May 9, 2023 at 9:07 am #10528ElRkParticipant
In the past months I’ve tried many times to read my data and also make clear how exactly R works.
When I open my results file in Excel, each four lines are dedicated to one experimental item-sentence.
Generally, the results appear in a very confusing way (i.e. not categorized) but I’ve managed to locate the info I need like the participant’s name, the item (under the column Penn Element Name) and the scale choice.
The problem is that when I run the file in R I can’t find anywhere the item’s information.
Do you have a possible solution in mind?
ElMay 9, 2023 at 10:18 am #10532JeremyKeymaster
You have four lines per trial because, by default, each PennController trial reports one line for when it starts and one line for when it ends. Then your trials use the
logcommand on an Image element and on a Scale element, which will report two more lines per trial, one for each: one line reports when the Image element was printed onto the page, the other line reports what value was selected on the scale + when the selection occurred. This way, if you need to calculate the time difference between any pair of those four events, you have the option to do so. If all you care about is the value that was selected on the scale, you can just subset the table to those rows that report about the Scale element
This page of the advanced tutorial illustrates how one can parse the data in R. In particular,
read.pcibexwill automatically name the table’s columns using the comment lines from the results file (the line starting with
#) and the command
filtercan be used to subset to certain rows only, for example
filter(PennElementName == "side-by-side" | PennElementName == "selection") %>%keeps only the rows that correspond to the elements named “side-by-side” or “selection” in the tutorial’s study
JeremyMay 9, 2023 at 11:22 am #10534ElRkParticipant
Thanks a lot Jeremy! I’ll certainly use this command.
I have one last question. In the results section of pcibex, some numbers appear under the forth column which is named as ‘Item’.
What are they referring to? I haven’t understood.
For example, a certain element which is 35th in the row in the experiment has the number 35 under ‘PennElementName’ and the number 40 under ‘Item’.May 9, 2023 at 11:31 am #10535JeremyKeymaster
The “Item” column reports the item number according to Ibex. The other number corresponds to the PennController trial’s index: since PennController trials do not necessarily constitute the entirety of the study’s items (and certainly, in your study, you have non-PennController items, for example the one labeled “intro” does not contain a PennController trial but instead contains a single Message controller) the two number do not necessarily align. The example you mention means that the 40th item in your study corresponded to the 35th PennController trial, and that there likely are at least 5 items that are not PennController trials (I say “likely” because there are other potential sources of mismatch between Ibex’s item numbers and PennController’s trial numbers, eg. one could create some PennController trials first in a certain order and pass them as Ibex items later in a different order)
- You must be logged in to reply to this topic.