This Appendix contains Table C.1, a list of item response rates for questions from the N-SSATS 2010 questionnaire. Item non-response was minimized through careful editing and extensive follow-up.
The item response rate for the 2010 N-SSATS averaged 98.6 percent across 192 separate response categories.
Item non-response was 10 percent or more for only 5 of 192 separate response categories.
Imputation was conducted for 3 instances of client values and 257 cases of admissions values for facilities that were missing counts for one or more service types (i.e., hospital inpatient, residential [non-hospital], or outpatient treatment, and their subcategories). An imputed value for a service type was assigned if a facility reported that it provided the service but had not provided a client or admissions count for the service.
Imputation was also conducted for client count values identified as outliers. Outliers were defined as cases where the number of clients reported exceeded the 75th percentile of the distribution of client counts for all cases of one of the main service types. The 2010 value was compared with the previous year's value. If the difference exceeded 1.8 times the 2009 value in either direction, the value was imputed.
When available, client values and admissions values from up to five previous surveys were used to impute the missing counts. If historical data were not available, the average client value, stratified by State and facility operation, was used to impute the missing client counts. If a facility were unique in its State and facility operation category, values were imputed using average values for the State only. Missing client counts were imputed for each type of service (i.e., hospital inpatient detoxification, hospital inpatient treatment, residential [non-hospital] detoxification, etc.) and summed to the larger service type totals (total hospital inpatient clients, total residential [non-hospital] clients, and total outpatient clients), and finally to total clients.
