The following methodology was employed for data analysis and used to obtain a functional universe for sample selection purposes. The techniques in this analysis include, but were not limited to, exploratory analysis, comparative analysis, and descriptive analysis. We describe the results of the analysis at different phases, reasons for the analysis, and the sample methodology.
In the database of 347,716 opted-out records, there were a total of 71,179 unique VNS Registrant IDs. In order to understand the reasons for opt out, we examined the data and we found four different types of data values and blanks in the field “OPT-OUT REASON” to include invalid address, user choice, contact choice, no longer a victim, and blank fields.
We excluded the 52,029 VNS IDs that were associated with “Invalid Address” and 6,025 records that had no information in both the city and state columns of the database. The final universe of opt-out victims contained 13,125 records (opt-out victims) and only included records that had enough information for sample selection purposes.
In order to obtain a better response rate and useful responses to the questions, we used a stratified sampling plan. The universe of 13,125 unique Registrant IDs was stratified considering the combination of the number of notifications and the values in the “OPT‑OUT REASON” field for a total of 489 sample Registrant IDs selected.
Opt-out Survey Response Analysis
There were 58 respondents who responded to the opt-out survey. There were 14 respondents who did not respond to all or most of the questions in the survey. The responses from these 14 respondents were excluded from further analysis. A frequency table of the remaining 44 responses was saved and used for logical tests analyses.