Comparación de enfoques estadísticos para estimar la sobrevivencia de monos araña (Ateles geoffroyi)

Value of the Information & Strength of the Argument

The talk delivers strong empirical value through a unique, long-term dataset (24 years) on a vulnerable species. The argumentation is clear and well-structured, logically moving from the description of the data and inherent challenges (censoring, dispersal) to the methodological rationale for comparing frequentist and Bayesian approaches. The specific contribution lies in explicitly integrating expert elicitation to handle the ‘disappeared individuals’ uncertainty, a practical and often overlooked problem in field ecology. The presenter convincingly explains why the Kaplan-Meier estimator fails to provide comparable metrics for males and how Bayesian methods can incorporate external knowledge to fill these gaps. The presentation is grounded in the speaker’s own doctoral and postdoctoral research, adding authority to the discussion and offering a replicable framework for other studies facing similar data limitations.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is high, evidenced by the use of established statistical techniques (Kaplan-Meier, Cox models, Bayesian survival models) and a carefully designed expert questionnaire. The research benefits from collaborations with statisticians and ecologists from IIMAS-UNAM and other institutions, as mentioned by the presenter. The speaker references her own previous publications on social structure and survival, indicating an ongoing line of rigorous research. The title is perfectly matched to the content; the presentation does precisely what it promises: it compares statistical approaches for estimating survival. The structure of the seminar, with clear objectives and preliminary results, maintains scientific integrity. Methods are explained sufficiently to allow replication by other researchers in the field.

The title accurately describes the content, which systematically compares frequentist and Bayesian approaches to estimate spider monkey survival.

8/10

The presentation is based on 24 years of longitudinal field data (1997-2020) on a spider monkey group, involving 127 individuals. It uses standard frequentist (Kaplan-Meier) and Bayesian parametric survival models, and integrates expert elicitation to handle censored individuals. The speaker is a postdoctoral researcher with relevant publications. Results are preliminary but methodologically sound.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction and logistical announcements for the departmental seminar series.
• 04:00Formal introduction of speaker Cristina Jasso del Toro and her academic background.
• 08:00Definition of survival analysis and its key components (time-to-event, censoring).
• 25:00Comparison of frequentist and Bayesian statistical interpretations and methodologies.
• 30:00Discussion of specific challenges in estimating survival for wild animals (small samples, rare deaths).
• 35:00Presentation of the study species (spider monkeys) and the 24-year dataset (127 individuals, 49 males, 78 females).
• 45:00Presentation of Kaplan-Meier survival curves, showing limitations for male survival estimation.
• 50:00Explanation of the expert elicitation process using a 12-scenario questionnaire for disappeared individuals.
• 1:00:00Integration of expert opinion as prior information into Bayesian survival models.
• 1:10:00Presentation of preliminary Bayesian model results and comparison with frequentist outputs.