Duke University

Evidence-adjusted career presence. The 0.84× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 0.16×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.189. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 0.69× to 0.91×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.

Evidence-adjusted career presence. The 1.76× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 1.83×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.909. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 1.69× to 1.80×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.

Evidence-adjusted career presence. The 1.72× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 3.00×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.362. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 1.44× to 2.17×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.

Evidence-adjusted career presence. The 1.01× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 1.05×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.249. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 1.01× to 1.02×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.

Evidence-adjusted career presence. The 1.01× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 1.07×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.124. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 1.00× to 1.02×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.

Evidence-adjusted career presence. The 0.81× headline is a modelled rating, not an observed employment rate.

Raw observed/expected ratio: 0.00×. The same continuous evidence adjustment applies to every school: 1 + evidence weight × (raw ratio − 1). Evidence weight: 0.187. Small expected cohorts and disagreement between cohort definitions reduce this weight. Traditional reputation does not change a direct rating.

With alternative stability settings, this rating ranges from 0.64× to 0.90×. This is a model-sensitivity check, not a confidence interval or range of hiring odds.

Position ranges group the model's ordering for readability; they are not statistical confidence intervals. Traditional reputation is a proxy, not evidence of placement in this industry or programme availability.