Analyse Umpire Feedback Bias
Budget: £20 – £250 GBP
I oversee a cricket league that assigns 50 umpires to as many as sixteen matches each season. After every game both team captains complete an online survey, giving each on-field umpire a 1–5 rating and the option to add open comments. We also log the match result and note any disciplinary reports that arise.
I need a statistician who can tell me—with evidence—whether these ratings are systematically influenced by a team’s result (win / loss) or its disciplinary record. In short, do victorious captains reward umpires with higher scores and do sanctioned teams punish them with lower ones? The conclusions will directly affect how we promote or retain our officials, so the analysis must be rigorous and clearly communicated.
You will receive this season's data set containing game IDs, the two teams, the two umpires, win/loss flag, disciplinary flag, numeric ratings and any text comments. Using the toolset you prefer—R, Python (pandas, statsmodels), SPSS, or similar—please:
• quantify the presence (or absence) of rating bias linked to match outcome and disciplinary incidents
• control for repeated measures (same umpire, same team) so results are robust
• summarise findings in a concise report with visualisations and plain-language conclusions
• recommend any adjustments to our feedback process that would reduce bias going forward
If additional data preparation is needed, let me know early so I can supply it.
I need a statistician who can tell me—with evidence—whether these ratings are systematically influenced by a team’s result (win / loss) or its disciplinary record. In short, do victorious captains reward umpires with higher scores and do sanctioned teams punish them with lower ones? The conclusions will directly affect how we promote or retain our officials, so the analysis must be rigorous and clearly communicated.
You will receive this season's data set containing game IDs, the two teams, the two umpires, win/loss flag, disciplinary flag, numeric ratings and any text comments. Using the toolset you prefer—R, Python (pandas, statsmodels), SPSS, or similar—please:
• quantify the presence (or absence) of rating bias linked to match outcome and disciplinary incidents
• control for repeated measures (same umpire, same team) so results are robust
• summarise findings in a concise report with visualisations and plain-language conclusions
• recommend any adjustments to our feedback process that would reduce bias going forward
If additional data preparation is needed, let me know early so I can supply it.