39 Replication and Synthetic Data

Access to comprehensive data is pivotal for replication, especially in the realm of social sciences. Yet, often the data are inaccessible, making replication a challenge (G. King 1995). This chapter dives into the nuances of replication, the exceptions to its norms, and the significance of synthetic data.

39.1 The Replication Standard

Replicability in research ensures:

Credibility and comprehension of empirical studies.
Continuity and progression in the discipline.
Enhanced readership and academic citations.

For a research to be replicable, the “replication standard” is vital: it entails providing all requisite information for replication by third parties. While quantitative research can, to some extent, offer clear data, qualitative studies pose complexities due to data depth.

39.1.1 Solutions for Empirical Replication

Role of Individual Authors:
- Authors need to vouch for the replication standard for enhancing their work’s credibility.
- Archives like the Inter-University Consortium for Political and Social Research (ICPSR) serve as depositories for replication datasets.
Creation of a Replication Data Set:
- A public data set, consisting of both original and relevant complementary data, can serve replication purposes.
Professional Data Archives:
- Organizations like ICPSR provide solutions to data storage and accessibility problems.
Educational Implications:
- Replication can be an excellent educational tool, and many programs now emphasize its importance.

39.1.2 Free Data Repositories

Zenodo: Hosted by CERN, it provides a place for researchers to deposit datasets. It’s not subject-specific, so it caters to various disciplines.
figshare: Allows researchers to upload, share, and cite their datasets.
Dryad: Primarily for datasets associated with published articles in the biological and medical sciences.
OpenICPSR: A public-facing version of the Inter-University Consortium for Political and Social Research (ICPSR) where researchers can deposit data without any cost.
Harvard Dataverse: Hosted by Harvard University, this is an open-source repository software application dedicated to archiving, sharing, and citing research data.
Mendeley Data: A multidisciplinary, free-to-use open access data repository where researchers can upload and share their datasets.
Open Science Framework (OSF): Offers both a platform for conducting research and a place to deposit datasets.
PubMed Central: Specific to life sciences, but it’s an open repository for journal articles, preprints, and datasets.
Registry of Research Data Repositories (re3data): While not a repository itself, it provides a global registry of research data repositories from various academic disciplines.
SocArXiv: An open archive for the social sciences.
EarthArXiv: A preprints archive for earth science.
Protein Data Bank (PDB): For 3D structures of large biological molecules.
Gene Expression Omnibus (GEO): A public functional genomics data repository.
The Language Archive (TLA): Dedicated to data on languages worldwide, especially endangered languages.
B2SHARE: A platform for storing and sharing research data sets in various disciplines, especially from European research projects.

39.1.3 Exceptions to Replication

Some exceptions to the replication standard are:

Confidentiality: Sometimes data, even fragmented, is too sensitive to share.
Proprietary Data: Data sets owned by entities might restrict dissemination, but usually, parts of such data can still be shared.
Rights of First Publication: Embargos might be set, but the essential data used in a study should be accessible.

39.2 Synthetic Data: An Overview

Synthetic data, modeling real data while ensuring anonymity, is becoming pivotal in research. While promising, it has its own complexities and should be approached with caution.

39.2.1 Benefits

Privacy preservation.
Data fairness and augmentation.
Acceleration in research.

39.2.2 Concerns

Misconceptions about inherent privacy.
Challenges with data outliers.
Models relying solely on synthetic data can pose risks.

39.2.3 Further Insights on Synthetic Data

Synthetic data bridges the model-centric and data-centric perspectives, making it an essential tool in modern research. Analogously, it’s like viewing the Mona Lisa’s replica, with the real painting stored securely.

Future projects, such as utilizing the R’s diamonds dataset for synthetic data generation, hold promise in demonstrating the vast potentials of this technology.

For a deeper dive into synthetic data and its applications, refer to (Jordon et al. 2022).

39.3 Application

The easiest way to create synthetic data is to use the synthpop package. Alternatively, you can do it manually

library(synthpop)
library(tidyverse)
library(performance)

# library(effectsize)
# library(see)
# library(patchwork)
# library(knitr)
# library(kableExtra)

head(iris)
#>   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#> 1          5.1         3.5          1.4         0.2  setosa
#> 2          4.9         3.0          1.4         0.2  setosa
#> 3          4.7         3.2          1.3         0.2  setosa
#> 4          4.6         3.1          1.5         0.2  setosa
#> 5          5.0         3.6          1.4         0.2  setosa
#> 6          5.4         3.9          1.7         0.4  setosa

synthpop::codebook.syn(iris)
#> $tab
#>       variable   class nmiss perctmiss ndistinct
#> 1 Sepal.Length numeric     0         0        35
#> 2  Sepal.Width numeric     0         0        23
#> 3 Petal.Length numeric     0         0        43
#> 4  Petal.Width numeric     0         0        22
#> 5      Species  factor     0         0         3
#>                             details
#> 1                  Range: 4.3 - 7.9
#> 2                    Range: 2 - 4.4
#> 3                    Range: 1 - 6.9
#> 4                  Range: 0.1 - 2.5
#> 5 'setosa' 'versicolor' 'virginica'
#> 
#> $labs
#> NULL

syn_df <- syn(iris, seed = 3)
#> 
#> Synthesis
#> -----------
#>  Sepal.Length Sepal.Width Petal.Length Petal.Width Species

# check for replciated uniques
replicated.uniques(syn_df, iris)
#> $replications
#>   [1]  TRUE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE  TRUE FALSE FALSE
#>  [13]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE
#>  [25] FALSE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
#>  [37] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
#>  [49] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
#>  [61] FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE  TRUE
#>  [73] FALSE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE
#>  [85] FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE  TRUE
#>  [97] FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE  TRUE FALSE FALSE
#> [109] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
#> [121] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE
#> [133] FALSE  TRUE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
#> [145] FALSE FALSE FALSE FALSE FALSE FALSE
#> 
#> $no.uniques
#> [1] 148
#> 
#> $no.replications
#> [1] 17
#> 
#> $per.replications
#> [1] 11.33333


# remove replicated uniques and adds a FAKE_DATA label 
# (in case a person can see his or own data in 
# the replicated data by chance)

syn_df_sdc <- sdc(syn_df, iris, 
                  label = "FAKE_DATA",
                  rm.replicated.uniques = T)
#> no. of replicated uniques: 17

iris |> 
    GGally::ggpairs()


syn_df$syn |> 
    GGally::ggpairs()

lm_ori <- lm(Sepal.Length ~ Sepal.Width + Petal.Length , data = iris)
# performance::check_model(lm_ori)
summary(lm_ori)
#> 
#> Call:
#> lm(formula = Sepal.Length ~ Sepal.Width + Petal.Length, data = iris)
#> 
#> Residuals:
#>      Min       1Q   Median       3Q      Max 
#> -0.96159 -0.23489  0.00077  0.21453  0.78557 
#> 
#> Coefficients:
#>              Estimate Std. Error t value Pr(>|t|)    
#> (Intercept)   2.24914    0.24797    9.07 7.04e-16 ***
#> Sepal.Width   0.59552    0.06933    8.59 1.16e-14 ***
#> Petal.Length  0.47192    0.01712   27.57  < 2e-16 ***
#> ---
#> Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#> 
#> Residual standard error: 0.3333 on 147 degrees of freedom
#> Multiple R-squared:  0.8402, Adjusted R-squared:  0.838 
#> F-statistic: 386.4 on 2 and 147 DF,  p-value: < 2.2e-16

lm_syn <- lm(Sepal.Length ~ Sepal.Width + Petal.Length , data = syn_df$syn)
# performance::check_model(lm_syn)
summary(lm_syn)
#> 
#> Call:
#> lm(formula = Sepal.Length ~ Sepal.Width + Petal.Length, data = syn_df$syn)
#> 
#> Residuals:
#>      Min       1Q   Median       3Q      Max 
#> -0.79165 -0.22790 -0.01448  0.15893  1.13360 
#> 
#> Coefficients:
#>              Estimate Std. Error t value Pr(>|t|)    
#> (Intercept)   2.96449    0.24538  12.081  < 2e-16 ***
#> Sepal.Width   0.39214    0.06816   5.754  4.9e-08 ***
#> Petal.Length  0.45267    0.01743  25.974  < 2e-16 ***
#> ---
#> Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#> 
#> Residual standard error: 0.3658 on 147 degrees of freedom
#> Multiple R-squared:  0.8246, Adjusted R-squared:  0.8222 
#> F-statistic: 345.6 on 2 and 147 DF,  p-value: < 2.2e-16

Open data can be assessed for its utility in two distinct ways:

General Utility: This refers to the broad resemblances within the dataset, allowing for preliminary data exploration.
Specific Utility: This focuses on the comparability of models derived from synthetic and original datasets, emphasizing analytical reproducibility.

For General utility

compare(syn_df, iris)

Specific utility

# just like regular lm, but for synthetic data
lm_syn <- lm.synds(Sepal.Length ~ Sepal.Width + Petal.Length , data = syn_df)
compare(lm_syn, iris)
#> 
#> Call used to fit models to the data:
#> lm.synds(formula = Sepal.Length ~ Sepal.Width + Petal.Length, 
#>     data = syn_df)
#> 
#> Differences between results based on synthetic and observed data:
#>              Synthetic  Observed        Diff Std. coef diff CI overlap
#> (Intercept)  2.9644900 2.2491402  0.71534988       2.884829  0.2640608
#> Sepal.Width  0.3921429 0.5955247 -0.20338187      -2.933611  0.2516161
#> Petal.Length 0.4526695 0.4719200 -0.01925058      -1.124602  0.7131064
#> 
#> Measures for one synthesis and 3 coefficients
#> Mean confidence interval overlap:  0.4095944
#> Mean absolute std. coef diff:  2.314347
#> 
#> Mahalanobis distance ratio for lack-of-fit (target 1.0): 3.08
#> Lack-of-fit test: 9.23442; p-value 0.0263 for test that synthesis model is
#> compatible with a chi-squared test with 3 degrees of freedom.
#> 
#> Confidence interval plot:


# summary(lm_syn)

You basically want your lack-of-fit test to be non-significant.

38 Sensitivity Analysis/ Robustness Check

A Appendix