What software program are your tasks in / what is E-Prime?
Our lab uses E-Prime software to design and run most of our tasks. E-Prime is a Psychology Software Tools (PST) program specifically designed to run experiments. It is not free and does require a license. For more information, go to www.pstnet.com/eprime.cfm. Note that we are in no way affiliated with E-Prime except in the capacity of a customer. We do not receive free or discounted E-Prime licenses or other benefits from PST.
Can I obtain your tasks in a program other than E-Prime?
Our lab primarily used E-Prime to program and run our experiments. A few of our tasks were also programmed for Inquisit and Mel, which we have on our site for download. Please note we are unable to provide support for our tasks that are in a software package other than E-Prime. Recently, some labs have made online versions of our tasks and others available. We cannot verify the validity of these tasks, however, so we cannot endorse their use.
How can I obtain E-Studio files of your tasks?
We are happy to send you the E-Studio file if you create and sign a brief formal agreement saying:
- 1) You will not change the nature of the task
- 2) You will not distribute the task
- 3) You will cite our lab in future publications in which you used our tasks
- 4) You will report any changes you made to the task so that your results can be replicated
Note - We do not have access to E-Studio files for the Oswald et al. (2015) shortened tasks. Please email Tom Redick for those.
Can I make edits to your tasks?
We have no issue with researchers making slight modifications. For example translating the task to another language, making modifications to the instructions, adding a questionnaire, making aesthetic font/color changes, etc. However, we request that you do not change the nature of the task such that the task is no longer measuring what it was designed to measure (i.e., if you convert the processing trials in the Operation Span to geometry problems or change the memory stimuli to words, this is no longer the Operation Span task). If you do make any modifications, again please report any changes you made in future publications/presentations so that your results can be replicated.
Are E-Prime 2 and E-Prime 3 compatible?
Yes. PST ensured its users that compatibility issues would be less problematic for E-Prime 3, and thus you can run and convert E-Prime 2 files in E-Prime 3.
Will you have tasks available for E-Prime 3 soon?
Maybe. We are still using E-Prime 2 since 3 was released after we had already programmed for the current study. After this study is over, we will consider upgrading to E-Prime 3 and making these tasks available. Though while we have found E-Prime 3 to be stable and reliable in our testing, we are not a fan of the new interface. For this reason and because E-Prime 2 files are compatible with E-Prime 3, there is a chance that we will not have E-Prime 3 files up for a while.
Are E-Prime 1 and E-Prime 2 compatible?
No, E-Prime 1 and E-Prime 2 are not compatible in any way. Seeing as we only have E-Prime 2, we cannot edit or open any of our E-Prime 1 files.
Are older versions of E-Prime 2 compatible with newer ones?
You can run older files if you have a newer version - an error message may pop up saying that the file is too old, however just hit “Okay” and the file will run just fine. However, you cannot run versions newer than what you have (e.g., you cannot run 188.8.131.523 files with 184.108.40.206).
Do I have to have E-Prime Professional to run your tasks?
Some of our tasks were created using professional, particular tasks with sound files. These tasks can not be run in standard E-Prime. However, other researchers have had luck emailing PST about this issue, and they have agreed to make standard versions of professional tasks for people.
How long does each task take?
The approximate amount of time it takes for 95% of subjects to complete task:
- Symmetry Span & Rotation Span: 20 minutes
- Running Span: 25 minutes
- Operation Span & Reading Span: 30 minutes
At the end of the experiment, a red screen appears saying to get the experimenter. How do I exit this screen to access the E-Data (.edat2) file?
Most of our tasks are designed so that once the red screen comes up, the ‘Q’ key (for quit) will exit the program and generate the E-Data file.
I have a data file in the form of a .txt file but I cannot find the .edat2 (E-Data) file. Can I read the .txt file? Why does this happen?
E-Prime only generates an .edat2 file if the experiment is successfully terminated at the end (by hitting the ‘Q’ key), but it always generates a .txt file. If you abort an experiment and still wish to access the data, you can easily turn the .txt file into an .edat2 file by using the E-Recovery program that is part of the E-Prime software package.
I cannot run the E-Run file because the file is from a newer version of E-Prime.
E-Prime files from newer versions are not typically compatible with older versions. If you are using 220.127.116.11, for instance, you may have issues running 18.104.22.1683 files. In this case, you can either request the password for the 2.0.8 tasks or you can request the E-Studio 22.214.171.1243 files and recompile the E-Run file on your own machines. Also note that if you have a valid E-Prime license, you can go to their website to obtain the most recent version of E-Prime for free.
In the E-Data files, I see a lot of variables. What do they mean and where do I find the span score?
The variables are different depending on the task. That being said, you should be primarily interested in the partial score (which has more variance and thus allows for better discrimination between high and low ability participants; please see Conway et al. (2005) for a detailed discussion of the span tasks, including scoring issues).
The partial span score is the total number of items recalled in the correct order on memory trials. For the absolute score, the participant only received credit if they were 100% accurate within that trial (i.e., set size). For example, if the participant is performing a trial of set size 3 and answers all 3 correctly, both the absolute and partial score will be a 3 for that trial. However, if the participant answers 2 of the 3 correctly, the partial score will be a 2 and the absolute score will be a 0 (because the participant was not 100% accurate on that trial). The partial score will always be equal to or higher than the absolute score.
For most tasks, you will see variables called “SpanAbsoluteScore” and “SpanPartialScore” (e.g. in the Symmetry Span these are “SspanAbsoluteScore” and “SspanPartialScore”).
For other tasks, you will see variables called “SpanScore” and “SpanTotal” (e.g. in the Reading Span these are, “RspanTotal” and “RspanScore”). The variable ending in “Total” is the partial score, and the variable ending in “Score” is the absolute score.
What about errors on the processing (distractor) trials?
You may be interested in errors on the processing trials (e.g. the math problem in the Operation Span, the symmetry judgment in the Symmetry Span) as well as the span score. The processing trials have numerous functions, but primarily they are designed to prevent the participant from rehearsing information in short-term memory to ensure that the span score reflects processes associated with executive attention and working memory rather than short-term retention and rehearsal. The score for the processing trials will show up as (in the case of the Symmetry Span) “SymmetryCorrect” or (in the case of the Reading Span) “ReadingErrorTotal.” Therefore it may give you the number of correct processing trials and/or errors on processing trials, and each task has a different amount of processing trials. Note – only our complex span tasks have processing trials, and thus the running span and counting span do not have them.
Our lab’s policy is generally to remove subjects whose accuracy on the processing trials is below 85%. This is admittedly an arbitrary rule of thumb, and it’s not so much the actual accuracy that matters but more so if the participant was attending to the processing trials. Thus a participant with low arithmetic abilities may score 80-85% on the math problems in the Operation Span while putting full effort into the problems. If this were the case, we would keep this participant’s data. We highly recommend having the experimenter observe each participant on at least some of the processing trials to get an idea of if the participant is putting effort into these trials or not.
How much time are participants given on the processing (distractor) trials?
The participant is given their mean + 2.5 standard deviations to respond to each processing trial. For example, if the participant’s mean RT is 1000ms to make a symmetry judgment (in the Symmetry Span) with a standard deviation of 200ms, then they will be given 1500ms to respond to each symmetry judgment. The baseline value used to create the response deadline is calculated from each participant’s responses on the practice processing trials, and so it is important that the participant is giving full attention and effort to the practice trials. If during the experiment it appears the participant is not being afforded enough time on the processing trials, it is likely that they were responding too quickly during the practice trials and the task may need to be restarted.
We have no issue with researchers making slight modifications (e.g. translations, adding a questionnaire, aesthetic changes, altering set sizes) to our tasks. We are happy to send you the E-Studio file if you fill out and sign a brief formal agreement saying you will not change the nature of the task, you will not distribute the task, you will cite our lab in publications using our task, and you will provide a detailed outline of any changes you made to the task so that any results you find can be replicated. You can contact (firstname.lastname@example.org) for the E-Studio files.
What are the shortened tasks that are on your website?
These are tasks designed by researchers trying to shorten the amount of time the complex span tasks take without affecting the psychometric properties of the tasks. The Foster et al. shortened tasks are from our lab, and allow the researcher to indicate how many blocks (out of 3) they want to use for each task. The Oswald et al. were created in part by former lab members, and were shortened in a different manner as Foster et al. We do not have E-Studio files of the Oswald et al. tasks. For access to those, please email Tom Redick.
Shortened complex span tasks can reliably measure working memory capacity. Memory and Cognition.
Our shortened complex span tasks allow the researcher to indicate how many blocks (out of 3) they want to use for each task. Using a single complex span task is not appropriate for measuring working memory capacity as a construct - and we strongly suggest using multiple tasks if you would like to obtain a reliable and valid estimate of working memory capacity (these are true of both the original and shortened versions of our tasks). Thus the shortened tasks are designed to be used together to allow researchers to obtain a valid estimate of working memory capacity while taking less time than using the full-length versions of all tasks.
In our recently published paper, a graph and other data are presented that show how much time certain combinations of blocks of tasks take along with how reliable and valid these combinations are. We strongly recommend taking a look at the paper if you want to use the shortened tasks.
The development of a short domain-general measure of working memory capacity. Behavior Research Methods.
Automated operation, symmetry, and reading spans described elsewhere (Unsworth et al., 2005,Memory; Unsworth et al., 2009, Memory; Redick et al., 2012, European Journal of Psychological Assessment) were modified in the following ways:
- - For operation and reading span, instead of presenting 15 trials (3 each of list lengths 3, 4, 5, 6, & 7) with a maximum possible score of 75, 6 trials are presented (2 each of list lengths 4, 5, & 6) with a maximum possible score of 30
- - For symmetry span, instead of presenting 12 trials (3 each of list lengths 2, 3, 4, & 5) with a maximum possible score of 42, 6 trials are presented (2 each of list lengths 3, 4, & 5) with a maximum possible score of 24.
- - For all tasks, the number of storage-only and processing-and storage practice trials is 2 (versus 4 and 3, respectively).
Currently, there are no normative data available for these versions of the tasks. Study 2 of Oswald et al. (2015) contains descriptive statistics and correlations from a sample of N = 172 Purdue University undergraduates.
If I use your tasks, what articles should I cite?
If you are publishing a study in which you used our tasks, please cite the following:
If you used any of our tasks:
An automated version of the operation span task. Behavior Research Methods, 37, 498 - 505.
If you used shortened versions of tasks:
Shortened complex span tasks can reliably measure working memory capacity. Memory and Cognition.
If you want to know more about the reliability and validity of our tasks, please reference and cite:
Measuring working memory capacity with automated complex span tasks. European Journal of Psychological Assessment 28,164-171.
For another detailed discussion of the tasks, please reference and cite:
Working memory span tasks: A methodological review and user's guide. Psychonomic Bulletin and Review, 12, 769 - 786.