🧌This "bestiary" can be a vital resource for researchers, educators, and reviewers to recognize, understand, and mitigate QRPs, ultimately raising the standard of psychological research. Read the full article here: journals.sagepub.com/doi/10.1177/...

Aleksandrina Skvortsova, Samuel Alarie, Katherine Graves, Lukas K. Sotola, @davidmoreau.bsky.social , @evarubinova.bsky.social

🙏Huge thanks to all co‐authors: @amathia-deluxe.bsky.social , @mamsmandrill.bsky.social, @lukaswallrich.bsky.social, Kathleen Schmidt, Tal Waltzer, Jason W. Payne, Biljana Gjoneska, @yashvin.bsky.social, @andrewang.bsky.social, Daniel Scharfenberg, Gabriella Tyson, @ufangyang.bsky.social, ...

😮‍💨But there's hope! Our study highlights key preventive measures to avoid QRPs: transparent reporting, preregistration, blind data analysis, following best practices, and performing robustness checks. These foster research integrity!

⚡Why should we care? QRPs pose a substantial threat to scientific conduct and quality. They can lead to biased statistical error rates, reduced replicability & reproducibility, biased effect size estimates, and inflated credibility, undermining research trustworthiness.

🕵️How detectable are QRPs? Our study identified 18 QRPs detectable directly from a publication📰, while 19 more need supplementary info like preregistration or open data. We found 97 unique clues🔍! This info can serve as a checklist for human reviewers & prompts for AI risk assessment.

☂️To bring clarity, we introduced "QRP umbrella terms" that group conceptually related QRPs. For example, #Phacking 🪓consists of 13 specific QRPs, and #CherryPicking 🍒covers 7. This disaggregation helps pinpoint specific behaviors.

🧑‍🔬We collected and categorized 40 distinct QRPs for quantitative research in psychology, spanning all stages of the scientific enterprise from planning to publication. Many of these QRPs were previously unlisted, despite being recognized (e.g., visualizing data in a misleading way 📊).

❓So, what are QRPs? Our community consensus study agreed on this definition: "Ways of producing, maintaining, sharing, analyzing, or interpreting data that are likely to produce misleading conclusions, typically in the interest of the researcher". Crucially, they don't include fraud or random error.