# My First PhD Study: Contactless

There is a reason I was so keen on writing a series about payment methods and the pain of paying. This is what my own PhD research is about. I am looking at the effect of contactless on expenditure and expenditure recall. In this article I will quickly summarise how my first study has gone.
__Procedure __

When it comes to recall, one of the easiest things to do is just ask. As such, we did. Customers of the on-campus grocery store were approached by a research assistant. They were asked to fill in a short survey. The participant was given the survey, after handing the receipt to the assistant. Participants were asked to indicate their age, gender and whether the participant was a student.

To test for the relationship between payment method and expenditure, participants were asked which method of payment they had used. The survey was matched with the receipt, which showed the actual expenditure of the individual and was used to verify the method of payment indicated on the survey.

To test for the relationship between payment method and expenditure recall, participants were asked to estimate their expense in the store. The completed survey was then exchanged for a choice of snack (chocolate/fruit).

*Measures*
This study focussed on the method of payment and the recall error (expenditure – estimate of expenditure). Receipts were matched to the survey, to later determine the deviation between money spent (receipt) and perceived expenditure (survey).

*Participants *
A total of 3050 individuals (56.5 % female, mean age = 21.4, age range 12 - 83) of which 94% were students, completed the survey. Of those, 6 were incomplete, 15 had missing receipts and 7 were unusable due to answers such as (“I did not pay, I robbed the store”). After applying some other exclusion criteria such as an expenditure over £30, (contactless payments are not able to exceed this limit), the analysis was conducted on the remaining 2,923.

__Results__
Without getting too technical and without posting 67 different graphs I will quickly outline the results we have found.

*Expenditure*
Looking at expenditures we found that people spend less using contactless debit cards compared to pin-verified debit cards. The same was true for credit cards. Comparing contactless debit cards to cash, people spend more using contactless debit cards. Cash is still the best way to spend least.

*Expenditure Recall*
There were three different measures for expenditure recall accuracy: expenditure error, absolute expenditure error and the probability of correct expenditure recall.

Looking at expenditure recall error we found a smaller recall error with people using contactless debit cards compared to pin-verified debit cards. The same was true for credit cards. Comparing contactless debit cards to cash, there was a slightly larger recall error using cash, which is very surprising. Unfortunately, the differences in recall error were so small they are statistically non-significant. This means the real difference might as well be zero.

The findings for expenditure recall error replicated when measuring absolute expenditure recall error. An interesting finding to mention is that the recall error for contactless was negative, meaning that people overestimated their expenditure rather than underestimating it.

Looking at the probability of correct expenditure recall we find that there is no difference in the probability of correct recall for both pin-verified and contactless debit cards. For credit cards, people had a higher probability of correct recall using contactless. Comparing cash to contactless debit cards, people had a lower probability of correct recall using contactless debit cards.

*Covariation*

As exciting as the results above sound, we need to be careful intepreting them. Within the collected data, we find a strong indication of a *one-item-effect*. Meaning, that when purchasing one item, recall is incredibly accurate. It has an accuracy rate of 84.2%. Whereas the accuracy rate drops to 51.1% for two or more item purchases.

This one-item-effect also shows up in regressions. Regressions for all forms of recall indicate a significant role of the number of items purchased, yet none of the payment methods are significant in impacting either expenditure, or any form of expenditure recall. The only variable that does seem to have an impact is that of the Number of Items purchased.

Now this is not necessarily a strange result to find. It is seemingly in line with theories on short-term memory, which postulate that the more complex something is to remember, the less accurately we will remember it. If we have to only remember the price of one item, that is quite easy, but trying to remember the price of four different items and their cumulative expense, increases complexity massively. As such, the result fits the known literature.

__Caution and Conclusion__
There is a lot of information in this article about the set-up and results of the study. The take-away message is the following: when comparing pin-verified to contactless cards, in terms of expenditure and expenditure recall, contactless outperforms pin-verification. Contactless methods show lower expenditures, lower recall errors and higher probabilities of correct recall. When comparing contactless methods to cash, cash outperforms contactless. Cash shows lower expenditures, lower recall errors and higher probabilities of correct recall. This is in line with the pain of paying. However, most of these effects are negated by the number of items purchased, which provides a stronger explanation for the decrease in recall accuracy.

There is caution in interpreting these results. What I have described above is a simplification of the statistics run on the data. However, the biggest caution is that of the sample the results are derived from. The main group tested on was students, who mainly popped in to buy lunch, meaning that on average they bought about three items. It is not that difficult to remember the prices of three items, especially not for individuals in higher education. As a result, we are cautious in interpreting these findings, and urge for replications of this study with participants who are not students and who buy 10+ items, to strengthen the results.

I hope this summary of my study has given you a glimpse into what it is a PhD student does. How we approach topics and go about researching them. Let me know if anything is unclear, you’d like to know more or you have a suggestion for future research!