Protocol for tagging papers
You will see something like:
Click on the
Unreviewed link on the top navbar. and you'll see all the publications that still need to be processed. You can select a subset of the database by adding keywords:
Otherwise you can select one of the whole list.
Click on any of these publications to start processing it.
Find the full text of the publication
First try clicking on the
Title and Author link to search for the publication - oftentimes ASU Libraries will have a subscription to the journal that allows you to download a PDF or view the publication online. You may need to refine the search parameters or pick the correct paper out of several search results. If this doesn't work you can also try to find it via Google Scholar, or try to access the journal itself via the ASU Library / Google Scholar. If the ASU Library does not provide access to the publication you can also request a copy via https://illiad.lib.asu.edu/illiad/ - If you're completely unable to access the text of the paper, add a note and
Flag the paper.
Skim the paper
Does the paper base its findings on a computational model? Does it describe a computational model? False positives include review papers and "meta" papers that discuss methodologies for model development. Sometimes there are also papers with mathematical proofs without any agent-based model, which is also a type of paper we do not consider. Since we look at the replicability of model results, a good indicator for a relevant paper is that model results are shown in graphs and/or tables.
If the answer is NO, change the status to “Not Applicable” on the page.
If the answer is YES try to locate the following data:
Corresponding author name and email
Find the corresponding author (they will sometimes have a small envelope icon next to their name, or they will be listed explicitly as the corresponding author). Enter their name in the
Contact author name field, and their email in the
Contact email field.
Model source code URL
Does the paper make reference to the model source code being available? Sometimes a link is provided in the paper, referring to an online appendix, or a webpage of the author. Check if these pages still exist.
If the paper says that the code is available at http://www.princton.edu/~jsmith/mymodel.html check that URL in your browser. If the URL gives a 404 error or does not contain the actual source code for the model, click the
Add Note button at the top right of the screen and note that there was a URL listed but it did not resolve properly. Please include the broken URL as well in the note.
Otherwise, enter the URL to the source code in the
Code archive url field, e.g.,
If you cannot locate a URL for the model code, leave the entry for
Code archive URL blank.
The modeling platform is the programming language and framework used to implement the model. Common modeling platforms include NetLogo and RePast, a more comprehensive list is maintained at https://www.comses.net/resources/modeling-platforms/.
If there is no platform listed, a paper may list the programming language used to implement the model, this
could be something like
Java, Python, Visual Basic, C, C++, C#, etc. Note that you see sometimes that the model is written in for example Java using Mason libraries. In that case we list the platform Mason and not also the java language.
If you don't find any platform or programming language information, leave this field blank.
NOTE: This is an autocomplete field, please look for and use existing values instead of creating new ones. You can also enter multiple values.
The sponsor is the funding agency or organization that provided funds to conduct the research. This information is typically listed in the acknowledgements. If there are multiple sponsors, create multiple tags.
There is no need to list specific grant numbers. If a specific branch of the funding agency is listed, enter the broader funding agency, so if you see something like
National Institutes of Health Small Business Innovative Research (SBIR) just enter
National Institutes of Health.
NOTE: This is an autocomplete field, please look for and use existing values instead of creating new ones. You can also enter multiple values. NOTE2: Since we look for external funding, we do not include departmental funding or grants from the university itself.
Model Documentation Protocol
How is the actual model itself documented? There are several options here.
The ODD is a structured narrative to organize the model description. The model description follows a systematic sequence of items (typically using the heading names as listed in the Figure below). First it discusses the purpose of the model, then the state variables, then the process overview, etc.
When a model is described using
ODD it will usually make reference to
ODD explicitly. Most papers will fall into the
Other Narrative category, where they describe the model, just not in the same structured narrative format as
AORML is special case of UML, called
Agent Oriented UML and is rarely seen.
If you have provided a URL for the software code, you also include Source Code in the list of Model Documentation.
Finalize the status
When you have finished all the entries and you have a url for the model code, you can change the status to
Reviewed: Publication meta data reviewed and verified by CoMSES
If you have any questions about the publication or are unsure about it, change the status of the Flag (right bottom of the screen) and use
Add Notes to record your questions.