training_5.html

<!DOCTYPE html>

<html>

<head>

<meta charset="utf-8" />
<meta name="generator" content="pandoc" />
<meta http-equiv="X-UA-Compatible" content="IE=EDGE" />


<title>Data Sharing</title>

<script src="site_libs/header-attrs-2.16/header-attrs.js"></script>
<script src="site_libs/jquery-3.6.0/jquery-3.6.0.min.js"></script>
<meta name="viewport" content="width=device-width, initial-scale=1" />
<link href="site_libs/bootstrap-3.3.5/css/flatly.min.css" rel="stylesheet" />
<script src="site_libs/bootstrap-3.3.5/js/bootstrap.min.js"></script>
<script src="site_libs/bootstrap-3.3.5/shim/html5shiv.min.js"></script>
<script src="site_libs/bootstrap-3.3.5/shim/respond.min.js"></script>
<style>h1 {font-size: 34px;}
       h1.title {font-size: 38px;}
       h2 {font-size: 30px;}
       h3 {font-size: 24px;}
       h4 {font-size: 18px;}
       h5 {font-size: 16px;}
       h6 {font-size: 12px;}
       code {color: inherit; background-color: rgba(0, 0, 0, 0.04);}
       pre:not([class]) { background-color: white }</style>
<script src="site_libs/jqueryui-1.11.4/jquery-ui.min.js"></script>
<link href="site_libs/tocify-1.9.1/jquery.tocify.css" rel="stylesheet" />
<script src="site_libs/tocify-1.9.1/jquery.tocify.js"></script>
<script src="site_libs/navigation-1.1/tabsets.js"></script>
<link href="site_libs/highlightjs-9.12.0/default.css" rel="stylesheet" />
<script src="site_libs/highlightjs-9.12.0/highlight.js"></script>
<script src="site_libs/kePrint-0.0.1/kePrint.js"></script>
<link href="site_libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
<script defer data-domain="cghlewis.github.io/mpsi-data-training" src="https://plausible.io/js/plausible.js"></script>
  

<style type="text/css">
  code{white-space: pre-wrap;}
  span.smallcaps{font-variant: small-caps;}
  span.underline{text-decoration: underline;}
  div.column{display: inline-block; vertical-align: top; width: 50%;}
  div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
  ul.task-list{list-style: none;}
    </style>

<style type="text/css">code{white-space: pre;}</style>
<script type="text/javascript">
if (window.hljs) {
  hljs.configure({languages: []});
  hljs.initHighlightingOnLoad();
  if (document.readyState && document.readyState === "complete") {
    window.setTimeout(function() { hljs.initHighlighting(); }, 0);
  }
}
</script>


<style type = "text/css">
.main-container {
  max-width: 940px;
  margin-left: auto;
  margin-right: auto;
}
img {
  max-width:100%;
}
.tabbed-pane {
  padding-top: 12px;
}
.html-widget {
  margin-bottom: 20px;
}
button.code-folding-btn:focus {
  outline: none;
}
summary {
  display: list-item;
}
details > summary > p:only-child {
  display: inline;
}
pre code {
  padding: 0;
}
</style>


<style type="text/css">
.dropdown-submenu {
  position: relative;
}
.dropdown-submenu>.dropdown-menu {
  top: 0;
  left: 100%;
  margin-top: -6px;
  margin-left: -1px;
  border-radius: 0 6px 6px 6px;
}
.dropdown-submenu:hover>.dropdown-menu {
  display: block;
}
.dropdown-submenu>a:after {
  display: block;
  content: " ";
  float: right;
  width: 0;
  height: 0;
  border-color: transparent;
  border-style: solid;
  border-width: 5px 0 5px 5px;
  border-left-color: #cccccc;
  margin-top: 5px;
  margin-right: -10px;
}
.dropdown-submenu:hover>a:after {
  border-left-color: #adb5bd;
}
.dropdown-submenu.pull-left {
  float: none;
}
.dropdown-submenu.pull-left>.dropdown-menu {
  left: -100%;
  margin-left: 10px;
  border-radius: 6px 0 6px 6px;
}
</style>

<script type="text/javascript">
// manage active state of menu based on current page
$(document).ready(function () {
  // active menu anchor
  href = window.location.pathname
  href = href.substr(href.lastIndexOf('/') + 1)
  if (href === "")
    href = "index.html";
  var menuAnchor = $('a[href="' + href + '"]');

  // mark the anchor link active (and if it's in a dropdown, also mark that active)
  var dropdown = menuAnchor.closest('li.dropdown');
  if (window.bootstrap) { // Bootstrap 4+
    menuAnchor.addClass('active');
    dropdown.find('> .dropdown-toggle').addClass('active');
  } else { // Bootstrap 3
    menuAnchor.parent().addClass('active');
    dropdown.addClass('active');
  }

  // Navbar adjustments
  var navHeight = $(".navbar").first().height() + 15;
  var style = document.createElement('style');
  var pt = "padding-top: " + navHeight + "px; ";
  var mt = "margin-top: -" + navHeight + "px; ";
  var css = "";
  // offset scroll position for anchor links (for fixed navbar)
  for (var i = 1; i <= 6; i++) {
    css += ".section h" + i + "{ " + pt + mt + "}\n";
  }
  style.innerHTML = "body {" + pt + "padding-bottom: 40px; }\n" + css;
  document.head.appendChild(style);
});
</script>

<!-- tabsets -->

<style type="text/css">
.tabset-dropdown > .nav-tabs {
  display: inline-table;
  max-height: 500px;
  min-height: 44px;
  overflow-y: auto;
  border: 1px solid #ddd;
  border-radius: 4px;
}

.tabset-dropdown > .nav-tabs > li.active:before {
  content: "";
  font-family: 'Glyphicons Halflings';
  display: inline-block;
  padding: 10px;
  border-right: 1px solid #ddd;
}

.tabset-dropdown > .nav-tabs.nav-tabs-open > li.active:before {
  content: "&#xe258;";
  border: none;
}

.tabset-dropdown > .nav-tabs.nav-tabs-open:before {
  content: "";
  font-family: 'Glyphicons Halflings';
  display: inline-block;
  padding: 10px;
  border-right: 1px solid #ddd;
}

.tabset-dropdown > .nav-tabs > li.active {
  display: block;
}

.tabset-dropdown > .nav-tabs > li > a,
.tabset-dropdown > .nav-tabs > li > a:focus,
.tabset-dropdown > .nav-tabs > li > a:hover {
  border: none;
  display: inline-block;
  border-radius: 4px;
  background-color: transparent;
}

.tabset-dropdown > .nav-tabs.nav-tabs-open > li {
  display: block;
  float: none;
}

.tabset-dropdown > .nav-tabs > li {
  display: none;
}
</style>

<!-- code folding -->


<style type="text/css">

#TOC {
  margin: 25px 0px 20px 0px;
}
@media (max-width: 768px) {
#TOC {
  position: relative;
  width: 100%;
}
}

@media print {
.toc-content {
  /* see https://github.com/w3c/csswg-drafts/issues/4434 */
  float: right;
}
}

.toc-content {
  padding-left: 30px;
  padding-right: 40px;
}

div.main-container {
  max-width: 1200px;
}

div.tocify {
  width: 20%;
  max-width: 260px;
  max-height: 85%;
}

@media (min-width: 768px) and (max-width: 991px) {
  div.tocify {
    width: 25%;
  }
}

@media (max-width: 767px) {
  div.tocify {
    width: 100%;
    max-width: none;
  }
}

.tocify ul, .tocify li {
  line-height: 20px;
}

.tocify-subheader .tocify-item {
  font-size: 0.90em;
}

.tocify .list-group-item {
  border-radius: 0px;
}


</style>


</head>

<body>


<div class="container-fluid main-container">


<!-- setup 3col/9col grid for toc_float and main content  -->
<div class="row">
<div class="col-xs-12 col-sm-4 col-md-3">
<div id="TOC" class="tocify">
</div>
</div>

<div class="toc-content col-xs-12 col-sm-8 col-md-9">


<div class="navbar navbar-default  navbar-fixed-top" role="navigation">
  <div class="container">
    <div class="navbar-header">
      <button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-bs-toggle="collapse" data-target="#navbar" data-bs-target="#navbar">
        <span class="icon-bar"></span>
        <span class="icon-bar"></span>
        <span class="icon-bar"></span>
      </button>
      <a class="navbar-brand" href="index.html"></a>
    </div>
    <div id="navbar" class="navbar-collapse collapse">
      <ul class="nav navbar-nav">
        <li>
  <a href="index.html">About</a>
</li>
<li>
  <a href="training_00.html">Data Management</a>
</li>
<li>
  <a href="training_0.html">Project Management</a>
</li>
<li>
  <a href="training_1.html">Module 1</a>
</li>
<li>
  <a href="training_2.html">Module 2</a>
</li>
<li>
  <a href="training_3.html">Module 3</a>
</li>
<li>
  <a href="training_4.html">Module 4</a>
</li>
<li>
  <a href="training_5.html">Module 5</a>
</li>
<li>
  <a href="training_6.html">Module 6</a>
</li>
      </ul>
      <ul class="nav navbar-nav navbar-right">
        
      </ul>
    </div><!--/.nav-collapse -->
  </div><!--/.container -->
</div><!--/.navbar -->

<div id="header">


<h1 class="title toc-ignore">Data Sharing</h1>

</div>


<p><br></p>
<hr />
<div id="overview" class="section level2">
<h2>Overview</h2>
<hr />
<p>Data sharing has become one of the most discussed areas around data
management. With the growing number of funders requesting a data sharing
plan (i.e. NIH), more people want to know how and why they should share
their study data. In a 2019 <a
href="https://journals.library.ualberta.ca/istl/index.php/istl/article/view/12">paper</a>,
Pasek and Mayer found that data curation and re-use was cited as the
area most needing improvement by graduate students across two
universities. And in a way data sharing is a great jumping off point for
many people to think about data management. With an end goal in mind,
researchers can start to think about what structures they need to put in
place to curate data that will be acceptable for data sharing.</p>
<hr />
</div>
<div id="what-is-data-sharing" class="section level2">
<h2>What is Data Sharing</h2>
<hr />
<p>While data sharing may sometimes encompass things such as sharing
correlation matrices, summary tables, or study results, that is not what
this module is about. This module is about sharing raw, item and case
level, primary data collected as part of a research study. It can also
include extant data collected and added to data that you collect. Beyond
that overview, what types of data you share, how you share, and where
you share, may depend more on things like your funder, your budget, your
project/participants, and your field. However, this module will provide
many recommendations for what and how to share your data that will lead
to the most benefits for you, your project, and your field.</p>
<hr />
</div>
<div id="share" class="section level2">
<h2>Why Share Data?</h2>
<hr />
<div id="ten-reasons" class="section level3">
<h3>Ten Reasons</h3>
<p>There are many reasons to consider sharing your research study data.
Ten compelling reasons are listed below.</p>
<ol style="list-style-type: decimal">
<li>Required by Funder</li>
</ol>
<p>There are many reasons to share data and the number one reason may be
that it is required by your funder. Almost all funders that education
researchers work with now require data sharing in some form as part of
their data management plan and the requirements are only getting more
explicit and serious over time as agencies are better understanding the
consequences of not sharing data and the benefits of sharing.</p>
<ol start="2" style="list-style-type: decimal">
<li>Engage in Open Science</li>
</ol>
<p>While this term can be a catch all for many things, what I mean here
is that sharing data allows researchers to review each others findings.
This includes confirming and challenging the findings of others, (<a
href="https://ies.ed.gov/funding/pdf/EDPlanPolicyDevelopmentGuidanceforPublicAccess.pdf">IES</a>).
And through this process we enhance scientific rigor and
reproducibility. “When a result can be reproduced by multiple
scientists, it validates the original results and readiness to progress
to the next phase of research”, (<a
href="https://www.nih.gov/research-training/rigor-reproducibility">NIH</a>).</p>
<ol start="3" style="list-style-type: decimal">
<li>Accelerate Discovery</li>
</ol>
<p>This is a reason mentioned in <a
href="https://grants.nih.gov/grants/guide/notice-files/NOT-OD-21-013.html">NIH’s
documentation</a>. Sharing data promotes reuse for future studies by
providing access to high-value datasets. This allows researchers to
develop and answer questions to pressing education issues in less time
than if they had to collect new original data.</p>
<ol start="4" style="list-style-type: decimal">
<li>Promote Collaboration</li>
</ol>
<p>Sharing your data provides the opportunity for you to collaborate
with new and existing collaborators to develop new research questions
and studies with the data you have shared.</p>
<ol start="5" style="list-style-type: decimal">
<li>Combine datasets to gain new insights</li>
</ol>
<p>This is an idea mentioned in <a
href="https://ies.ed.gov/funding/datasharing_policy.asp">IES
materials</a>. Similar to collaboration, sharing your data provides the
opportunity to combine datasets across studies, allowing new
discoveries. Combining datasets may also help those who have datasets
with low-statistical power, and without combining datasets, are unable
to answer questions of interest, (<a
href="https://edarxiv.org/2x3cu/">Logan, Hart &amp;
Schatschneider</a>).</p>
<ol start="6" style="list-style-type: decimal">
<li>Save resources</li>
</ol>
<p>Similar to accelerate discovery, allowing other researchers to use
your data, rather than collect new original data saves time, energy and
money. Publicly archiving your data allows others to see what has
already been collected, reducing duplicate data collection efforts.</p>
<ol start="7" style="list-style-type: decimal">
<li>Increase citation rate</li>
</ol>
<p>Sharing your dataset with an identifier such as persistent unique
identifier (ex: DOI), allows you and others to cite your data in
publications. You can track these citations to see the reach of your
work.</p>
<ol start="8" style="list-style-type: decimal">
<li>Support equitable access to high-quality datasets</li>
</ol>
<p>Similar to save resources, sharing data is a means to provide
equitable access of high-quality datasets for early career scholars,
students, and underrepresented researchers, (<a
href="https://edarxiv.org/2x3cu/">Logan, Hart &amp; Schatschneider</a>).
Not all researchers have the budget, staff, or connections to collect
high-quality datasets, but sharing your datasets freely and openly
allows everyone the opportunity to work with these data. Sharing data
also serves as an important resource for training future
researchers.</p>
<ol start="9" style="list-style-type: decimal">
<li>Diversity of thought</li>
</ol>
<p>Sharing data encourages diversity of analysis and opinions. Other
researchers may have novel questions they can answer with your data that
you have never considered.</p>
<ol start="10" style="list-style-type: decimal">
<li>Transparency leads to better data management</li>
</ol>
<p>One unintended consequence of data sharing may be that it promotes
good data management practices. Knowing that you will eventually be
sharing your data and documentation with others outside of your team may
motivate researchers to think hard about how to organize their data
management practices in a way that will produce data that they trust to
share with the outside world.</p>
<p>Resources:</p>
<p>📑 <a
href="https://venngage.net/ps/Uz65vsS6hI/new-10-benefits-of-data-sharing">Ashley
Edwards 10 Benefits of Data Sharing</a><br />
📑 <a href="https://edarxiv.org/2x3cu/">Data Sharing in Education
Science</a><br />
📑 <a href="http://www.withinandbetweenpod.com/">Within and Between
podcast: The what, why, how, and hesitations of data sharing</a></p>
<hr />
</div>
</div>
<div id="principles" class="section level2">
<h2>Principles</h2>
<hr />
<div id="fair" class="section level3">
<h3>FAIR</h3>
<p>As a reminder of what we covered in the <a
href="https://cghlewis.github.io/mpsi-data-training/training_00.html">Data
Management Module</a>, in 2016 the <a
href="https://www.go-fair.org/fair-principles/">FAIR Principles</a> were
published in Scientific Data, outlining 4 guiding principles for
scientific data management and stewardship. These principles should be
referred to when choosing when, where and how to share your research
data.</p>
<p>F: Findable</p>
<p>All data should be findable through a persistent identifier and have
good data documentation, aka metadata. As we move towards automation in
our work and life, the need for machine-readable metadata becomes more
prevalent for automatic discovery of data.</p>
<p>A: Accessible</p>
<p>You data is accessible if humans can access your data. This can mean
your data is available in a repository or through a request system.</p>
<p>I: Interoperable</p>
<p>Use standardized vocabularies as well as formats. Both humans and
machines should be able to read and interpret your data. Software
licenses should not pose a barrier to usage. Data should be available in
open formats that can be accessed by any software such as .csv, .txt,
.dat, etc. Furthermore, thorough data documentation should accompany
data.</p>
<p>R: Reusable</p>
<p>Your metadata should provide information on the broad context of your
project as well as your data collection to allow for accurate use of
your data. You should also have clear licensing for data use.</p>
<p>Additional resources:</p>
<p>📑 <a
href="https://podcasts.apple.com/us/podcast/episode-4-the-what-why-how-and-hesitations-of-data-sharing/id1517310439?i=1000483521281">Within
&amp; Between podcast</a><br />
📑 <a
href="https://journals.sagepub.com/doi/10.1177/2515245917746500">Practical
Solutions for Sharing Data and Materials from Psychological
Research</a></p>
<hr />
</div>
</div>
<div id="planning-for-data-sharing" class="section level2">
<h2>Planning for Data Sharing</h2>
<hr />
<p>While Data Sharing may often happen at the end of a project, planning
for data sharing should happen at the beginning. For many funders,
you’ll be required to write a brief overview of your data sharing plan
in your data management plan (DMP) as part of you grant proposal, and
will also be allowed to submit DMP associated costs in your grant
application budget requests. DMPs are often a supplement/appendix to
your grant application and restricted to anywhere from 1-2 pages (NIH,
NIJ, and NSF) to a 5 page maximum (IES). For most funders these DMPs are
not part of the scoring process, but they are reviewed by a panel or
program officer. Some funders may provide feedback and/or ask for
revisions if they believe your plan and/or your budget/associated costs
are not adequate.</p>
<p>What to include in a DMP varies some across funding agencies.
<strong>While you should check each funding agency’s site for their
specific DMP requirements</strong>, this comparison table provides an
overview of 11 common categories covered in a data management plan and
whether four large funding agencies ask applicants to address these
categories in their data management plans, as well as any additional
guidance they provide for each category.</p>
<p><strong>Categories to Include in a Data Management Plan and Guidance
Provided by Funder</strong></p>
<div
style="border: 1px solid #ddd; padding: 0px; overflow-y: scroll; height:800px; ">
<table class="table table-bordered" style="margin-left: auto; margin-right: auto;">
<thead>
<tr>
<th style="text-align:left;color: white !important;background-color: #2f4a60 !important;position: sticky; top:0; background-color: #FFFFFF;">
Content Category
</th>
<th style="text-align:center;color: white !important;background-color: #2f4a60 !important;position: sticky; top:0; background-color: #FFFFFF;">
IES - Data Management Plan
</th>
<th style="text-align:center;color: white !important;background-color: #2f4a60 !important;position: sticky; top:0; background-color: #FFFFFF;">
NIH - Data Management and Sharing Plan
</th>
<th style="text-align:center;color: white !important;background-color: #2f4a60 !important;position: sticky; top:0; background-color: #FFFFFF;">
NIJ - Data Archiving Plan
</th>
<th style="text-align:center;color: white !important;background-color: #2f4a60 !important;position: sticky; top:0; background-color: #FFFFFF;">
NSF - Data Management Plan
</th>
</tr>
</thead>
<tbody>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Roles and Responsibilities
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>What are staff roles in mgmt and long-term preservation of data?</p>
<p>Who ensures accessibility, reliability, and quality of data?</p>
Is there a plan if a core project member leaves the project or
institution?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Staff responsible for data creation and management should be
identified and their duties described.</p>
Identify any hardware/software required for managing data.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Type of Data to be Shared
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>How is data captured (surveys, assessments, observations)?</p>
<p>Will data be item-level and summary scores?</p>
<p>Will you share raw datasets and clean Datasets?</p>
What are the expected # of files? Expected # of participants in files?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Full item/case level dataset, not summary statistics or tables are
expected. This includes primary data collected through the grant, as
well as data created by linking to extant data sources.</p>
<p>Share, at a minimum, the data underlying any peer-reviewed
publication, this includes “final analytic measures as well as the
source measures used to construct them”. However, researchers are
encouraged to share any data that will “inform the field more
broadly”.</p>
Providing access to final cleaned data is required. Including raw data
and derived variables is optional.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>All primary data collected through the grant regardless of whether
the data are used to support scholarly publications.</p>
Data should be of sufficient quality to validate and replicate research
findings.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>All cleaned data collected under the grant award.</p>
Data should include all variables used to produce analysis, tables, and
descriptive information provided in the final report (ex: computed,
derived, and weight variables - if applicable).
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
Share primary data collected during the course of the grant.
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Format of Data
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>Electronic?</p>
<p>Nonproprietary software format (ex: csv)?</p>
<p>More than one format (.sav and .csv)?</p>
Any related tools needed to manipulate shared data?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Must be an electronic format.</p>
<p>Advised to provide data in multiple formats (one being a
non-proprietary format).</p>
Identify any hardware/software required for sharing data.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Datasets should be provided in formats consistent with those used in
the community that the repository serves.</p>
Preferred non-proprietary formats.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>For Quantitative: SPSS preferred, Stata and SAS are acceptable.
Embedded variable and value labels, and missing values assigned.</p>
For Qualitative: txt, rtf, Microsoft Word or PDF, Excel, databases
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Documentation
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>What metadata will you create (data dictionaries, codebooks)?</p>
<p>Consider project level, dataset level and variable level
documentation.</p>
<p>What format will it be in (xml, csv, pdf)?</p>
What other documentation do you plan to include when sharing data (code,
data collection instruments, protocols)?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Provide sufficient documentation to support responsible use by other
researchers.</p>
<p>“Information can be embedded directly into the file or can be
included with it as a separate file (e.g., a ReadMe.txt file, a .pdf,
etc.).”</p>
“Documentation should include a summary of the purpose of the data
collection, methodology and procedures used to collect the data, timing
of the data collection, as well as details of the data codes, definition
of variables, variable field locations, and frequencies.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
Examples include: Methodology and procedures used to collect the data,
Data labels, Definitions of variables, Any other information necessary
to reproduce and understand the data
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Documentation should include: IRB approved protocol, Final project
report/journal article, IRB approved consent (if applicable), Data use
agreements (if applicable, Description of file formats, data anomalies,
frequently used terms/acronyms, instructions for merging files,
documentation on decisions, guidance for using weights (if applicable),
citations, contact information, codebooks if variable and value labels
are not embedded in the data files, syntax files containing statistical
programming code as well as data manipulation code.</p>
PDF format preferred: Word, RTF and ASCII accepted
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Standards
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
Any data/documentation standards being used (ex: DDI, International
standard)?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“There are emerging metadata standards in many fields, but currently
there is not a set of standards in education research.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“While many scientific fields have developed and adopted common data
standards, others have not. In such cases, the Plan may indicate that no
consensus data standards exist for the scientific data and metadata to
be generated, preserved, and shared.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“DMPs submitted to EHR should be appropriate to the data being generated
and reflect the procedures, standards and best practices developed by
the communities of practice in the area of research being proposed.”
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Method of Data Sharing
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>How will you share data (institutional repository, data archive, PI
website)?</p>
<p>Will data be restricted and is a data enclave required?</p>
Is a data agreement required?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Highly encouraged to deposit data in publicly accessible
repositories. However, other methods may be used including the PI taking
responsibilitiy for data sharing or some combination of repository and
PI data sharing.</p>
Datasets should be discoverable and citable (ex: metadata and DOI).
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Strongly encourage the use of established repositories to the extent
possible, specifically, domain-specific repositories where possible.</p>
<p>Data should be findable and identifiable (ex: via a persistent unique
identifier or other standard indexing tools).</p>
While data kept by the researcher or institution, provided on request is
not preferred, NIH recognizes and respects that many communities (ex:
AI/AN communities) may want to manage, preserve and share their own
data.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>In most cases, the NIJ requires grantees to deposit their data in the
National Archive of Criminal Justice Data (NACJD), which is hosted by
ICPSR.</p>
If this is not the appropriate repository for the datasets, the data
archiving plan should include submission of study level information and
link to data location.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“If data or products are to be preserved by a third party, please refer
to their preservation plans if available.”
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Circumstances preventing sharing
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>Do you have any data covered by FERPA/HIPAA that doesn’t allow data
sharing?</p>
<p>Do you work with any partners that do not allow you to share data
(ex: School districts, Tribal regulations)?</p>
Are you working with proprietary data?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>“Specify appropriate restrictions on access to and usage of the data
to ensure protection of human subjects while not unduly restricting
access to the data.”</p>
If a DMP states data cannot be shared, researchers must provide a
compelling rationale.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Will there be any restrictions on data collected from human
subjects?</p>
<p>“Any restrictions imposed by federal, Tribal, or state laws,
regulations, or policies, or existing or anticipated agreements.”</p>
Any limitations should be communicated to both the data repository as
well as communicated in the Data Management and Sharing Plan for review.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“Unless otherwise specified in writing by the NIJ grant manager, as
authorized by the appropriate NIJ authority, data submission is required
for all research, development, and evaluation awards, and the
requirement may not be unilaterally modified or waived.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Privacy and rights of participants
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>How will you maintain participant confidentiality during your project
and when data is shared, prevent disclosure of PII?</p>
Did participants sign informed consent? Did the consent communicate how
participant data are expected to be used and shared?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Protect confidentiality in final data and follow rules in accordance
with IRB and any state/federal laws and regulations.</p>
<p>Proxy IDs may be used to protect direct disclosure of participants in
data, but means of indirect exposure should be identified as well (ex:
small numbers, sample characteristics) and remedies provided.</p>
Consent forms and IRB approvals should reference future data sharing so
that participants, schools, and other participating organizations are
made aware of conditions that will be put in place to protect privacy
prior to any data collection.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
Describe informed consent as well as how you will protect privacy and
confidentiality “consistent with applicable federal, Tribal, state, and
local laws, regulations, and policies”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“All direct identifiers to be removed and all indirect identifiers to be
recoded to prohibit re-identification.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“Access and sharing of data and products should reflect appropriate
protections for IRB, privacy, confidentiality, data security, and
intellectual property.”
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Data Security
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
NO
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
<p>How will you keep data secure on site during the project.</p>
Consider IRB requirements.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Protect confidentiality in final data and follow rules in accordance
with IRB and any state/federal laws and regulations.</p>
<p>Proxy IDs may be used to protect direct disclosure of participants in
data, but means of indirect exposure should be identified as well (ex:
small numbers, sample characteristics) and remedies provided.</p>
Consent forms and IRB approvals should reference future data sharing so
that participants, schools, and other participating organizations are
made aware of conditions that will be put in place to protect privacy
prior to any data collection.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
You do not have to speak to security as NIH acknowledges that data
security fall within the purview of Institutional IRB (while the project
is active) and the repository (once data is shared).
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“All direct identifiers to be removed and all indirect identifiers to be
recoded to prohibit re-identification.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“Access and sharing of data and products should reflect appropriate
protections for IRB, privacy, confidentiality, data security, and
intellectual property.”
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Schedule for Data Sharing
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
When will you share data and for how long?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>No later than when the main findings from the final study dataset are
published in a peer-reviewed scholarly publication. However, researchers
may share data earlier as well, if appropriate.</p>
Data should be available for 10 years unless a shorter period of time is
required to comply with Federal or State laws.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Data should be shared no later than the time of an associated
publication or end of the grant period, whichever comes first.</p>
<p>A single project can share data at different times (ex: share data
underlying publication during the period of award but ALSO share data
that have not yet led to a publication by the end of the award
period).</p>
Data should be available as long as it is useful for the larger research
community.
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“Grant recipients are strongly encouraged to submit data sets 90 days or
earlier prior to the end of the award project period.”
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
“Access to data and products should be provided, and data and the
products of research shared, as soon as is reasonably possible.”
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #b0d2da !important;">
Pre-registration
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
YES
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
NO
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
NO
</td>
<td style="text-align:center;width: 3in; background-color: #b0d2da !important;">
NO
</td>
</tr>
<tr>
<td style="text-align:left;width: 3in; background-color: #7da1bf !important;">
Where and when will you pre-register your study?
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
<p>Causal impact studies must be pre-registered in a recognized study
registry, documenting their confirmatory research questions and planned
analytic activities.</p>
<p>Must be registered within the first year of the project.</p>
Example registry options: OSF, Registry of efficacy and effectiveness
studies (REES).
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
<td style="text-align:center;width: 3in; background-color: #7da1bf !important;">
</td>
</tr>
</tbody>
</table>
</div>
<p><br></p>
<p>However, even if your funder does not require a data sharing plan,
there are still many reasons to consider sharing your data, as we
covered <a href="#share">above</a>. Planning what, how, and when to
share before your project even begins is the best way to ensure you have
everything in order by the time you need to share your data. You can
always update your plan during or after your project completion. If data
sharing is required by your funder, it may be helpful to keep in contact
with your program officer regarding any potential changes throughout
your project.</p>
<p>Data Management Plan resources:</p>
<p>IES:</p>
<p>📑 <a
href="https://ies.ed.gov/funding/datasharing_implementation.asp">IES
Implementation Guide for Public Access to Research Data</a><br />
📑 <a
href="https://ies.ed.gov/funding/pdf/EDPlanPolicyDevelopmentGuidanceforPublicAccess.pdf">U.S.
Department of Education Plan and Policy Development Guidance for Public
Access</a><br />
📑 <a href="https://ies.ed.gov/funding/datasharing_faq.asp">IES Data
Sharing FAQ</a><br />
📑 <a href="https://ies.ed.gov/funding/researchaccess.asp">IES Policy
Regarding Public Access to Research</a><br />
📑 <a href="https://dmptool.org/template_export/1895.pdf">DMP Tool
Template IES</a></p>
<p>NIH:</p>
<p>📑 <a
href="https://sharing.nih.gov/data-management-and-sharing-policy/planning-and-budgeting-DMS/writing-a-data-management-and-sharing-plan">NIH
Writing a Data Management and Sharing Plan</a><br />
📑 <a
href="https://grants.nih.gov/grants/guide/notice-files/NOT-OD-21-013.html">Final
NIH Policy for Data Management and Sharing</a><br />
📑 <a href="https://osf.io/awypt">NIH DMSP Guidance for Data Support
Services Working Group</a></p>
<p>NIJ:</p>
<p>📑 <a href="https://nij.ojp.gov/funding/data-archiving">NIJ Data
Archiving Plan</a></p>
<p>NSF:</p>
<p>📑 <a href="https://www.nsf.gov/bfa/dias/policy/dmpdocs/ehr.pdf">Data
Management for NSF EHR Directorate Proposals and Awards</a><br />
📑 <a
href="Dissemination%20and%20Sharing%20of%20Research%20Results%20-%20NSF%20Data%20Management%20Plan%20Requirements">NSF
Dissemination and Sharing of Research Results</a><br />
📑 <a
href="http://datasharing.sparcopen.org/compare?ids=11&amp;compare=data">SPARC
Data Sharing Requirement Comparison</a></p>
<hr />
</div>
<div id="best-practices" class="section level2">
<h2>Best Practices</h2>
<hr />
<p>While your funder may have guidance as well as requirements for your
data sharing plan, there are also generally accepted best practices that
you should consider when you construct your plan. Following required
guidelines and best practices will help you provide data that is useful
and accessible to researchers.</p>
<p>However, the <a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">Institute of
Education Sciences</a> put it best when they said, keep the big picture
in mind. They listed out 4 big ideas to consider when planning for data
sharing:</p>
<ol style="list-style-type: decimal">
<li><p>Focus on sharing well-organized and well-documented data. Include
all documentation necessary for someone with no familiarity with your
project to pick up your data and make sense of it. Consider everything
you can include so that future researchers aren’t reaching back out to
you with ongoing questions. Also, organize your data with a well
designed structure. Don’t share messy files that are inconsistent across
the project. Share files that are standardized, uniform, and can be
easily linked if necessary.</p></li>
<li><p>Commit to sharing some data or code to facilitate analysis. If
possible share more data beyond those used to produce a study’s main
findings. And if your data have restrictions and you are unable to
share, is there anything you can share that is still in the vein of open
science, such as code or aggregated data?</p></li>
<li><p>Don’t get stuck on one single way to share data. Each project is
unique and has it’s own opportunities and constraints. Think outside the
box about what means of data sharing works best for you, while also
considering how to maximize impact.</p></li>
<li><p>Not most, but some projects will need to consider the possibility
of tradeoffs. There may be times when sharing all of your data requires
you to share with restricted access, while removing some variables and
sharing only some of your data allows you to share the dataset openly.
Researchers may need to consider what makes the most sense here, and in
some circumstances you may be able to share data through of combination
of methods.</p></li>
</ol>
<div id="where-to-share-data" class="section level3">
<h3>Where to Share Data</h3>
<div id="repository" class="section level4">
<h4>Repository</h4>
<p>There are many reasons to choose a repository over alternatives such
as sharing from your personal website or providing data upon
request.</p>
<ol style="list-style-type: decimal">
<li>It’s a good open science practice</li>
</ol>
<p>Providing data in a repository allows your data to be publicly
accessible. While saying you provide data upon request or providing
access through a request system on your personal website is still data
sharing, it misses the mark on the FAIR principle of Accessible.
Repositories allow discoverability of your data through searchability.
As opposed to sharing on a personal website or sharing data on request,
sharing data in an open repository allows others to discover your data
merely by searching for topics of interest. Authors can also link to
their repository within publications.</p>
<ol start="2" style="list-style-type: decimal">
<li>Funders prefer it</li>
</ol>
<p>As open data sharing has become a requirement for many federal
funders, data management plans have started suggesting that PIs share
data in a repository vs other options, if at all possible. IES, NIH, and
NIJ all recommend the use of openly accessible repositories over the
alternative of having to contact PIs to request data.</p>
<ol start="3" style="list-style-type: decimal">
<li>Repositories provide persistent identifiers and licensing</li>
</ol>
<p>Repositories offer unique persistent identifiers which allow datasets
to be citable, meeting the FAIR principle of Findable. Researchers can
track the contributions of their dataset through the citation of their
dataset in published papers. Researchers can also add links to their
repository data in papers.</p>
<p>Repositories also offer licensing which allow you to make clear how
others can use your data. For example, LDbase, a Learning and
Development Data Repository, suggests that depositors use the Open Data
Commons Attribution License (ODC-by), which allows others to use, modify
and share the deposited data but users must cite the dataset when doing
so.</p>
<ol start="4" style="list-style-type: decimal">
<li>Using a repository provides a “hands-off” approach to data
sharing</li>
</ol>
<p>While preparing your materials for deposit in a repository may seem
like a lot of upfront work, it is actually the least burdensome route
for future you and your staff. Using a repository takes away the
long-term commitment of your staff to respond to data requests and build
datasets for requestors. In this respect, it also reduces cost. Costs
associated with paying staff to respond to these data requests. Also
costs for continuing to protect your data from “obsolescence, loss or
irreversible damage” from humans or technology, (<a
href="https://www.icpsr.umich.edu/files/deposit/dataprep.pdf">ICPSR</a>).
Making sure your file types are becoming obsolete or protecting your
data from corruption all take resources. <a
href="https://simplystatistics.org/posts/2020-04-30-asymptotics-of-reproducibility/">Roger
Peng</a> likens this to “maintenance capital expenditure” in the
business world. How much do I need to invest to keep up my promise of
data sharing. Using a repository reduces this investment because,
depending on the repository, they may take care of some or all of this
for you.</p>
<ol start="5" style="list-style-type: decimal">
<li>Repositories provide viable options for restricted data</li>
</ol>
<p>Some data may need to be for restricted use only, not available to
the general public, either because of participant confidentiality
issues, third-party licensing, or due to consent language/use agreements
that restrict data sharing.</p>
<p>For restricted use data, some repositories provide the option to have
data housed in a secure data enclave which stores and protects data with
the highest security. This data can be accessed through an application
process where applicants agree to strict requirements to preserve
confidentiality. “Some repositories assess applicants against a set of
criteria, while others require the original researcher to grant
permission for its use”, (<a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES</a>).
Some repositories may provide onsite data enclaves as well as remote
secure access data enclaves. Remote enclaves do not allow users to
download data to their computer but instead the user can analyze the
data remotely.</p>
<ol start="6" style="list-style-type: decimal">
<li>Repositories may have resources to guide you through the data
sharing process</li>
</ol>
<p>While not all repositories have physical people available to help
curate your data, repositories typically have detailed guidelines about
what you <strong>should</strong> deposit and what formats those files
should be in. This guidance can be very helpful for those who are new to
data sharing and feeling overwhelmed by the process. It’s a good
practice, that if you plan to share your data in a specific repository,
to reach out to that repository early. Work with them to develop your
Data Management Plan (DMP), or review the requirements of the repository
before completing your DMP so that you know what you will be expected to
have together at the end of your study. This early investigation into a
repository can help you organize your data management processes
throughout the study. You can also start developing a schedule for
ongoing deposits of data and documentation if that is something you want
to consider doing throughout your study.</p>
<p><strong>Considerations for Choosing a Repository</strong></p>
<ol style="list-style-type: decimal">
<li>Domain specific vs General</li>
</ol>
<p>One of the first choices you might consider is, do I want to use a
domain-specific repository, relevant to my field, or a more general
repository? Domain-specific repositories may be of more interest to
researchers in your field and may be the best option to help facilitate
discovery of your datasets. Using a domain specific repository can also
help you ensure you are preserving data and providing documentation
according to recognized standards in your field. However, there are many
other considerations to take into account in addition to discoverability
so weigh all options.</p>
<ol start="2" style="list-style-type: decimal">
<li>Cost</li>
</ol>
<p>Consider the costs to both store your data and for users to access
your data. There may also be costs associated with additional services
such as data curation.</p>
<ol start="3" style="list-style-type: decimal">
<li>Data Access</li>
</ol>
<p>Consider how users access the data in the repository. Is data
accessible to a wide audience? Are users required to have an account to
access data or is that optional?</p>
<p>Is it only data download or are there other ways to work with the
data online as well?</p>
<p>Does the repository have appropriate procedures to protect access to
data? Does it offer options for restricted use files? Do researchers set
the terms of use for restricted files or does the repository?</p>
<ol start="4" style="list-style-type: decimal">
<li>Data Security</li>
</ol>
<p>For data security, consider how the data is stored and backed up.
Also consider what standards the repository meets to ensure it is
protecting participant confidentiality.</p>
<p>How is data stored? What storage is used and how often is data backed
up? Is data encrypted during transfer and/or at rest?</p>
<p>Does the repository meet <a
href="https://en.wikipedia.org/wiki/General_Data_Protection_Regulation#General_provisions">GDPR
(General Data Protection Regulation) standards</a> for data
protection?</p>
<p>Does the repository help you review your data for missed
identifiers?</p>
<ol start="5" style="list-style-type: decimal">
<li>Types and Formats of Data Accepted</li>
</ol>
<p>Check for size limits of both the data files and the entire project.
Also check which file types are allowed. Certain repositories may have
preferences for file types for both data files and documentation
files.</p>
<ol start="6" style="list-style-type: decimal">
<li>Assignment of DOI</li>
</ol>
<p>Does the repository assign a direct object identifier (DOI) which
allows researchers to track the use and contributions of their dataset?
This is an important deciding factor. A DOI allows your dataset to be
findable, one of the FAIR principles. They can make your datasets easier
to find and can be used to measure the citation impact of your
dataset.</p>
<ol start="7" style="list-style-type: decimal">
<li>Viable Length of Deposit</li>
</ol>
<p>Does the repository have the resources to stay afloat for the
appropriate length of time to meet your funder’s data sharing
requirement? (i.e. 10 years for an IES grant)</p>
<ol start="8" style="list-style-type: decimal">
<li>Versioning and Removal</li>
</ol>
<p>Do depositor’s have permission to freely update and remove data from
the repository? Does the repository version amended data? Do they notify
users, who have already downloaded data, that there is a new version
available?</p>
<ol start="9" style="list-style-type: decimal">
<li>Additional Services</li>
</ol>
<p>Does the repository provide curation services to ensure high quality
data is deposited? Ensuring data are in interoperable formats, with
appropriate metadata, and identifiers removed? Or is that up to the
depositor?</p>
<ol start="10" style="list-style-type: decimal">
<li>Linking Capabilities</li>
</ol>
<p>Does the repository allow you to link to other projects,
publications, code, or data? Does it allow you to link to your code on
sites like GitHub? Does it allow you to link to other projects within
the repository? Does it allow you to link to other projects or
pre-registrations outside of the repository? Does it allow you to link
to publications?</p>
<ol start="11" style="list-style-type: decimal">
<li>CoreTrustSeal</li>
</ol>
<p>Is the repository certified by the <a
href="https://www.coretrustseal.org/">CoreTrustSeal</a>? This is not a
must-have, but can be an additional assurance to both you and your
funders that the repository you are choosing has been vetted by an
independent authority to ensure your data will be managed and curated
with high standards and that you will have continued access to your
data.</p>
<p><strong>Repository Options</strong></p>
<p>Currently there are very few education specific repositories. Both <a
href="https://www.ldbase.org/">LDbase</a> and <a
href="https://www.icpsr.umich.edu/web/pages/deposit/index.html">ICPSR</a>
are recommended by IES. LDbase being truly for education researchers,
and ICPSR being a little more general, for social sciences. There are
many general repository options though. A few well-known ones include <a
href="https://dataverse.harvard.edu/">Harvard DataVerse</a>, <a
href="https://datadryad.org/stash">Dryad</a>, <a
href="https://osf.io/">OSF</a>, <a
href="https://zenodo.org/">Zenodo</a>, and <a
href="https://figshare.com/">FigShare</a>.</p>
<p><a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">Institute of
Education Sciences</a> put together a table comparing 5 commonly used
repositories.</p>
<p>*Note: There is an error for LDbase. They <strong>do</strong> allow
you to update and remove data.</p>
<p><img src="img/repo_table.PNG" width="70%" style="display: block; margin: auto;" /></p>
<p>Resources:</p>
<p>📑 <a href="https://www.re3data.org/">Registry of
repositories</a><br />
📑 <a href="https://evodify.com/free-research-repository/">Great
overview and comparison of 4 well-known repositories</a><br />
📑 <a href="https://cdsweb.u-strasbg.fr/CTS-CDS">Additional Information
on CoreTrustSeal</a></p>
</div>
<div id="other-means-of-data-sharing" class="section level4">
<h4>Other means of data sharing</h4>
<p>There may be reasons that you cannot share with a repository; maybe
an agreement made with a study partner, or you are just not ready to
make the leap to depositing in a repository. Other options do exist.
While these methods may technically meet the requirements of your
funder, keep in mind that all of these options:</p>
<ul>
<li>Reduce the visibility of your data. How will you notify researchers
that your data is available?<br />
</li>
<li>Most likely increase the burden on you or your team to respond to
requests and maintain data over time.</li>
<li>Reduce accessibility of your data. These methods are likely to
increase burden on your requestor who has to complete informal data
requests, which are not always a streamlined process. This may reduce
the number of people who use your data.<br />
</li>
<li>Will be difficult to track the use of your datasets over time.
Without assigning a DOI to your dataset, you will need to rely on
requestors to notify you of future publications. Be sure to have a data
request agreement in place that allows requestors to state the purpose
of their request and their publication intentions.</li>
<li>Increase risks in data sharing. With no repository managing the
secure storage and secure file transfers, your team will need to
consider how to continue to securely store your data as well as how to
securely provide access to data for approved users.</li>
</ul>
<p>Non-repository means of data sharing:</p>
<ol style="list-style-type: decimal">
<li><p>Deposit your data with an Institutional Archive. This method has
several benefits compared to the others listed below, such as reducing
the burden on your staff and securely storing your data. However,
institutional repositories may provide less services than a public
repository and the data will less discoverable to others outside of your
institution.</p></li>
<li><p>Deposit your data with an agency that you partnered with for the
study (i.e. a School District). All data requests will go through that
partner.</p></li>
<li><p>Share your data through a lab, personal, or project website.
While this seems like an accessible option, you still need to remember
the long term costs associated with maintaining this data and responding
to data requests.</p></li>
<li><p>Informal peer-to-peer sharing, (<a
href="https://ukdataservice.ac.uk/learning-hub/research-data-management/plan-to-share/how-to-share-data/">UK
Data Service</a>). This may work well for peers but doesn’t make a
broader audience aware of the availability of your data and keeps the
burden of data maintenance and responding to requests on you and your
staff. Also, with no formal request system in place, this makes it
difficult for the requestor to know who to contact to request the data
or what information is needed to retrieve the data.</p></li>
<li><p>Supplemental materials attached to an article or stored on a
publisher’s website.</p></li>
<li><p>The least recommended option here is to have a statement in your
publication that says “data available upon request” and then provide an
email for people to reach out to request data. There are countless
horror stories where people have reached out to authors about getting
data and they’ve either never heard back or were told that the data was
not in fact, available. It is best to already have a system already in
place and direct people to that system, either a repository, your
personal website with a data request form, etc.</p></li>
</ol>
<p><a
href="https://arxiv.org/ftp/arxiv/papers/2110/2110.00888.pdf">Borghi and
Gulick</a> put together a flow chart for choosing where to deposit your
data that can be helpful to navigate your options. Although, to the
point about restricted data, while the authors refer you to
institutional libraries, we’ve seen that some public repositories do
allow you to submit restricted data to be accessed in enclaves.</p>
<p><img src="img/repo_flow.PNG" width="80%" style="display: block; margin: auto;" /></p>
</div>
</div>
<div id="what-data-to-share" class="section level3">
<h3>What data to share</h3>
<div id="considering-what-to-share" class="section level4">
<h4>Considering what to share</h4>
<p><strong>Consent</strong></p>
<ul>
<li><p>The first thing to review before sharing any data is, what did
you say you were going to do in your approved IRB consent form, or more
importantly, what language you did NOT use.</p>
<ul>
<li>Make sure you didn’t use any language that suggested you are “not
going to share data” or that you “plan to destroy data” at a certain
point. You want your consent language to be broad and not restrict your
data sharing to aggregate tables.</li>
</ul></li>
<li><p>You want your consent language to discuss confidentiality but
also mention that there are still risks.</p>
<ul>
<li>Make sure your consent form explains that your team will make all
reasonable efforts to de-identify data and keep it secure, but it is
impossible to guarantee that your data may never be linked back to you.
However, due to the extensive de-identification process, the harm or
risk to the participant would be minimal.</li>
<li>If your consent form does not warn about the possibility of
re-identification then, in addition to removing direct identifiers from
your data, you will need to really consider small sample sizes and
combinations of variables that might allow some to re-identify a
participant.</li>
</ul></li>
<li><p>It’s a bonus if you used language that explicitly states that you
plan to publicly share data in a trusted repository and participant’s
data may be used for future research.</p></li>
<li><p>When putting your data together for sharing, review permission to
share and remove any consented participant who did not give permission
for you to share their data. This information <em>should</em> have been
tracked in your participant tracking database at the time of consent
collection.</p></li>
<li><p>Refer back to <a
href="https://cghlewis.github.io/mpsi-data-training/training_2.html">Module
2</a> for Meyer’s dos and don’ts of consent form language.</p></li>
</ul>
<p>Resources for additional information on consent language:</p>
<p>📑 <a href="https://edarxiv.org/2x3cu/">Data Sharing in Education
Science</a><br />
📑 <a
href="https://www.icpsr.umich.edu/web/pages/datamanagement/confientiality/conf-language.html">ICPSR</a><br />
📑 <a
href="https://figshare.com/articles/preprint/Working_with_your_IRB_Obtaining_consent_for_open_data_sharing_through_consent_forms_and_data_use_agreements/13215305/1">Shero
and Hart: Working with your IRB</a></p>
<p><strong>Funder requirements</strong></p>
<p>You’ll also want to review what you laid out in your data management
plan in your project proposal. What did you say you would share
(although this can be adjusted) and what did your funder suggest you do
share? Most funders will state that at <em>minimum</em>, the data
underlying your peer-reviewed, published study findings should be
shared. However, most funders will also note that you should consider
sharing more than that.</p>
</div>
<div id="data" class="section level4">
<h4>Data</h4>
<p>Researchers should consider sharing all primary data collected
through the project, including both raw (item-level) and derived
variables and can also consider sharing any existing data linked to
original data. If extant data is included, make sure to review any
existing data licenses associated with that data that may impact data
sharing. There may also be some item-level data that cannot be shared.
Review copyright for published scales to see what is allowed. Some paid
assessments may only allow you to share derived scores or un-labelled
item-level information (ex: “item 1 of assessment” vs “the actual
question wording”) (<a
href="https://figshare.com/articles/presentation/Data_Sharing_and_Data_Shared/15040740">Jessica
Logan</a>).</p>
<p><strong>Format</strong></p>
<p>You’ll need to consider what format to share your data in. This
encompasses many decisions.</p>
<ol style="list-style-type: decimal">
<li>Raw vs Clean vs Analytic files</li>
</ol>
<ul>
<li><p>The <strong>true raw data</strong> is the data that comes
directly from the source with no manipulations. The problem with this
raw data is, it’s not usually in a user-friendly format. Metadata is not
usually applied, variable names aren’t decipherable, and values may be
confusing. The other thing about the true raw data is, you may have not
yet removed participants who are not in the study, removed duplicate
instances, or removed identifying variables. This data should never
leave your secure data storage area, approved by IRB, in your
lab.</p></li>
<li><p>The next level is <strong>de-identified raw data</strong>. This
is your raw data as mentioned above, but with identifiers removed,
duplicates removed, and checked to ensure that only consented
participants are included. While this file <em>can</em> be shared, I
still don’t find this file very useful for those wanting to use your
data.</p></li>
<li><p>The third level is a <strong>cleaned data file</strong>. This is
your raw data file, checked to ensure that only consented participants
are included, de-identified, and cleaned to a usable standard. If you
refer back to our discussion on <a
href="https://cghlewis.github.io/mpsi-data-training/training_4.html#Data_cleaning_steps">data
cleaning steps</a>, it would include the transformations mentioned
there. Think of a dataset with interpretable and consistent variable
names, with metadata added (such a variable and value labels), and coded
consistently. This is a data file that is “ready for analysis”. This
means the dataset should be in the minimally accepted format for a
future researcher to be able to pick up the dataset and do some basic
exploratory analysis with the dataset without having to dig deep into
the documentation. This is a broad, clean, general dataset that ANYONE
can use. It will include all item-level data as well as any derived
scores (sum scores, means) or created variables (ex: cohort, time)
necessary to complete the file. My recommendation is that
<strong>this</strong> is the file you share.</p></li>
<li><p>The fourth level is an <strong>analytic file</strong>. While
people use this term in different ways, what I am referring to here is a
file prepared for a specific analysis that is based on the
aforementioned cleaned data file. These are files that begin as the
cleaned data file mentioned above and then may be further reduced in
size (either in cases and/or columns), and may have additional
transformations that are <em>created solely for the purpose of a
specific analysis</em>. You can think of this as a file used as the
basis of a specific publication. While I do recommended this file to be
shared as well, I view this as a secondary source, shared for the
purposes of allowing replication of your study, rather than for the
purposes of people using your data for their own future
analyses.</p></li>
</ul>
<ol start="2" style="list-style-type: decimal">
<li>Merged vs Unlinked</li>
</ol>
<p>If you are collecting data across time or collecting different forms
within time but across participants, you will need to consider whether
or not you want to merge files before sharing or provide distinct files
that users can merge on their own.</p>
<ul>
<li><p><strong>Merging datasets</strong> may be a great options for
longitudinal studies with many waves of data collection. Merging files
across time can reduce the burden for researchers who want to view all
data across time for each unit of analysis. You may consider making a
merged file for each participant level (student level dataset, teacher
level dataset, school level dataset). While education researchers
typically merge their datasets in wide format, where each case is one
row, there may be reasons to merge your data in long format. Merging in
long format is more efficient for data storage and may make data
contents more clear. Variable names stay the same across time (rather
than have time appended to them), and each case has a row of data for
each time point they were in the study.</p></li>
<li><p><strong>Uploading datasets separately</strong> is another option
though. Keeping data in separate files reduces file size and may make it
easier to upload and access data. It may also make it easier for
researchers to narrow down files to their variables of interest. If they
are only interested in one form from the student dataset, having the
forms separated allows researchers to download just the file for that
one form, rather than downloading the larger dataset and dropping
variables not relevant to their research. If you decide to share
datasets separately in repository, consider developing a folder and file
naming structure that allows researchers to know what they are working
with. You will also need to ensure that your files have identifiers to
link datasets. Consider adding a README to your documentation that lays
out how files can be linked through identifiers.</p></li>
</ul>
<pre><code>        levelName
1 project-name   
2  ¦--files_1920 
3  ¦   ¦--student
4  ¦   °--teacher
5  °--files_2021 
6      ¦--student
7      °--teacher</code></pre>
<ol start="3" style="list-style-type: decimal">
<li>File format</li>
</ol>
<ul>
<li><p>Most funders, including <a
href="https://ies.ed.gov/funding/datasharing_implementation.asp">IES</a>,
prefer data to be deposited in an electronic, rectangular format. It is
also recommended to provide data in at least one non-proprietary format
that is not restricted to a specific software. This not only provides
access to a broader audience, those without specific software licenses,
but also protects against technological obsolescence. Open formats
include text files like .csv, tab-delimited, or ASCII. This
recommendation meets the FAIR principle of interoperable.</p></li>
<li><p>It is nice, however, to <strong>also</strong> share data in
formats that have embedded metadata, created using programs such as SAS,
Stata, or SPSS. You can either export a file that allows embedded
metadata (.sas7bdat, .dta, .sav), or you can create raw ASCII files
along with setup files for people to run in specified statistical
programs. Providing your data in more than one format (non-proprietary
and proprietary), can give your user options and also protect your data
from obsolescence.</p></li>
<li><p>It is important to review your repository data format
requirements. <a
href="https://www.icpsr.umich.edu/web/pages/deposit/index.html">ICPSR</a>
for example, prefers submissions of files with embedded metadata, such
as SPSS, Stata or SAS files. They then use these files and curate them
into ASCII data with setup files to accompany statistical programs. The
National Archive of Criminal Justice Data used by NIJ prefers SPSS
formats for quantitative data, but will also accept Stata or SAS
files.</p></li>
</ul>
<p><strong>Confidentiality</strong></p>
<p>No matter what data you share or what format you share it in, you
need to consider confidentiality. All data you share should be
de-identified. This includes removing direct identifiers as well as
reducing the risk of confidentiality breaches through indirect
identifiers or deduction.</p>
<ol style="list-style-type: decimal">
<li>De-Identified Data</li>
</ol>
<ul>
<li><p>Direct Identifiers</p>
<ul>
<li><p>These are personally identifiable information (PII) like we
discussed in <a
href="https://cghlewis.github.io/mpsi-data-training/training_2.html">Module
2</a>. These are typically identifiers collected for data collection and
project implementation purposes but not necessary for secondary analyses
and should be removed before sharing data. These include the <a
href="https://www.luc.edu/its/aboutits/itspoliciesguidelines/hipaainformation/18hipaaidentifiers/">18
HIPAA identifiers</a> as well as <a
href="https://studentprivacy.ed.gov/content/personally-identifiable-information-education-records#glossary-node-210">FERPA
Identifiers</a>, such as:</p>
<ul>
<li>Name</li>
<li>Email</li>
<li>Phone Number</li>
<li>Social Security Numbers</li>
<li>State, District, or School level assigned Student IDs</li>
<li>IP Address</li>
</ul></li>
<li><p>If schools and districts were promised confidentiality, names of
schools and districts will need to be removed as well.</p></li>
<li><p>In replacement of names, random study IDs (either numeric or
alphanumeric) should be used to allow linking across forms and time, as
well as clustering of observations as needed. IDs should not be assigned
in a logical order, such as alphabetical. They should also be completely
distinct from identifying information such as name and birth date. If
IDs were assigned in a non-random way during the study, or IDs were
shared throughout the study with participants, consider assigning yet
another set of study IDs for the purposes of publicly sharing data. You
will want to track both sets of these IDs in your participant tracking
database.</p></li>
</ul></li>
<li><p>Indirect Identifiers</p>
<ul>
<li><p>These are identifiers that may not independently link to an
individual in your study but through combinations of these variables, or
small group sizes, individuals may be identified.</p>
<ul>
<li>Birth date</li>
<li>Small counts of gender, race, special education, or other
demographics (Review categorical variables with categories of 5 or less,
some people may see 3 or less)</li>
<li>Crosstabs of variables (ex: grade level and special education
status)</li>
<li>Geography (ex: zipcode)</li>
<li>Exact dates of events</li>
</ul></li>
<li><p>A few methods, <strong>there are many more</strong>, to account
for these variables are:</p>
<ul>
<li><p>Top or Bottom Coding: This is a method of restricting the upper
and lower range to mask outliers. For something like income you may top
code to “$150k or higher”.</p></li>
<li><p>Combine/Collapse categories: For categorical variables with
categories containing Ns of 5 or less, consider combining categories
into an “other” category. If issues of small sample size occur with
crosstabs, you can also consider merging two or more variables into a
new summary variable.</p></li>
<li><p>Aggregation: Identifier variables here are replaced with higher
level aggregations of information. For example you may aggregate zip
code to county or state level geography or aggregate birth date to year
of birth.</p></li>
<li><p>Recode continuous variables: Continuous variables with outliers
can be collapsed into categories to reduce the chances of participant
deduction.</p></li>
<li><p>Remove/redact variables: This often happens with variables such
as birth date. Most researchers want birth date for an age calculation
so it is feasible to calculate age variables using your dates of
interest (ex: assessment date) and then remove both birth dates and
assessment dates before sharing data.</p></li>
</ul></li>
</ul></li>
<li><p>Open-ended comments/questions</p>
<ul>
<li>These types of open-ended text boxes need to be considered for any
identifying information. You can either choose to remove all open text
variables before sharing data or code all responses into a categorical
variable.</li>
</ul></li>
</ul>
<p>Once your data has been de-identified, this data is ready to ready to
be released on the internet. If no other restrictions exist for this
data (i.e. a data use or sharing agreement) the data follows the Release
and Forget model. There is little control a data owner has once the data
is released into the world outside of any licenses that have been
assigned, and any terms of use requestors agree to when accessing the
data.</p>
<p>Resources on de-identification:</p>
<p>📑 <a
href="https://figshare.com/articles/preprint/De-Identification_Guide/13228664">Schatsneider,
Edwards, and Shero have a great paper on data
de-identification</a><br />
📑 <a
href="https://www.povertyactionlab.org/sites/default/files/research-resources/J-PAL-guide-to-deidentifying-data.pdf">J-PAL
has an excellent table of direct and indirect identifiers, as well as
recommended de-identification methods on page 10</a><br />
📑 <a
href="https://infograph.venngage.com/ps/5p6yjaAGTSs/new-5-things-to-check-for-data-deidentification">Great
recommendations on de-identification of data from LDbase</a><br />
📑 <a
href="https://www.icpsr.umich.edu/files/deposit/dataprep.pdf">ICPSR has
a discussion on treating indirect identifiers</a><br />
📑 <a href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES
has a section on Disclosure Risk Management in their Sharing Study Data
Guide</a><br />
📑 <a
href="https://the-turing-way.netlify.app/reproducible-research/rdm/rdm-personal.html">The
Turing Way provides examples of anonymizing data</a><br />
📑 <a
href="https://nvlpubs.nist.gov/nistpubs/ir/2015/NIST.IR.8053.pdf">National
Institute of Standards and Technology De-Identification Guide</a></p>
<ol start="2" style="list-style-type: decimal">
<li>Identifiable Data</li>
</ol>
<ul>
<li><p>If you are unable to remove all identifiers, if your consent
agreements do not allow you to share data, or if identifiers are
integral for linking datasets, then you have the option of creating a
restricted use file. These files can be shared through a data enclave.
Some repositories may provide both the option of onsite data enclaves as
well as virtual secure access enclaves. Using remote access enclaves,
users are not allowed to download data but are able to analyze data
remotely. Onsite data enclaves typically have security measures in place
such as no windows, key card entry, and a computer with no internet
access. These enclaves are also typically monitored by archive staff.
Access to these restricted data are permitted only through an
application process where requestors complete a detailed data use
agreement.</p></li>
<li><p>Even if you need to create a restricted access file, it is
important to consider creating a second, de-identified file as well to
maximize public benefit. The public access version may have more
variables removed for categories collapsed to create anonymity while the
restricted file will leave limited identifying information in the data.
However, be sure to consider all possible disclosure risks before
sharing to ensure that someone with access to both the restricted and
public files are not able to identify individuals. Also make sure that
no inconsistencies between files are created during the
de-identification process.</p></li>
</ul>
<p><a
href="https://www.childandfamilydataarchive.org/cfda/archives/CFDA/studies/38290/summary">The
United States Department of Health and Human Services</a> has a great
example of depositing restricted access research study data in a
repository (Child and Family Data Archive). You can see here that
various metadata are available for the study but restricted data are not
available for direct download.</p>
<p><img src="img/restricted_use.PNG" width="100%" style="display: block; margin: auto;" /></p>
<p><img src="img/restricted_use2.PNG" width="95%" style="display: block; margin: auto;" /></p>
<ol start="3" style="list-style-type: decimal">
<li>Data Use vs Data Sharing Agreements</li>
</ol>
<ul>
<li><p><strong>Data Use Agreements</strong> are legally binding formal
agreements (contract) used when sharing data with restricted access. <a
href="https://www.ldbase.org/resources/best-practices/data-use-vs-data-sharing">Florida
State University</a> has an example of a DUA here.</p>
<ul>
<li><p>The topics typically outlined in a DUA include:</p>
<ul>
<li>Terms and conditions of the use of the data, including prohibiting
any attempt to re-identify the data, linking to other data or
redistribution</li>
<li>Review access controls</li>
<li>Obligations for keeping the data secure during time of access</li>
<li>Liability of harm from using the data</li>
<li>Publication rules</li>
<li>Privacy rights</li>
</ul></li>
<li><p>And the user is typically also required to provide the following
information:</p>
<ul>
<li>What is their research plan and reasons for needing access to
confidential information</li>
<li>A data protection plan for keeping data secure</li>
<li>IRB approval</li>
</ul></li>
</ul></li>
<li><p><strong>Data Sharing Agreement</strong> are documents that
researchers may wish to have users complete when they want to list the
terms of use of their data. This is a less formal agreement and can be
used when you are sharing de-identified data. LDbase has a great <a
href="https://www.ldbase.org/resources/templates/example-data-sharing-agreement">sample
Data Use Agreement</a>. Reasons for using a data sharing agreement might
include:</p>
<ul>
<li>You are still actively analyzing your data and want to limit
access</li>
<li>You want users to agree to specific terms of use, such as adding you
as a co-author on papers using your data or ensure users cite your data
in publications</li>
<li>While your data may be de-identified there may still be qualitative
data or small sample sizes that could potentially make participants
identifiable so you want to restrict who uses the data and ensure users
have a plan to maintain confidentiality</li>
<li>You want to limit the types of research projects that use your
data</li>
</ul></li>
</ul>
<p>Resources:</p>
<p>📑 <a
href="https://www.icpsr.umich.edu/files/deposit/dataprep.pdf">ICPSR
provides more detail on Data Use Agreements</a><br />
📑 <a
href="https://www.ldbase.org/resources/best-practices/data-use-vs-data-sharing">LDbase
explains data use vs data sharing agreements</a><br />
📑 <a
href="https://research.unc.edu/wp-content/uploads/sites/61/2013/04/CCM3_039360.pdf">UNC
Data Use Agreement Flow Chart</a></p>
<p><br></p>
</div>
</div>
<div id="what-documentation-to-share" class="section level3">
<h3>What documentation to share</h3>
<p>I think the general rule is, share as much documentation as you
possibly can! But to narrow that down, I’ve provided some general
guidance below.</p>
<div id="considering-what-to-share-1" class="section level4">
<h4>Considering what to share</h4>
<p><strong>What is suggested from your funder?</strong></p>
<p>If you completed a data management plan as part of your grant
proposal, it’s likely that your funder asked you to very likely that
your funder provided some general guidance on what type of documentation
they expect you to share. Depending on the document you read, IES
generally suggests you share things such as codebooks that describe each
measure, syntax files used to convert raw data to analytic data, and
general project level data that allows users to know to better
understand the purpose of the project and what activities happened
during the grant. See <a
href="https://cghlewis.github.io/mpsi-data-training/training_5.html#Planning_for_Data_Sharing">DMP
table</a> above to read more about what funders suggest.</p>
<p><strong>What is required from your repository?</strong></p>
<p>Each repository may also mandate specific documentation or may
suggest types of documentation to provide. Review the requirements of
your repository early on in your project so you can plan for appropriate
documentation. <a
href="https://www.icpsr.umich.edu/web/pages/deposit/index.html">ICPSR</a>
for example provides suggested documentation to include, such as
codebooks, instruments, README files, project summary documents and
publications. And they also have an intake form that allows you to enter
project level metadata when you submit your data to their system.</p>
</div>
<div id="types-of-documentation-to-share" class="section level4">
<h4>Types of documentation to share</h4>
<p>Similar to what we discussed in <a
href="https://cghlewis.github.io/mpsi-data-training/training_1.html#Metadata">Module
1</a>, documentation can be categorized into 3 different types of
metadata. Consider sharing something from each of these categories to
give users the best overview of your project and how to use your
data.</p>
<ol style="list-style-type: decimal">
<li>Project level documentation</li>
</ol>
<p>This documentation should give a user a holistic picture of the
project: What was the purpose, who participated, and what activities
happened during the study. Some of this documentation may be directly
entered into your repository (check before developing this as a
standalone document), or it may also be provided as a standalone
document. For example, both <a
href="https://www.icpsr.umich.edu/web/pages/deposit/deposit-fields.html">ICPSR</a>
and <a
href="https://www.ldbase.org/resources/user-guide/information-to-gather/project-metadata">LDbase</a>
have an intake form that allows you to enter this information directly
into the system upon submission.</p>
<p>Sections to consider including in this document:</p>
<ul>
<li>Funding Source<br />
</li>
<li>Overview of study<br />
</li>
<li>Intervention and Research Design<br />
</li>
<li>Setting and Sample<br />
</li>
<li>Project Timeline (what data was collected, on whom, and when)<br />
</li>
<li>Measures Used in Study (including citations)<br />
</li>
<li>Overview of Procedures (Recruitment, Consent, Inclusion/Exclusion
Criteria, Randomization, Data Collection, Data Preparation and
Processing)</li>
</ul>
<p>Supplemental project level documents that can also be included
are:</p>
<ul>
<li>Consort Diagrams<br />
</li>
<li>Data Collection Instruments<br />
</li>
<li>A Table of Response Rates for Instruments</li>
</ul>
<ol start="2" style="list-style-type: decimal">
<li>Dataset level documentation</li>
</ol>
<p>This documentation provides a big picture sense of what datasets are
included in the project and how they are connected. The type of dataset
level documentation to consider including when data sharing is a README.
As discussed in <a
href="https://cghlewis.github.io/mpsi-data-training/training_1.html#README">Module
1</a>, a README is a plain text document that can serve many purposes.
It literally is an indicator to users to “Read Me” before moving forward
with this data. In the context of data sharing, your README should
include things such as:</p>
<ul>
<li>The file name of each included dataset<br />
</li>
<li>How datasets can be linked together (linking variable)<br />
</li>
<li>Ns of each dataset<br />
</li>
<li>Any acronyms used in each dataset worth knowing (ex: st_ = student
measures, tch_ = teacher measures)</li>
</ul>
<p>See p.58 of the <a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES Sharing
Study Data Guide</a> for an example of a README.</p>
<ol start="3" style="list-style-type: decimal">
<li>Variable level documentation</li>
</ol>
<p>This is documentation that allows users to know exactly what is
included in your data files and how to interpret each column. There are
two types of variable level documentation to consider including.</p>
<ul>
<li><p>Data dictionary: This rectangular document describes every
variable in each dataset, how the variables are coded, and any
transformations made to any variables. This can also include information
such as scales associated with each variable, survey skip patterns,
variable universe, and missing values assigned. See p.46 of the <a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES Sharing
Study Data Guide</a> for an example of a simple data
dictionary.</p></li>
<li><p>Codebook: This is a text file or pdf that provides summary level
information about the associated dataset and variables. Without ever
having to open the data file, a user can see summary statistics about
the data and make decisions based on that information. This file should
include every variable in the data and provide information such as the
name, label and type, as well as summary statistics for each variables
(Ns, percents, means, ranges, and percent missing). See p.48 of the <a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES Sharing
Study Data Guide</a> for an example of a codebook.</p></li>
</ul>
<p><strong>Note:</strong> These terms are often interchangeable. You
will notice that in the <a
href="https://ies.ed.gov/ncee/pubs/2022004/pdf/2022004.pdf">IES Data
Sharing Study Data Guide</a>, they refer to a Codebook as as Summary
Statistics Document and they refer to a Data Dictionary as a Codebook.
Depending on what organization you work with, different terms may be
used, but the ideas are all still the same.</p>
<ol start="4" style="list-style-type: decimal">
<li>Other documentation to include:</li>
</ol>
<ul>
<li>Publications: These can be either shared in the project or linked
within the project to another site where the publication lives.</li>
<li>Code: In terms of code, I believe there are two types of code that
can be shared:
<ol style="list-style-type: decimal">
<li><strong>Raw data cleaning code</strong>: This is the code used to
transform your true raw data to your cleaned data file. I am a believer
that you <strong>do not need to share this code</strong>. It is often
not something anyone would be interested in and there could potentially
be 20 or more syntax files to clean each form of data for each wave of
data collection. If you want to share the transformations you performed
on the raw data to get to your clean file, I would suggest adding these
transformations to your data dictionary. It will be easier to read (not
in code) and less overwhelming for the user. You simply add a column to
your data dictionary that is called something along the lines of
“recodes” or “transformations”. <strong>See example below</strong>. You
can also consider including a data cleaning plan, discussed in <a
href="https://cghlewis.github.io/mpsi-data-training/training_4.html">Module
4</a>, either in replacement of adding your transformations to your data
dictionary, or in addition to. Providing a data cleaning plan gives a
linear look at the order of transformations that you performed.</li>
<li><strong>Analysis code</strong>: This is the code used to transform
your cleaned data file into your analytic file, created solely for the
purpose of some analysis. This is usually code you use to narrow the
sample, narrow the variables you use, and do additional transformations
specifically for your analysis. If you plan to share any analysis level
files, I do recommend sharing this code. However, before sharing this
code, you need to review this code for identifiers. Are there any names
or other direct identifiers (either in the code or the comments)
included in the code that need to be removed before sharing? Also plan
to remove any absolute paths to your personal working directories and
replace with relative paths.</li>
</ol></li>
</ul>
<p>Ex: Data Dictionary with Transformations</p>
<div id="yvamwdabos" style="overflow-x:auto;overflow-y:auto;width:auto;height:auto;">
<style>html {
  font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, 'Helvetica Neue', 'Fira Sans', 'Droid Sans', Arial, sans-serif;
}

#yvamwdabos .gt_table {
  display: table;
  border-collapse: collapse;
  margin-left: auto;
  margin-right: auto;
  color: #333333;
  font-size: 16px;
  font-weight: normal;
  font-style: normal;
  background-color: #FFFFFF;
  width: auto;
  border-top-style: solid;
  border-top-width: 2px;
  border-top-color: #A8A8A8;
  border-right-style: none;
  border-right-width: 2px;
  border-right-color: #D3D3D3;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #A8A8A8;
  border-left-style: none;
  border-left-width: 2px;
  border-left-color: #D3D3D3;
}

#yvamwdabos .gt_heading {
  background-color: #FFFFFF;
  text-align: center;
  border-bottom-color: #FFFFFF;
  border-left-style: none;
  border-left-width: 1px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 1px;
  border-right-color: #D3D3D3;
}

#yvamwdabos .gt_title {
  color: #333333;
  font-size: 125%;
  font-weight: initial;
  padding-top: 4px;
  padding-bottom: 4px;
  padding-left: 5px;
  padding-right: 5px;
  border-bottom-color: #FFFFFF;
  border-bottom-width: 0;
}

#yvamwdabos .gt_subtitle {
  color: #333333;
  font-size: 85%;
  font-weight: initial;
  padding-top: 0;
  padding-bottom: 6px;
  padding-left: 5px;
  padding-right: 5px;
  border-top-color: #FFFFFF;
  border-top-width: 0;
}

#yvamwdabos .gt_bottom_border {
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
}

#yvamwdabos .gt_col_headings {
  border-top-style: solid;
  border-top-width: 2px;
  border-top-color: #D3D3D3;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  border-left-style: none;
  border-left-width: 1px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 1px;
  border-right-color: #D3D3D3;
}

#yvamwdabos .gt_col_heading {
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: normal;
  text-transform: inherit;
  border-left-style: none;
  border-left-width: 1px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 1px;
  border-right-color: #D3D3D3;
  vertical-align: bottom;
  padding-top: 5px;
  padding-bottom: 6px;
  padding-left: 5px;
  padding-right: 5px;
  overflow-x: hidden;
}

#yvamwdabos .gt_column_spanner_outer {
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: normal;
  text-transform: inherit;
  padding-top: 0;
  padding-bottom: 0;
  padding-left: 4px;
  padding-right: 4px;
}

#yvamwdabos .gt_column_spanner_outer:first-child {
  padding-left: 0;
}

#yvamwdabos .gt_column_spanner_outer:last-child {
  padding-right: 0;
}

#yvamwdabos .gt_column_spanner {
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  vertical-align: bottom;
  padding-top: 5px;
  padding-bottom: 5px;
  overflow-x: hidden;
  display: inline-block;
  width: 100%;
}

#yvamwdabos .gt_group_heading {
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: initial;
  text-transform: inherit;
  border-top-style: solid;
  border-top-width: 2px;
  border-top-color: #D3D3D3;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  border-left-style: none;
  border-left-width: 1px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 1px;
  border-right-color: #D3D3D3;
  vertical-align: middle;
}

#yvamwdabos .gt_empty_group_heading {
  padding: 0.5px;
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: initial;
  border-top-style: solid;
  border-top-width: 2px;
  border-top-color: #D3D3D3;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  vertical-align: middle;
}

#yvamwdabos .gt_from_md > :first-child {
  margin-top: 0;
}

#yvamwdabos .gt_from_md > :last-child {
  margin-bottom: 0;
}

#yvamwdabos .gt_row {
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
  margin: 10px;
  border-top-style: solid;
  border-top-width: 1px;
  border-top-color: #D3D3D3;
  border-left-style: none;
  border-left-width: 1px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 1px;
  border-right-color: #D3D3D3;
  vertical-align: middle;
  overflow-x: hidden;
}

#yvamwdabos .gt_stub {
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: initial;
  text-transform: inherit;
  border-right-style: solid;
  border-right-width: 2px;
  border-right-color: #D3D3D3;
  padding-left: 5px;
  padding-right: 5px;
}

#yvamwdabos .gt_stub_row_group {
  color: #333333;
  background-color: #FFFFFF;
  font-size: 100%;
  font-weight: initial;
  text-transform: inherit;
  border-right-style: solid;
  border-right-width: 2px;
  border-right-color: #D3D3D3;
  padding-left: 5px;
  padding-right: 5px;
  vertical-align: top;
}

#yvamwdabos .gt_row_group_first td {
  border-top-width: 2px;
}

#yvamwdabos .gt_summary_row {
  color: #333333;
  background-color: #FFFFFF;
  text-transform: inherit;
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
}

#yvamwdabos .gt_first_summary_row {
  border-top-style: solid;
  border-top-color: #D3D3D3;
}

#yvamwdabos .gt_first_summary_row.thick {
  border-top-width: 2px;
}

#yvamwdabos .gt_last_summary_row {
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
}

#yvamwdabos .gt_grand_summary_row {
  color: #333333;
  background-color: #FFFFFF;
  text-transform: inherit;
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
}

#yvamwdabos .gt_first_grand_summary_row {
  padding-top: 8px;
  padding-bottom: 8px;
  padding-left: 5px;
  padding-right: 5px;
  border-top-style: double;
  border-top-width: 6px;
  border-top-color: #D3D3D3;
}

#yvamwdabos .gt_striped {
  background-color: rgba(128, 128, 128, 0.05);
}

#yvamwdabos .gt_table_body {
  border-top-style: solid;
  border-top-width: 2px;
  border-top-color: #D3D3D3;
  border-bottom-style: solid;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
}

#yvamwdabos .gt_footnotes {
  color: #333333;
  background-color: #FFFFFF;
  border-bottom-style: none;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  border-left-style: none;
  border-left-width: 2px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 2px;
  border-right-color: #D3D3D3;
}

#yvamwdabos .gt_footnote {
  margin: 0px;
  font-size: 90%;
  padding-left: 4px;
  padding-right: 4px;
  padding-left: 5px;
  padding-right: 5px;
}

#yvamwdabos .gt_sourcenotes {
  color: #333333;
  background-color: #FFFFFF;
  border-bottom-style: none;
  border-bottom-width: 2px;
  border-bottom-color: #D3D3D3;
  border-left-style: none;
  border-left-width: 2px;
  border-left-color: #D3D3D3;
  border-right-style: none;
  border-right-width: 2px;
  border-right-color: #D3D3D3;
}

#yvamwdabos .gt_sourcenote {
  font-size: 90%;
  padding-top: 4px;
  padding-bottom: 4px;
  padding-left: 5px;
  padding-right: 5px;
}

#yvamwdabos .gt_left {
  text-align: left;
}

#yvamwdabos .gt_center {
  text-align: center;
}

#yvamwdabos .gt_right {
  text-align: right;
  font-variant-numeric: tabular-nums;
}

#yvamwdabos .gt_font_normal {
  font-weight: normal;
}

#yvamwdabos .gt_font_bold {
  font-weight: bold;
}

#yvamwdabos .gt_font_italic {
  font-style: italic;
}

#yvamwdabos .gt_super {
  font-size: 65%;
}

#yvamwdabos .gt_two_val_uncert {
  display: inline-block;
  line-height: 1em;
  text-align: right;
  font-size: 60%;
  vertical-align: -0.25em;
  margin-left: 0.1em;
}

#yvamwdabos .gt_footnote_marks {
  font-style: italic;
  font-weight: normal;
  font-size: 75%;
  vertical-align: 0.4em;
}

#yvamwdabos .gt_asterisk {
  font-size: 100%;
  vertical-align: 0;
}

#yvamwdabos .gt_slash_mark {
  font-size: 0.7em;
  line-height: 0.7em;
  vertical-align: 0.15em;
}

#yvamwdabos .gt_fraction_numerator {
  font-size: 0.6em;
  line-height: 0.6em;
  vertical-align: 0.45em;
}

#yvamwdabos .gt_fraction_denominator {
  font-size: 0.6em;
  line-height: 0.6em;
  vertical-align: -0.05em;
}
</style>
<table class="gt_table">
  
  <thead class="gt_col_headings">
    <tr>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">var_name</th>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">label</th>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">type</th>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">original_value</th>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">transformation</th>
      <th class="gt_col_heading gt_columns_bottom_border gt_left" rowspan="1" colspan="1">final_value</th>
    </tr>
  </thead>
  <tbody class="gt_table_body">
    <tr><td class="gt_row gt_left">item1</td>
<td class="gt_row gt_left">I like pizza</td>
<td class="gt_row gt_left">numeric</td>
<td class="gt_row gt_left">1 = agree, 2 = neutral, 3 = disagree</td>
<td class="gt_row gt_left">1 -&gt; 3, 2 -&gt; 2, 3 -&gt; 1</td>
<td class="gt_row gt_left">1 = disagree, 2 = neutral, 3 = agree</td></tr>
    <tr><td class="gt_row gt_left">item2</td>
<td class="gt_row gt_left">I don't like school</td>
<td class="gt_row gt_left">numeric</td>
<td class="gt_row gt_left">1 = disagree, 2 = neutral, 3 = agree</td>
<td class="gt_row gt_left">NA</td>
<td class="gt_row gt_left">NA</td></tr>
  </tbody>
  
  
</table>
</div>
<p><br></p>
</div>
<div id="format-of-documentation" class="section level4">
<h4>Format of Documentation</h4>
<p><strong>Repository Requirements</strong></p>
<p>If you are sharing data with a repository, check their requirements
before creating documentation. Each repository may have required or
suggested documentation formats. For example, <a
href="https://www.icpsr.umich.edu/web/pages/deposit/index.html">ICPSR</a>
requires documentation to be submitted in Word, ASCII or DDI XML format
and they in turn convert your documentation to XML or PDF format. <a
href="https://nij.ojp.gov/funding/data-archiving">NIJ</a> which has it’s
own suggested repository, National Archive of Criminal Justice Data (and
ICPSR sponsored archive), requests documentation to be in PDF, Word, or
ASCII format. You’ll also want to check if there are any metadata
standards your repository complies with, such as <a
href="https://www.icpsr.umich.edu/web/pages/datamanagement/lifecycle/metadata.html">DDI
standards</a>, and if that requires any additional considerations on
your part or if the archive does the work to make sure standards are
met.</p>
<p><strong>General Suggestions</strong></p>
<p>In general, sharing documentation in non-proprietary formats is
recommended. These are files anyone can open no matter what software
they own. PDFs and text files are generally good formats. If a
rectangular format is required, for documents such as metadata, csv
files can be created and still opened in any spreadsheet program.</p>
<p><br></p>
</div>
</div>
<div id="when-to-share-data" class="section level3">
<h3>When to Share Data</h3>
<p>Similar to what data and documentation to share, in terms of when to
share data, you want to consider both your funder’s requirements, your
repository’s suggestions, as well as your own personal preferences.</p>
<p><strong>Funder Requirements</strong></p>
<p>Your funder will most certainly have a suggested time frame for when
they expect your data to be shared. <a
href="https://ies.ed.gov/funding/datasharing_implementation.asp">IES</a>
for example states that study data must be shared no later than when you
publish the main findings of your study. This is a typical
recommendation from funders.</p>
<p><strong>Repository Suggestions</strong></p>
<p>If you are depositing your data in a repository, you may also find
they provide guidance on when to share your data. <a
href="https://www.icpsr.umich.edu/files/deposit/dataprep.pdf">ICPSR</a>,
at minimum, requires that users deposit data before the end of a project
so that any issues with the data can easily be resolved. However, they
also recommend to begin uploading data as soon as possible after each
wave of data collection. If you choose this method, ongoing data
deposit, you will want to consider if you would like to embargo your
data. Embargo means you deposit your data with delayed dissemination,
delaying the public release of your data until a later time point (for
example, after your study is complete). This allows you to continually
deposit your data (to reduce workload at the end of your study), while
also allowing you sole access to your data while your project is active.
If you choose to deposit data on an ongoing basis, make sure that your
repository allows you to edit, update and version your data if errors
are found or changes are needed at any point during the project.</p>
<p><br></p>
<p>Overall Helpful Data Sharing Resources:</p>
<p>📑 <a href="https://www.youtube.com/watch?v=24LTVqHi7hw">Sara Hart
gave a talk at the 2022 COS Unconference on Data Sharing that is very
insightful</a><br />
📑 <a
href="https://www.icpsr.umich.edu/files/deposit/dataprep.pdf">ICPSR’s
Data Preparation and Archiving Guide</a><br />
📑 <a href="https://www.ldbase.org/resources">LDbase has tons of helpful
guides</a></p>
<hr />
</div>
</div>
<div id="retraction-and-revisions" class="section level2">
<h2>Retraction and Revisions</h2>
<hr />
<p>Data Sharing is a great service to the larger community and that
service it is built upon trust. It’s built upon <strong>you</strong>
trusting that your own data is accurate and represents the true
information that you collected. It is also built upon users putting
<strong>their</strong> trust in you and your data; that the data you
provide are accurate and free from errors, and that the findings you
publish are based off of data that are free from errors and
manipulation.</p>
<p>This means, when you find errors in the data that you have publicly
shared, you have an obligation to do something about it.</p>
<p><strong>If your data is deposited in a repository:</strong></p>
<ol style="list-style-type: decimal">
<li><p>Consider making a comment in your project, notifying users of
errors in the data. If the repository requires/allows users to make an
account before accessing the data, they may have a system to email
current users to let them know a new comment was added to a project they
have downloaded.</p></li>
<li><p>If the errors in your data is fixable, take time to go back to
the raw data and re-clean, making the appropriate edits.</p></li>
<li><p>Upload a new version of your data to the repository. Make
comments about the revisions so that users know what changes have been
made between the previous and new version of your data.</p></li>
</ol>
<p><strong>If your data is not in a repository:</strong></p>
<ol style="list-style-type: decimal">
<li>Do the same as above but instead, consider personally reaching out
to anyone that has submitted a data request to you, to make them aware
of errors and changes in the data.</li>
</ol>
<p>If you have used your data in a publication, you will also need to
consider retracting the paper from the journal. Contact your journal to
make them aware of the errors you found. Consider the story of <a
href="https://laskowskilab.faculty.ucdavis.edu/2020/01/29/retractions/">Dr. Kate
Laskowski</a>, who began finding errors in her data, after publishing
with it, and decided she had no other choice but to retract her
publication. You can find the retracted article <a
href="https://www.journals.uchicago.edu/doi/full/10.1086/708066?mobileUi=0">here</a>.</p>
<p>An <a
href="https://www.sciencedirect.com/science/article/pii/S0213911118300724">article</a>
in Science Direct found that across a three year span, of all the
articles retracted in the journal PubMed, data processing errors were
the number two reason for journal retractions. Errors happen, we are all
human, but it is best to take the time and care to manage your data
during your project so that retractions are not necessary later on.</p>
<p>To learn more about reasons for retractions, you can read the blog <a
href="https://retractionwatch.com/">Retraction Watch</a> or search their
<a
href="https://retractionwatch.com/retraction-watch-database-user-guide/">database</a>.</p>
</div>


</div>
</div>

</div>

<script>

// add bootstrap table styles to pandoc tables
function bootstrapStylePandocTables() {
  $('tr.odd').parent('tbody').parent('table').addClass('table table-condensed');
}
$(document).ready(function () {
  bootstrapStylePandocTables();
});


</script>

<!-- tabsets -->

<script>
$(document).ready(function () {
  window.buildTabsets("TOC");
});

$(document).ready(function () {
  $('.tabset-dropdown > .nav-tabs > li').click(function () {
    $(this).parent().toggleClass('nav-tabs-open');
  });
});
</script>

<!-- code folding -->

<script>
$(document).ready(function ()  {

    // temporarily add toc-ignore selector to headers for the consistency with Pandoc
    $('.unlisted.unnumbered').addClass('toc-ignore')

    // move toc-ignore selectors from section div to header
    $('div.section.toc-ignore')
        .removeClass('toc-ignore')
        .children('h1,h2,h3,h4,h5').addClass('toc-ignore');

    // establish options
    var options = {
      selectors: "h1,h2,h3",
      theme: "bootstrap3",
      context: '.toc-content',
      hashGenerator: function (text) {
        return text.replace(/[.\\/?&!#<>]/g, '').replace(/\s/g, '_');
      },
      ignoreSelector: ".toc-ignore",
      scrollTo: 0
    };
    options.showAndHide = true;
    options.smoothScroll = true;

    // tocify
    var toc = $("#TOC").tocify(options).data("toc-tocify");
});
</script>

<!-- dynamically load mathjax for compatibility with self-contained -->
<script>
  (function () {
    var script = document.createElement("script");
    script.type = "text/javascript";
    script.src  = "https://mathjax.rstudio.com/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML";
    document.getElementsByTagName("head")[0].appendChild(script);
  })();
</script>

</body>
</html>