A big thanks to my supervisors @gudruneisele.bsky.social @ginettelafit.bsky.social @oliviajkirtley.bsky.social @inezgermeys.bsky.social for all their support along the way :)

On a more personal note, I am grateful to have had the opportunity to set up my own study during my PhD. Although there were some bumps in the road, I have learned a lot—and it has been extremely rewarding to analyze data that I collected myself!

More broadly, we argue that ensuring that ESM studies generate as much high-quality data as possible without excessive burden to participants is essential, and incentives represent a key lever for researchers to optimize their ESM protocols.

In the absence of any significant effects on data quality and participant experiences, we discuss other factors—at the beep (e.g., social context), person (e.g., clinical symptoms), and study (e.g., gamification) level—that might influence these outcomes in ESM research.

We found that paying participants per beep increases compliance as compared to bulk payment with or without personalized feedback—and without harming data quality. Interestingly, we found no significant effects for personalized feedback on data quantity, data quality, or participant experiences.

Only if you help me code it👀

So my intuition is that (the strength of) the relationship between the compliance and the payment matters! But curious to hear other ideas as well :)

E.g., payment structures with only one/few critical compliance threshold(s) would be in the same category as payment per beep). In the SHARE study, we chose to compare a fixed payment (not conditional on compliance) to an extreme form of incremental payment, i.e., payment per beep (all beeps count).

Glad you were able to access the poster!

I think that the lack of effect in previous meta-analyses could be partially explained by the fact that they typically group together all incremental payment structures regardless of how ”strongly” the payment is tied to the compliance rate.

Stay tuned for the preprint!

5/ Want to dive deeper? Read the full preprint here: osf.io/preprints/ps... 🔍

With @gudruneisele.bsky.social @lisapeeters.bsky.social @oliviajkirtley.bsky.social @ginettelafit.bsky.social Inez Myin-Germeys

@ccp-kuleuven.bsky.social

4/ Our findings emphasize that careless responding in ESM isn’t just noise—it follows systematic patterns over time. To improve data quality, researchers should account for these temporal trends.

3/ Interestingly, these indicators were only weakly related—meaning they capture different types of carelessness. This highlights why relying on a single measure might not be enough—using multiple indicators (or other methods) provides a more complete picture.

2/ BUT: Not all indicators performed equally well. Response time, within-beep standard deviation, and the inconsistency index captured carelessness, but occasion-person correlation was not well-suited for an ESM context.

1/ Our results indicate that careless responding in ESM is not stable over time but increases across days and fluctuates within days.