The effect of simulated censored data on estimates of heritability of longevity in the Thoroughbred racing industry.

The research investigates the influence of censored data on heritability estimates of longevity in Thoroughbred racing horses. Elements affecting longevity and variances in heritability scores with different degrees of data censoring were studied using a simulated data set modelled on the Irish Thoroughbred industry.

Objective of the Study

• The aim of the study was to analyze the impact of censored data on the estimates of heritability of longevity in Thoroughbred horses. The longevity was defined as the length of a productive racing life of a horse, which tends to be influenced by numerous factors.

Methodology

• The research utilized a simulated data set that was modeled on the basis of the Irish Thoroughbred industry.
• Researchers created several scenarios for their analysis, each representing different levels of censoring of performance data.
• For every created scenario, the heritability of longevity was estimated and then compared with the estimated heritability for the complete data set, which was 0.120.

Findings

• Upon analyzing the data, it was found that the estimates of heritability were lower than expected when data was censored or omitted especially from poor-performing animals.
• Specifically, the heritability estimates were 0.107, 0.106, and 0.082 with 10%, 20%, and 25% of data, respectively, being censored.
• This suggests that censoring data introduces bias to the estimation of heritability of longevity.

Conclusion

• The study concluded that for completeness in reporting and to reduce bias in the estimation of heritability of longevity, it is important to include a complete data set without censorship, especially data from poor performing horses.