RESEARCH SUPPORT
Economics Replication Package Support
From Code to README
For economists whose journal has asked for data and code: scripts arranged in one run order, a README that a data editor can follow, data vintages recorded, and a clear statement of how others can reach any restricted data.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Is Economics Replication Package Support?
Economics replication package support covers the part of a paper that a journal’s data editor reruns: code arranged in one run order, a README, a record of the data sources and vintages, an access statement for restricted files, and a check that every table and figure regenerates. The package is judged by whether someone else can reproduce the numbers, so it is best assembled while the estimates are being settled and not after acceptance. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, assigns the package to in-house PhD-level statisticians. The specification and the identification argument are the authors’ economics, and the package documents them as they stand.
WHEN TO ENGAGE
Working Paper Situations That Call for Support
Credibility is set by choices made long before the table appears, and changes to data or code late in the process are the hardest to defend.
Code Spread Across Coauthors Scripts sit on several coauthors' machines in different versions, and one shared folder, one run order and one set of package versions must be agreed before the tables are final.
Journal Data Policy Applies The target journal asks for data and code with a README, and cleaning and estimation steps spread across many files must be consolidated into one documented run order.
Restricted Data Involved Part of the data is licensed or confidential and cannot be shared, so the package must explain how others can obtain access and which results the code reproduces without it.
Revision Needs New Estimates A referee has asked for another specification, a different clustering level or a placebo test, and the additions must be documented without disturbing the existing tables.
PACKAGE CONTENTS
What Goes Into the Replication Package
A working paper's estimates and its package are built together, since a table that cannot be traced to code is a weak claim however sound the model.
Run Order and Master Script
Cleaning, merging, estimation and table steps are placed in one stated order behind a single entry point, with relative file paths, so the package runs on a machine other than yours.
Software Environment
The software version and every user-written command or library are recorded, or installed by the script, so that the environment can be rebuilt when the data editor reruns the code.
Restricted and Licensed Data
Where files cannot be shared, the package states how others can apply for access, which results the code reproduces without them and what the data custodian cleared for release.
Robustness Tables
Alternative samples, lags, functional forms and standard-error choices are tabulated with intervals, so that a reader sees which conclusions depend on the specification.
Data and Code Policy
The journal's policy is read first, such as the AEA data and code availability policy, and the package is checked against what the data editor requires, so that gaps are found before the data editor's own run.
Journal Table Format
Tables follow the conventions of Economics, with coefficients, standard errors and intervals presented as the target journal prefers.
REFEREE AND REPLICATION RISKS
Why Packages Fail a Data Editor's Rerun
Most failures surface only when someone else runs the code, and each one below can be found and corrected before the package is submitted.
Paths That Exist Only on One Machine File paths, working folders or installed packages exist only on the author's computer, so the code stops, or quietly behaves differently, when the data editor runs it elsewhere.
Random Steps Without a Seed Bootstrap, simulation or sampling steps run with no recorded seed, so a rerun gives slightly different standard errors and the printed tables cannot be matched exactly.
Tables Copied by Hand Numbers are typed from the output into the manuscript, so a late correction changes the code but not the paper, and the printed table no longer matches what the package produces.
Hand-Edited Intermediate Files A cleaned file was edited in a spreadsheet at some point, so the package code cannot rebuild it from the raw data and the data editor cannot reproduce the tables.
Data Vintage Unrecorded Administrative or national accounts series are revised after release, and when the package does not record which vintage was used, a rerun on the current release gives different numbers.
Favorable Specifications Only Many specifications were tried and only the favorable ones reach the paper, so the table overstates how stable the finding is, whereas the full set shows its range.
DELIVERABLES
Code, README and Tables for Submission
Taken together, these files trace each number in the paper to the code and data behind it, including results that rest on files that cannot be shared.
Cleaning and Merge Log A record of every step from the raw files to the analysis sample, with the observations dropped or merged at each step and the reason for it.
Estimation Code in Run Order Code that runs from the raw or access-controlled data to every table and figure in a stated order, so that the paper can be reproduced from a single entry point.
README for the Data Editor A replication package README with data sources and access conditions, software and package versions, the run order and the expected run time, in the layout that data editors usually ask for.
Coefficient and Robustness Tables Main results with intervals and the robustness set side by side, in the format of the target journal and with the sample behind each column.
Data Access Statement A description of any restricted or licensed data, where it can be obtained and which results the shareable files can reproduce without it.
Revision Round Tables Specifications requested in review, set beside the original columns, with a note on what each changes in the main conclusions.
ILLUSTRATIVE EXAMPLE
A Conditionally Accepted Paper Whose Code Would Not Rerun
An illustration of how a package is rebuilt after a journal's data editor asks for code that runs from the raw files.
ILLUSTRATIVE EXAMPLE · HOUSING MARKET RESEARCH · UNITED STATES
An assistant professor of economics had a paper on local housing markets, conditionally accepted, and the journal asked for a replication package. The code had grown over three years across a county price series, a commercial listings file and a spreadsheet of manual corrections.
The first rerun on a clean machine failed at the second script, because a folder path and one user-written command existed only on the author’s computer. The complication was the data. The price series had been revised since it was downloaded, the listings file was licensed and could not be redistributed, and one cleaning step had been done by hand in a spreadsheet. The package replaced the manual step with a documented rule, recorded the release of each series, added a master script with relative paths, and left the licensed file out with an access statement and a description of its layout. Every table was then regenerated and compared with the manuscript, and each value that differed was traced and explained.
Outcome: The economist received the master script, the rebuilt cleaning and merge log, the record of data releases, the README with the access statement, the regenerated tables and a reconciliation note listing every manuscript value that changed on rerun and the reason.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Estimation Tools and the Package Around Them
Economics journals accept code in several languages, and the estimators, the coauthors' habits and what a reader must rerun narrow the field to one.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCH RESPONSIBILITY
Identification Is an Economist's Argument
The research design, the economic reasoning and the claims in the paper belong to the authors, and estimation and documentation are the part supplied.
Agreed Statistical Support OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, defined in a scope that is confirmed before any work begins.
No Intellectual Contribution OnlineSPSS does not claim any intellectual contribution to your research. Authorship of the working paper or article remains yours.
Your Research Decisions The economic question, the identification argument, the interpretation of the estimates and the conclusions remain yours, with overall responsibility.
Confidential Handling Research data and project details are handled confidentially, project information is not published, and an NDA is available where required.
FREQUENTLY ASKED QUESTIONS
Packages, Restricted Data and Robustness
Restricted data, the amount of robustness a paper needs and the timing of journal checks, answered for economists.
What does economics replication package support cover?
OnlineSPSS prepares the package that a journal’s data editor reruns: code in one run order, a README, a record of data vintages, an access statement for restricted files and a check that every table regenerates. The specification and the identification argument stay the authors’ and are documented as they stand, and a first draft or a conditional acceptance are both workable points of entry.
How is a package made to run on another computer?
By removing everything that depends on the original machine. File paths are set relative to one root folder, a single master script calls each step in order, package and command versions are recorded or installed by the script, and the whole package is run once on a clean installation before submission, which is where hidden dependencies usually surface.
How much robustness does a working paper need?
No fixed number applies. The robustness set should answer the objections a referee would raise given the identification strategy, such as alternative samples, controls, clustering levels and placebo outcomes, reported compactly with intervals and not as an undifferentiated catalog.
Do I need the code ready before I submit?
Policies vary: some journals check the package after conditional acceptance and others at submission. Organizing the code while the estimates are still being settled is easier than reconstructing it later, and it exposes manual steps that would otherwise surface during the journal’s own run.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the draft paper, the data description and the existing code, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
An Estimation Path
a Data Editor Can Follow
When every table regenerates from one entry point, referees are left to argue about the economics rather than the arithmetic.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014