Exporting Study Data
Exporting the Data
During the study, participants' behavioral data on the website is recorded in the MongoDB database that you have set up. MongoDB is a NoSQL Server in which data is stored in BSON (Binary JSON) documents and each document is built on a key-value pair structure. The details of this are not too important if you are not familiar with MongoDB.
Instead, a script (data-export.js) is provided in The Truman Platform project folder that when run, exports basic variables and information about the participants' behavioral metrics on Truman into a readable csv file. This eliminates the need for you to go into the database (which can be overwhelming) and allow you to easily analyze participants' behavioral data at the end of your study.
How to export data into readable csv file
To run data-export.js to export participants' behavioral metrics data into a csv file:
Ensure the value to the key MONGODB_URI= in the
.envfile is the MongoDB URL connection string to your database. If you followed the instructions to installing Truman, this should already be set to the correct value.Create a folder named
./outputFilesin your base project directory. Exported csv files will go in this folder.Next, enter
node data-export.jsin your terminal/command prompt from your project directory. A new exported csv file is generated each time you enter this command and run the script.When the script is completed, you will find the csv file with the data in the folder
./outputFiles/truman-dataExport-.........csv.To understand this file, see below.
Data exported
In general, these are the types of behaviors calculated/ exported by the data-export.js file.
All interactions with posts and comments (including likes, flags, view times)
Participants' posting behavior
Actor based interactions (Block, report, follow, profile view)
Site log (time on site)
Page log (pages visited)
Participant information (MTurkID, username, experimental condition)
Below is a more detailed breakdown and description of each variable.
In the outputted csv file, each row is one participant and the value in a column corresponds to the variable column name.
Id
Participant's provided Mechanical Turk ID
ResponseID
Username
Participant's username
Condition
Participant's condition
NumUserPostsCreated
Number of posts made by the participant
NumUserCommentsCreated
Number of comments made by the participant
NumActorPostsLiked
Number of actor posts liked by the participant
NumActorPostsFlagged
Number of actor posts flagged by the participant
NumActorCommentsLiked
Number of actor comments liked by the participant
NumActorCommentsFlagged
Number of actor comments flagged by the participant
UserPostsCreated
Metadata of the participant-made posts, including: - text of the post - which day of the study the post was made (ex: 1, 2, 3, etc., where the day number is defined in 24-hour periods of the study) Formatted like: <TEXT>, on Day <DAY>
UserCommentsCreated
Metadata of the participant-made comments, including: - text of the comment - id of the post the comment was left on - which day of the study the comment was made (ex: 1, 2, 3, etc., where the day is defined in 24-hour periods of the study) Formatted like: <TEXT>, on Post <POST ID>, on Day <DAY>
ActorPostsLiked
List of ids of the actor posts the participant liked
ActorPostsFlagged
List of ids of the actor posts the participant flagged
ActorCommentsLiked
List of ids of the actor comments the participant liked
ActorCommentsFlagged
List of ids of the actor comments the participant flagged
ActorsBlocked
List of actor usernames the participant blocked
ActorsReported
List of actor usernames the participant reported, and the reason for reporting Formatted like: <USERNAME>, Reported for <REASON>
ActorsFollowed
List of actor usernames the participant followed
TimeOnSite
Total amount of time (in seconds) participant spent on site
PageLog
List of page URLs the participant visited
Analyzing the Data and Paying Participants
You can use programs like R or other analyzing tools to analyze the data, find interesting trends and statistical significance, and compare different conditions.
The key that unites the participants' data with any Qualtrics pre-study surveys or Qualtrics post-study surveys results that you have used in your study is the ResponseID (see Integrating Truman with Qualtrics for more details)
You may also use the Mechanical Turk ID value of participants and the exported data to identify and help with payments/ rewards after the study.
Last updated