I consider here registration drives that are distinct from the legislatively encouraged registration of 18 year old students, that is, not on 2019-12-09, 2020-01-03, 2020-01-08. The process of identifying indicators of these registration drives consists of the following steps, which are carried out separately for each county. The day-to-day variation in registrations is considerable, making the use of the usual time series methods cumbersome. Consequently, I rely on subjective judgment. The process is:
All of the f(GT) yield very nearly the same set of days with earlyDate TRUE.
This is biased towards single-day events. A drive spread out over several days may not be recognized as such. This is complicated by some of the drives appearing to have taken place over a few consecutive or nearby days. How should these be addressed? There seem to be few choices, including treating the days separately or combining them in some consistent fashion. I choose not to treat them separately because of the clutter that would cause, and the smaller numbers of young people who would be involved each of those days. Another complication are “joint” drives, those that involve older persons as well as younger. I consider these to constitute a separate category and I will exclude them from this present analysis. There is no hard and fast rule for distinguishing joint drives, but roughly speaking, I will consider them so if in the count for a day the number of older persons is more than half the number of young persons.
This selection process finds putative drives in 10 of the 100 North Carolina counties. The counties and days are shown below. Evidently, Mecklenburg was very active in March of 2018 and 2019.
If the registration drive process has effectiveness compared to not carrying it out, then there must be some increase in the number of young people who registered and subsequently voted. I will compare the voting rates of the registrants with their age peers who register in 2018, 2019 and 2020, but did not pre-register nor prticipated in the one-stop October 15-31, 2020, window.
Mecklenburg and Wake far outweigh the other counties in regard to population, those two counties being the largest in North Carolina. They constitute about twenty percent of the entire state, and twice as many as the other counties on this plot, as made available by the NC Office of State Budget and Management.
This plot has categories that overlap in order to display and contrast (“slice and dice”) the ways young people can register. “Other” are the young registrants who did not pre-register, that is, who did not participate in the highschool age early registration activity enabled by the Legislature.
It is evident that the highschool-age pre-registration consistently produces higher percentages of voting in these counties than do alternative registration methods. The brief drives that are the subject of this report appear to be associated with increased voting rates in most of these counties. Lenoir and Surry are the counties with the smallest populations
It is helpful to have a quantitative measure of the effectiveness of a registration campaign. A measure that is used in many disciplines is the efficacy rate. This is defined in the following way by supposing that a treatment (early registration) is being applied to part of a population, while the remaining part does not participate in early registration.
E = [(t/T) - (u/U)]/(u/U)
This is the difference in the proportion of treated registrants who vote and that of the untreated registrants who vote, divided by the latter quantity. This can also be written as
E = (t/T)/(u/U)) - 1
A positive value of E indicates some degree of success, while a negative value indicates failure.
The number of treated people who vote in excess of those who would have voted if there had been no treatment is computed as:
N = TE(u/U)
E = (700/1200)/(500/1500) - 1 = (0.583/0.333) - 1 = 1.75 - 1 = 0.75
N = 1200 * 0.75 * (500/1500) = 300
So if the 1200 treated people had voted at the rate of those untreated (0.33), only 400 (1200 - 300) would have voted.
However, this is a simplistic model and ignores complexity and possibly confounding factors. For instance, the early registrants may be more enthusiastic and might have self-selected and consequently voted at a higher rate regardless of early registration.
This report was run on 2021-02-14.