Remem­ber­ing What’s Mine: How the Own­er­ship Shop­ping Par­a­digm Went Online.

Tes­sa Clark­son:
Okay. Hi, every­one. Today, I’m going to be talk­ing a bit about my PhD research. Before I get into it, I’m just going to explain my top­ic of inter­est. I study some­thing called the self-ref­er­ence effect. The self-ref­er­ence effect is, if you haven’t heard of it before, a mem­o­ry bias. It is the ten­den­cy for peo­ple to encode infor­ma­tion dif­fer­ent­ly, depend­ing on how that infor­ma­tion is rel­e­vant to them. So typ­i­cal­ly, when peo­ple are asked to remem­ber or recall infor­ma­tion, they’re typ­i­cal­ly bet­ter at doing this when that infor­ma­tion was rel­e­vant to them in some way. In psy­chol­o­gy, there’s many dif­fer­ent ways you can mea­sure the self-ref­er­ence effect and oper­a­tional­ize it. One way you could do it is through own­er­ship. So this par­a­digm I’m about to describe was orig­i­nal­ly designed as a par­a­digm to occur between two par­tic­i­pants in the lab, or a par­tic­i­pant and oppos­ing confederate.

Tes­sa Clark­son:
If you can imag­ine a bird’s eye view, we’ve got a par­tic­i­pant and oppos­ing con­fed­er­ate here, they enter into a lab set­ting and they’re giv­en own­er­ship over one of two bas­kets. Then the exper­i­ment just shows them a bunch of visu­al stim­uli on pic­ture cards, and these items are then sort­ed as being owned by the self, so owned by the par­tic­i­pant, or owned by the oth­er, and owned by the con­fed­er­ate. So, that’s encod­ing, and then par­tic­i­pants are sur­prised with a sur­prise recog­ni­tion mem­o­ry test, where they’re shown all that same visu­al stim­uli again, plus new items they haven’t seen before, and they’re asked to iden­ti­fy whether or not they remem­ber that item or not. What the researchers found here was that self-owned items, that were encod­ed as being self-owned, were remem­bered more accu­rate­ly than oth­er owned items, and also rec­og­nized more quick­ly. This is the self-ref­er­ence effect through ownership.

Tes­sa Clark­son:
Since the devel­op­ment of that par­a­digm, oth­er researchers have been able to repli­cate this, but through a com­put­er­ized vir­tu­al ver­sion. This is the task that I used, or was using in the lab. Instead of hav­ing real bas­kets and real items, I would show par­tic­i­pants two bags on a screen. And if you can imag­ine what a tri­al would look like, if you look at that bot­tom square, you can see the bags fol­lowed by an item, then a col­or cue that would indi­cate own­er­ship, and the par­tic­i­pant would use com­put­er keys, the left or right arrow keys, to sort those items. Then I do the same thing again, show all those same items again, plus new items that they haven’t seen before, and ask them know who owned that item. Typ­i­cal­ly, we find the same effect, where stim­uli that’s encod­ed by the self is remem­bered bet­ter. So this was the evo­lu­tion of a par­a­digm that went from a face-to-face par­a­digm and became com­put­er­ized. Now I’m going to be talk­ing about how I took that task that we did in the lab and how I made it a web-based paradigm.

Tes­sa Clark­son:
So this is just one of the many self-ref­er­ence effect stud­ies that me and my lab have been doing, but this is one I did recent­ly. We used GORILLA, and so we cre­at­ed two online ver­sions of this shop­ping task. One ver­sion we had… Before the par­tic­i­pant did the task, they engaged in a con­ver­sa­tion­al sit­u­a­tion with the oth­er per­son that they were going to be doing the task with. They learned cer­tain details about that per­son, and learned about their inter­ests and val­ues and what­not. And in the sec­ond ver­sion of the exper­i­ment, we did­n’t tell them any­thing about the oth­er per­son. We just said, “You’ll be doing a shop­ping task with anoth­er per­son.” And they were com­plete­ly nameless.

Tes­sa Clark­son:
The results were real­ly inter­est­ing. What we found was that in the first exper­i­ment, as indi­cat­ed by this first white line here on the graph, was that when we gave par­tic­i­pants con­text and infor­ma­tion about that oth­er per­son, they remem­bered the oth­er per­son­’s items bet­ter than their own. So this is going in the oppo­site direc­tion of the self-ref­er­ence effect. And in the sec­ond ver­sion, that was stripped down, where we did­n’t pro­vide any infor­ma­tion at all, this nor­mal self ref­er­ence effect emerges. So we found this real­ly cool bound­ary effect, where we found the bound­ary of where the self-ref­er­ence effect no longer applied.

Tes­sa Clark­son:
Before we moved online in the lab, there were some con­cerns or anx­i­eties that we had, nat­u­ral­ly, that I think most peo­ple have. So now I just want to give some rec­om­men­da­tions for things that I’ve learned since going online, that I think would be help­ful for any­one who wants to do any sort of mem­o­ry behav­ioral task. So these are some things that I rec­om­mend and learned along the way.

Tes­sa Clark­son:
Uti­lize full screen if you can, if you’re doing a behav­ioral exper­i­ment. If you can get your par­tic­i­pants to engage in… If they’re using a brows­er to do the study, if you can min­i­mize dis­trac­tion as much as pos­si­ble, get them to engage in the full screen. But just make sure that you give them instruc­tions on how to exit full screen and regain con­trol, should they wish to with­draw at any time.

Tes­sa Clark­son:
Pilot your exper­i­ments. We talked about this, it was brought up yes­ter­day, how impor­tant pilot­ing is. But pilot your exper­i­ments, specif­i­cal­ly with non-lab mem­bers, to test the effi­ca­cy of your instruc­tions. If you’re not going to be there to ver­bal­ly instruct and guide your par­tic­i­pant through, test­ing with non-lab mem­bers is real­ly impor­tant to make sure that that task is ful­ly under­stood. Set time lim­its. Make sure that you set time lim­its for com­ple­tion, this will dis­cour­age inter­rup­tions and pre­vent peo­ple from walk­ing away. And if they do walk away, and the timer is up, then it’s real­ly easy to clean them out of the data. Use JPEG files. If you’re used to using E‑Prime or MATLAB, you might be more used to using bitmap images, but JPEGs work a lot faster online. They load a lot quick­er, so I would rec­om­mend that. And test dif­fer­ent browsers. When you’re pilot­ing, get peo­ple to try out dif­fer­ent browsers. I had a hard time with Inter­net Explor­er for some rea­son. You might be com­plete­ly fine with it, but just make sure you test out dif­fer­ent browsers when you’re piloting.

Tes­sa Clark­son:
These are some of the ben­e­fits that I’ve expe­ri­enced since going online with my research. It’s obvi­ous­ly very time effec­tive. This exper­i­ment that I just told you about before, prob­a­bly would­n’t have been able to be con­duct­ed if it weren’t for GORILLA. It was­n’t a part of my PhD tra­jec­to­ry, and we only have so many years to do a PhD, as you would know, so using GORILLA made it real­ly time effec­tive. We had this ques­tion that we want­ed to ask, and so we were able to do that thanks to online research.

Tes­sa Clark­son:
Obvi­ous­ly, the tasks are real­ly eas­i­ly repli­cat­ed, but it was also real­ly good for repli­cat­ing my own design, and then being able to iso­late a vari­able and change it and then send it off again. It was just real­ly quick and easy, and I real­ly think online research is great for that. It reduces exper­i­men­tal bias a lot as well, which I like about online research. If you’re not phys­i­cal­ly there to be able to influ­ence your par­tic­i­pant in a cer­tain way, then that reduces exper­i­menter bias, I think, espe­cial­ly for stud­ies in social cog­ni­tion. And I found real­ly good com­pa­ra­ble lab to online data. So my mem­o­ry accu­ra­cy over­all did­n’t dif­fer between online and in the lab. I’ll just quick­ly show you a com­par­i­son here. For my depen­dent vari­ables, you can see this is lab data here and online data. You can see by these p val­ues here, that the dif­fer­ences in mem­o­ry accu­ra­cy was com­plete­ly neg­li­gi­ble. So what we found in the lab and what we found online is real­ly com­pa­ra­ble, and that’s great.

Tes­sa Clark­son:
Thank you. That’s as much as I have time for at the moment, but I’d love to chat with you guys more about online research. So if you do want to reach out, please send me an email. You can use this QR code here to check out the shop­ping task, if you want to have a look at the back­end of that your­self. Thank you so much to GORILLA for mak­ing this all pos­si­ble. And my won­der­ful advi­so­ry team and my real­ly great lab, thank you so much for all your support.

Speak­er 2:
Thank you very much, Tes­sa. Thank you. I should have said at the top, we’ll have time for one ques­tion, so if any­body would like to ask a ques­tion, please put that in the Q&A now. If more peo­ple have ques­tions, keep ask­ing them, because Tes­sa will be able to answer them online when she’s fin­ished speak­ing. But if you don’t get a ques­tion up quick­ly, I’m going to ask the ques­tion, and you should­n’t be forced to lis­ten to me ask­ing ques­tions all after­noon. Just very quick­ly then, I was inter­est­ed in your point about browsers and how that can inter­act with the task. What’s your feel­ing for why does that hap­pen? Why do browsers make a difference?

Tes­sa Clark­son:
I think that just, maybe… I’m not real­ly sure why we had trou­ble with Inter­net Explor­er, I just… Yeah, I’m not real­ly sure how it works, but I know that maybe visu­al­ly it could impact what the study looks like. I think it’s impor­tant for peo­ple to know that dif­fer­ent devices and dif­fer­ent browsers could make things look a lit­tle dif­fer­ent. And GORIL­LA’s real­ly good at scal­ing your visu­al stim­uli to a ratio, scale and size and every­thing like that, but it can look a lit­tle dif­fer­ent, I guess, across browsers. Maybe some browsers are a lit­tle more sen­si­tive to going into full screen mode, or there might be dif­fer­ent things to con­sid­er there, so maybe that’s why, and I think that’s pos­si­bly why it’s real­ly good to just test out all the browsers. Then if one isn’t coop­er­at­ing, you can just very eas­i­ly, in GORILLA, not let par­tic­i­pants use that par­tic­u­lar brows­er. Yeah.

Speak­er 2:
Bril­liant. Thank you.