IAT experiment log file(Interpreting results data)

Comments

• Fab

  October 2021

  Hi @hanlyang,

  By default (i.e., if you select the "Log all values" option in the Logger object), the logged data will contain every single variable measured, which can seem a little overwhelming.

  

  If you're using OS, the default option (and safest way to go) is to log everything (default option). However, if you wish, you can disable the "Log all variables" option and use the variables inspector to drag into the Logger's list of variables those specific variables you wish to see in the log (you must be very cautious here not to forget anything crucial you might need later). If you're using OSWeb (generating a JATOS experiment to run it online), the compatibility check will tell you to select the variables you wish to log in order to decrease bandwidth.

  Regarding the meaning of the variables, some variables are constant and will be part of every OS experiment, others will be specific to your experiment.

  Some key constant ones include for example subject_nr, which is the subject number, or subject_parity (whether the subject's number is odd or even). You will also see variables representing general settings of the task (e.g., default font, font size, backend verison, etc.), variables for every object in the timeline and some of their properties (a count, which is the number of times that object is executed, a time, which correspond to the time at which that object was executed - time measured based on the computer internal clock). Certain variables indicate also properties of specific objects. For example, you'll see variables coding for the sound sampler's buffer size, number of audio channels, etc.

  Then you will also find a number of variables related to the variables you're using specifically in your experiment (whether you defined them in a loop table or using code - whether Python or Javascript). You'll find their value as well as when they occurred (time), how many times (count) etc.

  Then you have variables related to responses (including the actual value of the response logged (response), the response time (response_time, whether the response is correct (correct), etc.

  In addition to that, you also have the feedback variables corresponding to things like average response time or average % of correct responses across a block.

  I recommend that you read through this page for a description of the key variables:

  https://osdoc.cogsci.nl/3.3/manual/variables/

  Hope this helps.

  Fabrice.

  
• lvanderlinden

  October 2021

  Hi @hanlyang ,

  
  

  In addition to @Fab 's response:

  
  

  If you more or less followed the IAT tutorial from the documentation website, and are mainly interested in this contract

  

  the most important columns are:

  
  

  • "mapping", indicating whether the mapping was congruent or incongruent
  • "response_time", indicating how fast participants responded, and
  • "correct" (0 or 1), indicating whether participants responded correctly or incorrectly

  I guess you could/should also use the blocks in which participants only made positive/negative vs only young/old categorizations, as baseline performance.

  
  

  Does this help?

  
  

  Cheers,

  
  

  Lotje

  Did you like my answer? Feel free to
• hanlyang

  October 2021

  @Fab Thank you for the detailed information. It helped a lot sorting out the information I needed within the logger file.
• hanlyang

  October 2021

  @lvanderlinden Thank you for your comment.

  I got up with one question. You have mentioned a mapping column yet I couldn't find

  any information of setting a mapping command(sentence? How do I call it?).

  Is there a specific way in grouping the blocks?
• eduard

  October 2021

  Hi,

  with "mapping column" Lotje meant a column in the loop table that codes information on whether the combination of the other experimental variables that you code in the loop table are congruent or incongruent. This variable you have to create yourself in the loop table and fill the rows in the way that makes sense with your experimental design. If you have done that, the column mapping should also appear in your log file. If not, then it won't be there of course.

  Hope this helps,

  Eduard