Back

Improving Statistical Rigor in Animal Aging Research by Addressing Clustering and Nesting Effects: Illustration with the National Institute on Aging's Intervention Testing Program Data

Parker, E. S.; Golzarri-Arroyo, L.; Dickinson, S.; Henschel, B.; Becerra-Garcia, L.-E.; Mokalla, T. R.; Robertson, O. C.; Thapa, D. K.; Vorland, C. J.; Allison, D. B.

2025-03-17 systems biology
10.1101/2025.03.14.642436 bioRxiv
Show abstract

Clustering effects, such as those introduced by housing animals in shared cages, are often overlooked in preclinical lifespan studies, despite their potential to distort variance estimates and inflate Type I error rates, leading to misleading conclusions. This methodological oversight reduces statistical rigor and may undermine the reliability of findings. To address this gap, the current study examines the impact of accounting for clustering and nesting effects on lifespan analyses by comparing the results of statistical models which both account for and ignore these effects. Using 2019 data from the Interventions Testing Program (ITP), a large-scale initiative evaluating the effects of compounds on lifespan in UM-HET3 mice as a case study, we illustrate how different modeling approaches influence statistical estimates and conclusions. Clustering and nesting effects were addressed using linear mixed effects, and Cox frailty models, both of which explicitly account for cage-level dependencies and different levels of data nesting. Comparisons were made between unadjusted lifespan analyses and those incorporating clustering and nesting adjustments. The results of this case study indicate that properly adjusting for clustering and nesting effects can change the conclusions drawn from statistical significance tests as compared to unadjusted model approaches, and so it remains best practice to properly account for clustering and nesting to reduce the potential for inflated Type I error rates. These findings highlight the importance of accounting for clustering and nesting in preclinical research to ensure valid and robust statistical inference. By demonstrating the practical application of clustering adjustments, this work underscores the broader implications for improving reproducibility and rigor in lifespan studies and other experimental designs.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

1
GeroScience
97 papers in training set
Top 0.1%
33.5%
2
Experimental Gerontology
11 papers in training set
Top 0.1%
17.8%
50% of probability mass above
3
PLOS ONE
4510 papers in training set
Top 31%
4.9%
4
Aging Cell
144 papers in training set
Top 1%
4.0%
5
Frontiers in Aging Neuroscience
67 papers in training set
Top 0.8%
4.0%
6
The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences
22 papers in training set
Top 0.1%
3.6%
7
Neurobiology of Aging
95 papers in training set
Top 0.9%
3.1%
8
The Journals of Gerontology: Series A
25 papers in training set
Top 0.4%
2.5%
9
Aging
69 papers in training set
Top 1%
2.1%
10
Scientific Reports
3102 papers in training set
Top 53%
1.9%
11
eLife
5422 papers in training set
Top 41%
1.7%
12
npj Aging
15 papers in training set
Top 0.6%
1.4%
13
Brain Research
35 papers in training set
Top 1%
1.4%
14
Frontiers in Aging
10 papers in training set
Top 0.3%
1.1%
15
International Journal of Molecular Sciences
453 papers in training set
Top 14%
0.8%
16
Cell Reports
1338 papers in training set
Top 31%
0.8%
17
International Journal of Psychophysiology
14 papers in training set
Top 0.1%
0.8%
18
Frontiers in Medicine
113 papers in training set
Top 7%
0.8%
19
Alzheimer's Research & Therapy
52 papers in training set
Top 2%
0.8%
20
Bioinformatics Advances
184 papers in training set
Top 5%
0.7%
21
PLOS Computational Biology
1633 papers in training set
Top 27%
0.7%
22
BMC Bioinformatics
383 papers in training set
Top 8%
0.5%
23
Biology Methods and Protocols
53 papers in training set
Top 4%
0.5%