.5, 1), NA, NA))

# run function
plot_transition(data)

Prepare dataset to be used in relative.importance

Description

Prepares datasets to be in the format required by the function relative.importance. It is automatically called by relative.importance.

Usage

prepare.data(dataset)

Arguments

dataset

A data frame with 4 columns only, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

Value

dataset

The data frame given as the argument, with column names changed and missing values removed.

Author(s)

Christiana Kartsonaki

Examples

# generate a dataset
data <- data.frame(seq(1:10), rep(c(1, 2, 3), length.out = 10), 
	rbinom(1, n = 10, p = 0.7), c(rnorm(8, 0, 1), NA, NA))

# run function
data_clean <- prepare.data(data)

Description

Presents various estimates for measures of educational differentials, the relative importance of primary and secondary effects and corresponding standard errors and confidence intervals.

Usage

print_relimp(dataset)

Arguments

dataset

A data frame with 4 columns only, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

Value

Returns a more nicely presented version of the results given by relative.importance.

Author(s)

Christiana Kartsonaki

References

Kartsonaki, C., Jackson, M. and Cox, D. R. (2013). Primary and secondary effects: Some methodological issues, in Jackson, M. (ed.) Determined to succeed?, Stanford: Stanford University Press.

Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M. and Cox, D. R. (2005) On Class Differentials in Educational Attainment. Proceedings of the National Academy of Sciences, 102: 9730–9733

Jackson, M., Erikson, R., Goldthorpe, J. H. and Yaish, M. (2007) Primary and secondary effects in class differentials in educational attainment: The transition to A-level courses in England and Wales. Acta Sociologica, 50 (3): 211–229

See Also

relative.importance

Examples

# generate a dataset
set.seed(1)
data <- data.frame(seq(1:10), rep(c(1, 2, 3), length.out = 10),
	rbinom(1, n = 10, p = 0.7), c(rnorm(8, 0, 1), NA, NA))

# run function
print_relimp(data)

Relative importance of primary and secondary effects

Description

Calculates various estimates for measures of educational differentials, the relative importance of primary and secondary effects and corresponding standard errors and confidence intervals.

Usage

relative.importance(dataset)

Arguments

dataset

A data frame with 4 columns only, in the following order: 1: student's ID, 2: class, 3: transition (0 if not, 1 if yes) and 4: performance score.

Value

sample_size

Total number of individuals

no_classes

Number of classes

class_size

A list of no_classes elements, each element containing the size of each class

percentage_overall

Overall percentage that made the transition

percentage_class

A list of no_classes elements, each element containing percentage that made the transition for each class

fifty_point

50% point of transition

parameters

A data frame with the parameters of logistic regression (\alpha, \beta) and normal distribution (\mu, \sigma) for each class

transition_prob

A data frame with the transition probabilities

log_odds

A data frame with log odds of transition (diagonal elements: actual log odds for each class, off-diagonal: counterfactual log odds)

se_logodds

A data frame with the standard errors of the log odds of transition

ci_logodds

Approximate 95% confidence intervals for the log odds of transition

odds

Odds of transition

log_oddsratios

Log odds ratios

se_logoddsratios

Standard errors for the log odds ratios

ci_logoddsratios

Approximate 95% confidence intervals for the log odds ratios

oddsratios

Odds ratios

rel_imp_prim1

Estimates of the relative importance of primary effects using the first equation for calculating the relative importance

rel_imp_prim2

Estimates of the relative importance of primary effects using the second equation for calculating the relative importance

rel_imp_prim_avg

Estimates of the relative importance of primary effects using the the average of the two equations for calculating the relative importance

rel_imp_sec1

Estimates of the relative importance of secondary effects using the first equation for calculating the relative importance

rel_imp_sec2

Estimates of the relative importance of secondary effects using the second equation for calculating the relative importance

rel_imp_sec_avg

Estimates of the relative importance of secondary effects using the the average of the two equations for calculating the relative importance

se.ri.1

Standard errors of the relative importance estimates given by the first equation

ci.ri.1

Approximate 95% confidence intervals for the relative importance of secondary effects given by the first equation

se.ri.2

Standard errors of the relative importance estimates given by the second equation

ci.ri.2

Approximate 95% confidence intervals for the relative importance of secondary effects given by the second equation

se.ri.avg

Standard errors of the relative importance estimates given by the average of the two equations

ci.ri.avg

Approximate 95% confidence intervals for the relative importance of secondary effects given by the average of the two equations

Author(s)

Christiana Kartsonaki

References

Kartsonaki, C., Jackson, M. and Cox, D. R. (2013). Primary and secondary effects: Some methodological issues, in Jackson, M. (ed.) Determined to succeed?, Stanford: Stanford University Press.

Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M. and Cox, D. R. (2005) On Class Differentials in Educational Attainment. Proceedings of the National Academy of Sciences, 102: 9730–9733

Jackson, M., Erikson, R., Goldthorpe, J. H. and Yaish, M. (2007) Primary and secondary effects in class differentials in educational attainment: The transition to A-level courses in England and Wales. Acta Sociologica, 50 (3): 211–229

See Also

print_relimp, plot_transition

Examples

# generate a dataset
set.seed(1)
data <- data.frame(seq(1:10), rep(c(1, 2), length.out = 10), 
	c(rep(0, times = 3), rep(1, times = 7)), 
	c(rnorm(4, 0, 1), rnorm(4, 0.5, 1), NA, NA))

# run function
relative.importance(data)