Thursday 22 October 2020

E-Prime Go: The Tutorial

I wrote about E-Prime Go before (here and here), but as this infernal pandemic keeps plaguing us, it's been necessary to get it ready for prime time. Luckily, the team from PST has done an admirable job in ironing out most of its quirks and it's now doing a good job for those of us who can no longer rely on our behavioural labs. How do I go about getting my normal experiment online using E-Prime Go?

Step 1: Get your experiment ready. 

This is the easy step. First, we move all auxiliary files, like images and sounds, over to a specific locations if you haven't done so. All such files will be automatically packaged as one single file, so if you have big 'movie' files, you better prepare the subjects for a massive download. They don't need to be all in the same directory as such, as E-Prime Go will search for them, but it helps to keep track of things. 

1. In the experiment properties, I change the Display option 'match desktop resolution at runtime' to Yes (which PST does not recommend, but from experience I know not all monitors do a good job at changing resolutions, and I made sure my experiment looks OK at various different resolutions). 

2. I also turn the Sound device off by removing the checkmark next to it, because I am not using it. This means that if subjects have a browser running in the background or some music software, it's possible they can continue listening to the music. Obviously that isn't ideal, but I ask them to stop background programs anyway, so we'll just have to hope they comply. The reason I do this is because some audio interfaces that do not have an audio cable plugged in are promptly turned off completely by Windows, which means that E-Prime throws an error: You want to prime the audio system, but you don't even have a soundcard. It's made a mess of one my classes before and I'm not falling for it again.

3. My immense amount of code, is it ready for online? Turns out that yes, basically it always is (unlike Psychopy). Of course, if you use canvas drawing (I do), you will need to remember that pixel drawn at coordinate (1600,1) will be visible on one screen, but not on another (a good reason for NOT matching desktops!), and perhaps throw an error. Windows also does funky things when people have two screens running in extended mode. For this reason I tell people to turn presentation mode to single display mode by pressing the windows button and P until it says 'PC screen only'. 

And that's that.

Step 2: Load it in E-Prime Go and add additional information

Make sure you have the latest version as development is going at a pretty fast pace. I am now using E-Prime Go, freshly downloaded from PST's support page.

1. In the top right corner, open the experiment you want to be put online - only the compiled version (.ebs3) not the original code (.es3) is used. A popup will show for 'choose files to be included in your experiment'. As I mentioned, EPG does a pretty good job at sorting out what needs to be included, including various pictures I do use (700x700circle.png), a movie file (not shown in the image below), but also some obviously useless files, such as a previous, compressed version of the complete experiment ( So unselect these. Too much is better than not enough, but try to keep the size down for the occasional boomer who is still using a dialup modem. (apologies, bad joke). 

E-Prime Go: Loading up the experiment

2. Write the Disclaimer, Start, Stop, and Failure messages. In particular, the Disclaimer and Start messages are important, as these give the last chance you will be able to tell the participant important info such as their rights, and the need to turn the experiment off. I keep all the messages in a single .txt file near the experiment so I can copy-paste them for every new version of the experiment, and after a while I came up with the following. 

E-Prime Go: Information for the subject
Top left, disclaimer. Disclaimer suggests some sort of Google information blurb that absolves them from legal problems, but here, you might simply see it as the first screen participants see after downloading and running the experiment. I tell them this:
Thank you for agreeing to participate in this experiment. Before you begin the experiment, make sure:
1) You have no other software (e.g. browsers, Skype) running in the background. 
2) You have ... minutes available for doing the experiment. There will be opportunity for breaks.
3) Your environment is quiet and free from distractions. Please turn your phone in flight mode during the experiment. 
4) You are currently using only a single display. If you use multiple screens, please press WIN+P and change the projection mode to PC screen only (or go to Windows Display settings and change the multiple display settings to use only one screen). 
Pressing next will give important information on the nature of the experiment and your rights as a participant. Following, if you provide informed consent to continue with the experiment, you will be asked for a subject and session number. If the lab assistant provided you with a subject and session numbers, please use these. 

Bottom right, start instructions. So, in the disclaimer, I tell of everything they should do before even considering the experiment. I already gave the participants information on their rights along with the download link, so they should already be familiar, but we tend to think the rights are important, and I reserve these for the start instructions:

This is an experiment on time perception. [two sentence explanation]. 

The task will take ... minutes in total, so please make sure you have this much time available. If you still need to end participation prematurely, you can stop execution at any time by pressing CTRL+ALT+SHIFT, which will terminate the experiment. This will not save any data and you will not be able to continue where you left off. However, ending your participation is one of your rights and this will have no negative consequences for you. 

If you understand the nature of the experiment, please review your rights as a participant: 

1) You will be free to end participation at any time, including up to 3 months after participation and your data will be removed.

2) You are free to enquire for more information at any point (e.g. by email).

3) You will not be deceived at any point (we are not sneakily testing something else entirely).

4) Your data will be anonymized and your privacy protected.

Continuing your participation indicates you acknowledge and agree with these terms. Click on RUN TASK below to start the experiment. Please be patient, this may take a minute.

We all know E-Prime is pretty hard to stop once it's running, so I do tell them about the ctrl+alt+shift option, though I imagine they might have forgotten when it's actually needed. So far, however, I have had no problems. Finally, there are the two text fields that show up after the experiment. These are somewhat self-explanatory, but I'll quote what I have there anyway for completeness.

Top right: stop instructions. These show up if the experiment completed without errors. Depending on whether you use the later automatic upload solution of E-Prime Go (I'll explain later) or the manual one, you will need to inform participants in two ways. In the former case, they're basically done, and maybe they can inform you. In the latter case, you need to tell them where to find the data (writing @folder will make a clickable link to the data) and what to do with them (e.g. email them to you). I'm now using the automatic upload, and this is what I write:
Thank you for your participant. The experiment is now finished and the data are automatically uploaded. Please inform [my email] that you have finished. He will also be able to answer any questions you might have regarding the experiment. 
Actually, quite often they have questions, so I made a little google doc in which I explain to such subjects what I more or less expect from the data and why I'm doing this research anyway (for the l0lz, obvs). 

Bottom right: failure instructions. Obviously, this is bad news. But you want two things: to prevent more problems, and to keep your participant happy. If you just tell them to inform you they will tell you "it broke", which is not terribly informative. So I give them a few hints. 

Unfortunately something seems to have gone wrong and the task ended early. Please provide as much information as possible on why you think this happened and what error message you saw, perhaps including a picture (e.g. using your phone). I would be grateful if you could send these to [my email] so I can find out what happened and prevent future problems.

Then finally, at the bottom, note the Options: Automated or Manual. The manual option means you will send participants a file with an executable (either by email - bad idea, looks like a virus) or by placing it on google drive or such. As mentioned above, and explained in detail in my first post on this topic, you will also need some solution for uploading the data. There are many ways this can go wrong, so I recommend Automated. However, try to fill in the text fields first and run the experiment in manual at least once to see if you're happy with what it looks like.

Step 3: Upload it to 

The experiment is now ready to go online. I wrote a lot of text above, but that's just because I get paid by the word, so don't worry: steps 1 and 2 each take about 1 minute if you already have a previous text document with the disclaimer and start text. So too with the final step. Clicking on automated instead of Manual and pressing Pack gives you popup with email and password. If you have an E-Prime account, you fill those in, otherwise you may ask over at the local police office for information on how to turn yourself in. My account is still registered under an old email, so I suppose I shouldn't be too judgemental here!

Pressing OK, it packs the files and presents 3 options: go to and close this application, close this application, or return to this application. What to do if you want to go to but not close this application?! To be honest, instead of the 3 options with the 4th missing, how about just a button with "Take me to my experiment on"? Most of us know how to close windows. I go with the first option anyway (closing E-Prime Go against my will), which only means the default browser opens Once logged in (again), you see your experiments, including the new one we just made:

DrawingTime being the new experiment
All good. There is now also the previously missing possibility for deleting old experiments. Clicking on DrawingTime presents you with the meat of the matter (that can be soy if you're vegetarian like me). In the first tab, Experiment, there is just a big button that says 'copy download link'. You click it, and this is what you paste it in the email you send to participants. In the second tab, you'll see the data that have been collected up until now. Obviously, my new experiment doesn't have any, but its predecessor, the super originally named DrawingTask does:

Experimental data on the server

The coolest feature is the Last status, which shows 'completed' if the data are in (yay), but also will tell you if a subject has started, so you can keep an eye on the data collection progress, much like a devoted eBay consumer does on their products. 

Step 4: Send the link to participants
We're more or less done. There's still no data collection platform ala Amazon Turk (but more ethical), which is a shame, so we need to do our recruitment ourselves. After sending a request to mailinglist and getting some people in who are interested, I thus send an email like this:

What it is about:
[four sentences explaining what the experiment is about]
How long will it take and what to do if you need to interrupt:
[mentioned in the start instructions]
Before you begin make sure:
[mentioned in the disclaimer]  
What are my rights as a participant:
[the rights again] 
How do I begin?
Clicking on the following link indicates you agree with these terms:
[paste here the link]
At the beginning of the experiment, you will be asked to give a subject number. Pleasure use number [...]. Write this number down on a piece of paper (or any other surface!).
What do I get in return?
After you finish, email me to let me know, along with the random number you wrote down. Also let me either know whether you want to paid, or me to send the money to charity ([charity 1], [charity2]). I also have an easy-to-read explanation of what I’m hoping to find out, if you’re interested. 

Note the link, which is what you copied in Step 3. The subject number makes it easier to keep track of who's done and who needs to be reminded. Finally, I'm pretty proud of my new charity work, because it means my experiments now have at least some practical use (Amnesty and animal shelter are the current options). It's also vastly easier to pay one charity once than many participants personally, and it makes you feel better afterwards. I also got some participants because the charity itself decided to help me in recruitment, so it's really a win for everyone!

No comments:

Post a Comment