Statistics at the Zoo

Review: Statistics at the Zoo
Student’s Name
Affiliation
Course
Due Date
Review: Statistics at the Zoo
The article “Statistics at the Zoo” provides information on the novel ways that Chester Zoo has used statistical data to protect endangered species and enhance animal welfare (Holmes et al., 2020). Holmes et al. (2020) report that statistical studies have been used to promote animal welfare, understand the reproduction patterns of Asian elephants, save an endangered parrot, and measure the impact of zoo visits. While the statistical studies have informed the adoption of evidence-based courses of action at Chester zoo, a review of the studies indicates that collider bias may have compromised the reliability of such studies. Collider bias occurs when the process of selecting a subset of data for analysis influences the relationship between the variables being studied, leading to erroneous conclusions being drawn from the data (Herbert et al., 2020; Homberg & Andersen, 2022). Despite the utility of statistical studies highlighted by Holmes et al. (2020), collider bias in such studies could undermine the zoo's understanding of the effectiveness of measures informed by the data.
In a highlighted statistical study where data was used to promote the welfare of a female lion at Chester zoo, the use of a longitudinal study improves the reliability of the findings. A longitudinal study allows researchers to observe changes in the subjects over time, which can provide insights into the development and progression of a particular phenomenon (Blazar & Kraft, 2019). To this end, the researchers collected data on the female lion and concludes that feeding carcasses to the animal reduced pacing in the female lion (Holmes et al., 2020). While the study was informative, some factors point to the likelihood of collider bias in the described study. The sampling of data in the study was not random, and the researchers do not consider the likelihood that other factors may have contributed to the reduction in pacing behaviors among lions. Herbert et al. (2020) note that one can identify a study affected by collider bias by examining the extent to which the researcher relies on random sampling to identify participants in the study. In the study, the researcher has only one participant and thus raising questions about the generalizability of the study. To improve the study, the researcher should consider studying the entire population of lions at the zoo and randomly assigning them to two groups. One group should be the control group and the other should be the intervention group. In such a way, the Chester zoo will have more data about the impact of changing diet.
In the same vein, the statistical study measuring the educational impact of visiting a zoo had a high likelihood of collider bias due to issues with the methodology. While collecting the data using a pre-post survey, the researcher did not have a control group making it hard for one to conclude whether visiting the zoo was the only variable that could explain the increase in knowledge among visitors (Ro...