MODEL RELEASE NOTES


Model: KYDP 2024 General Turnout

Type: Vote History

Geography: Kentucky

Date: March 13, 2023 (v1)


Background and Goals


The Kentucky Democratic Party (KDP) 2024 General Turnout Model seeks to predict the likelihood of a registered voter turning out in the 2024 General Election. To train and inform this model, a random sample of 2020 general voters was used to identify commonalities and predict what other voters in Kentucky will do. 


Model Description + Process Overview


A logistic regression model was created to target and identify individuals who are likely to turnout and vote in Kentucky’s 2024 general election. The model was constructed using voters who were eligible to vote in the 2020 Kentucky General Election and assigning values based upon whether a voter cast a ballot in 2020 (“0” for non-voters and “1” for voters) as training data.  


  1. Targets: Voters who had voted in the 2020 General Election

  2. Non-Targets: Voters who did not vote in the 2020 General Election.


Model scores are expressed on a 0-100 scale that represents the probability that a person will vote in the 2024 General. The model was used to score the Kentucky voter file (March 1, 2023 version) and will be used to score new voters after each update. 


The model was built using rules-based classifiers on a variety of consumer, political, and demographic data. Each group was selected through a random sample of 10,000 voters (6030 targets, 2970 non-targets) to represent 60.3% turnout.  


The model building process determined the most appropriate variables in identifying the differences between target voters and non-target voters. The final score was generated using a linear regression. To validate the model, a randomly selected group of records was held out from the model building process. These holdout records were then scored and analyzed for accuracy. 

Model Evaluation, Validation, and Weights


Area Under Curve (AUC): 0.836

An AUC of .796 indicates that the model’s ranking of primary turnout likelihood aligned with actual primary voters in our testing samples 79.6% of the time. 


Precision: 0.8333

The frequency with which a model was correct when predicting the positive class.


Recall: 0.7301

Out of all the possible positive labels, how many did the model correctly identify?


Accuracy: 0.7506

Fraction of predictions the model got right




The model was validated by scoring the remaining records of General 2020 voters.  The holdout records were then ranked by model score, separated into ten score ranges, and evaluated. 


When looking at the score ranges, the model correctly identified 97.97% of general voters (693,186/707,539) in the 90+ range and 93.53% (493,199/527,331) in the 80-89.99 range.


In the bottom of the score ranges, 26.67% of 2020 General voters (4/15) were placed in the 0-9.99 range and 30.52% (13,619/44,622) were placed in the 10-19.99 range. 


Score Distribution


Of the million voters in Kentucky, the model predicts approximately 1,938,678 Kentucky voters are likely to vote in the 2024 General. If using a 50+ score as the basis for most likely turnout, that total is 1,707,712 voters.  


The following charts show the distribution of model scores for all voting-age persons in the Commonwealth of Kentucky. The charts show score ranges from 0-100 with the bottom chart showing further score breakdowns, with higher scores indicating a higher likelihood of voting in the 2024 General Election. 


Versioning Notes


Version #

Version Date

Notes

Turnout Expectation (Avg / 50+)

1

3/13/2023

Initial build of the model.

1,938,678/1,707,712


Usage Notes


To make the most out of this model, users should target voters with higher scores and work down when targeting voters. Please note that as the score range is lowered, the campaign will be targeting voters who are less likely to vote in the 2024 General .