Chapter 3. Methodologies and Data Collection

In this chapter

3.1 Overview
3.2 Child Support Functions and Processes
3.3 Time Estimates
3.4 Estimates from the Focus Groups (Professional Estimates)
3.5 Time Logs
3.6 Time Estimates from the Focus Groups and Time Logs
3.7 Alternative Methodology

3.1. Overview

A prerequisite for developing staffing standards is a complete understanding of the functions, processes and tasks that comprise operations in a child support office. That understanding is particularly important in Colorado where technician responsibilities vary significantly across counties based on differences in organizational structures and county demographics. Thus, the first section of this chapter provides a detailed list of functions, process and tasks. It also discusses some of the factors that lead to differences among offices in how these functions, processes and tasks are prioritized and implemented.

The second section focuses on measuring the amount of time spent on child support activities, both our approach to measurement and the time estimates that resulted from that approach. The study relies on two methodologies:

Professional Estimates, which are estimates provided by child support staff, supervisors or experts; or extracted from historical records.
Time logs. Workers record what they do and how long it takes to do it.

The final section explores an alternative methodology for determining staffing standards. Specifically, it uses available statistical data to examine the relationship between Federal and State performance indicators and staffing levels. County-level and state-level data are used to define the strength of the relationship between staffing levels and selected program performance measures (e.g., percent of cases under order). We translate that relationship into estimates of staffing levels Colorado would need to achieve its performance goals.

3.2. Child Support Functions and Processes

PSI staff drafted a list of child support functions, processes and tasks associated with IV-D operations based on:

Information from the Virginia Child Support Staffing Study
Information from a prior study we conducted of child support operations in Arapahoe County, Colorado, and
Information about case processing from our own privatized child support operations (either full or partial service) in 12 states.

We defined work into functions, processes and tasks based on the following definitions:

Function: Broad but definable area of child support enforcement. The specific processes within the function are unique enough to distinguish it from other broad functional areas. For example, Intake is a distinct function in child support.
Process: A process is any series of activities (or tasks) that when taken together, transform a series of inputs into an output producing a result of value to a customer. Within the Intake function, for example, opening a new non-IV-A case is a distinct process.
Task: An activity that when linked with other related activities makes up a process. For example, the distinct tasks associated with the opening of a new non-IV-A case might be (1) processing the IV-D application, (2) updating the system, and (3) gathering any relevant information from the custodial parent.

We sent the draft list of functions, process and tasks to the State project manager, MIG members and IV-D administrators in five counties—Weld County, Conejos County, Kit Carson County, La Plata County, and Routt County—for review. Revisions were made based on their comments. Exhibit 6 displays the finalized list of functions and processes for six functions: intake, locate, establishment, enforcement, fiscal and customer service. A detailed list of the tasks performed within each process is provided in Appendix A.

Table 3.1. Exhibit 6: Child Support Functions and Processes

Processes for Intake

Case Opening—non IV-A

Case Opening—IV-A

Case Opening – Foster Care

Non-cooperation

Good Cause

Case Opening – Interstate (Responding)

Case Opening – Medicaid

Processes for Locate

Locate NCP Address

Locate NCP Employer

Locate NCP Assets

Processes for Establishment

Establish Support Order

Establish Paternity Order

Responding Interstate Paternity and/or Support Establishment

Initiating Interstate Paternity and/or Support Establishment

Establish Paternity and/or Support – Judicial

Responding Interstate Establishment – Judicial

Establishment Case Management

Non-cooperation

Case Closure

Genetic Testing

Prepare for Hearing

Processes for Enforcement

Income Withholding

Review and Adjustment of Orders

Enforcement Case Management

Case Closure

Medical Support

Non-cooperation

Administrative Enforcement (e.g., license suspension, credit bureau reporting, etc…)

Prepare for Hearing

Contempt

Initiating Interstate Enforcement

Processes for Fiscal

Ledger Calculation

Allocation, Distribution, Disbursement

Processes for Customer Service

Phone Calls

Mail

Walk-ins

Generally, the people who reviewed the document agreed on the functions and processes, but differed somewhat in the number and content of the individual tasks within each process. To reconcile these inconsistencies, the list of tasks in Appendix A are more rather than less detailed. This detail captures the county differences reviewers reported about individual tasks. Some of the few, but notable differences in tasks reviewers noted included the following.

The more specialized a County's staff, the more developed and detailed the tasks within the processes tended to be.
While the tasks were essentially the same in all counties for order establishment, a county's use of a judicial as opposed to an administrative process was derived from local office practice and judicial preferences.
In a variety of enforcement areas, the smaller the county’s size, the greater the preference for negotiating with the obligor.

3.3. Time Estimates

As discussed earlier, we developed time estimates using two different methodologies: (1) professional time estimates; and (2) time logs. To obtain the professional time estimates, PSI staff conducted focus groups in six County IV-D offices: two large offices, two mid-sized offices and two small offices. We asked staff in three different offices—again, one large office; one mid-sized office and one small office—to complete self-administered time logs. The State project manager asked counties to volunteer for participation in the study. In turn, PSI staff contacted the counties to finalize the arrangements for data collection. Selected child support statistics of the counties participating in the study are provided below in Exhibit 7. Selected demographic characteristics for these counties are displayed in Appendix B.

Table 3.2. Exhibit 7: Selected IV-D Statistics of Participating Counties^[a]

Counties Participating in Focus Groups	Total Number of Cases (Dec 1999)	Number of FTES^[b] (2000)	Percent of Cases Under Order ( Dec. 1999)	Percent of Current Support Paid ( Dec. 1999)
Adams County	13,576	54.0	76%	47%
Denver County	31,360	141.3	77%	40%
Delta County	1,687	8.1	69%	58%
Eagle County	734	5.0	61%	48%
Grand County	265	1.5	77%	63%
Garfield County	1,826	8.3	78%	52%
Fremont County	2,371	9.0	81%	49%
Larimer County	6,963	28.0	75%	51%
Prowers County	1,359	3.9	68%	56%
State Total	160,057	612.1	72%	47%
^[b]Excludes attorneys. Information was calculated by the State office using County Annual Plans.

3.4. Estimates from the Focus Groups (Professional Estimates)

To obtain professional estimates at each of the six participating counties, PSI staff set up six focus groups with enforcement technicians and other staff at each of the County offices. Each focus group centered on one of the functions (i.e., intake, locate, establishment, enforcement, fiscal and customer service). The County administrator was asked to select appropriate staff for each focus group. PSI staff provided them with the following selection guidelines.

Up to 10 technicians per group (optimal number 5-8 but obviously less in the smaller counties).
The staff person should have at least 6 months experience in their current position.
The staff person’s primary work is in one or more of the six functions.

PSI staff led the focus groups through a roundtable discussion to reach consensus on the average time it takes to complete each process within a function. The protocol for the focus groups is displayed below in Exhibit 8. A detailed description of the focus group protocol is also provided in Appendix C. Generally, each focus group lasted no longer than two hours. All six focus groups took one to two days per County office. Focus groups were conducted in April and May 2000.

Table 3.3. Exhibit 8: Focus Group Protocol

1. Introduction (e.g. background and goals of project and focus group) and instructions.

2. Distribution of a set of 3 x 5” cards to each participant. Each card displays one process and the tasks associated with that process. Participants are asked to independently estimate the amount of time they need to complete the process from the first task to the last task on the card.

3. A round robin approach was used to capture the time estimates from each participant. Ranges and midpoints are calculated. Participants are asked whether the midpoint accurately reflects an average case. If not, more detail and discussion ensue until participants reach consensus on an average time.

4. Replicate the approach for the remainder of the processes in that function.

3.4.1. Nuances in Administering the Focus Groups

PSI staff had the following observations based on their facilitation of the focus groups.

Generally, consensus on the average amount of time required to complete each process was reached rather quickly.
Defining an "average intake" was difficult for most of the participants. As a result, PSI staff asked participants to estimate time for two separate intakes: (1) a "clean case" and (2) a "problem case." A clean case was defined as one where things went smoothly. A problem case was defined as a complex case that had difficulties, such as multiple identification numbers or an existing order in another jurisdiction.
Participants were also asked what proportion of their intakes was "clean' and what proportion was "problematic." This was used to construct a weighed time average.
Participants had similar difficulty defining an average case when estimating the amount of time to establish or modify an order. The difficulty was resolved when we discussed three types of establishment methods separately: (1) consent; (2) default; and (3) contested. Participants were asked the frequency they used each type to establish and modify orders. This information was used to form a weighed time average.
In some situations, it was necessary to walk through all of the tasks within a process and estimate the time for each task. The processes that required this level of detail varied among the counties. Generally, this occurred when the process was performed by only a few staff within a County or performed infrequently (e.g., good cause exemption in a rural County).
The “customer service” function was not always readily separable from the other functions, particularly in counties that do not have staff dedicated to this function. For example, telephone conversations and walk-ins with noncustodial and custodial parents may be viewed as part of the enforcement or other function and not customer service.
Some of the divisions of labor resulting from the County's organizational structure did not match the six functions and their respective processes. For example, in Adams County, the staff who conduct modifications are in the establishment unit. Hence, it was not appropriate to include this staff person in the establishment or enforcement focus groups because they did not perform any of the other processes in that function. In situations where staff were highly specialized, we interviewed them separately.
Some participants struggled somewhat in developing estimates for functions they perform infrequently (e.g., non-cooperation, good cause, and foster care openings in rural counties.)

3.5. Time Logs

The three counties participating in the time log study were asked to keep time logs for the last week of March and the second week of April. Staff were given a set of instructions and forms for the time log (displayed in Appendix D) and a list of the functions, processes and tasks about which we were gathering information from the focus groups (previously displayed in Exhibit 6 and Appendix A).

Staff were asked to:

keep daily time logs;
enter the amount of time spent on each child support process (as identified from the list);
enter the number of cases they worked on while performing a specific process (e.g., number of cases in which the staff person was trying to locate the noncustodial parent);
enter the number of completed processes during the time period (e.g., the number of cases in which the staff person actually located the noncustodial parent);
tally any interruption that lasted at least 10 minutes; and
log personal leave time code.

Case management, which is not always case specific, was considered a separate process. Staff noted how much time they spent on case management, but were not required to report the number of cases they worked on during that time.

3.6. Time Estimates from the Focus Groups and Time Logs

Exhibit 9 summarizes the time estimates from the focus groups and time logs. A more detailed summary of the time estimates including ranges is provided in Appendix E.

Table 3.4. Exhibit 9: Minutes Required to Complete Child Support Processes

	Focus Groups	Time Logs
	Average	Median	Average	Median
INTAKE
01 Case Opening-Non-IV-A	131	106	51	45
02 Case Opening-IV-A	129	98	40	40
03 Case Opening-IV-E	96	94	125	125
04 Non-cooperation	53	49	13	13
05 Good Cause	76	70	NC	NC
06 Case Opening-Interstate (Responding)	100	96	98	98
07 Case Opening-Medical	119	98	40	40
LOCATE
08 Locate NCP Address	56	54	5	7
09 Locate NCP Employer	57	55	9	5
10 Locate NCP Assets	51	18	5	5
ESTABLISHMENT
11 Establishment Support Order-administratively	273	262	72	77
12 Establish Paternity Order-administratively	295	308	39	40
13 Responding Interstate Establishment	329	290	59	59
14 Initiating Interstate Establishment	225	195	148	148
15 Establish –Judicial	373	387	212	212
16 Responding Interstate Establishment –Judicial	377	330	21	21
17 Case Management –Establishment (per day)	177	193	12	9
18 Non-Cooperation	36	35	NC	NC
19 Case Closure	27	23	12	8
20 Genetic Testing	86	85	113	113
21 Prepare for Hearing	98	93	68	88
ENFORCEMENT
22 Income Withholding	41	23	14	14
23 Review & Adjustment Orders	236	226	135	67
24 Enforcement Case Management (per day)	167	178	95	43
24 Enforcement Case Management (per month)	912	900	NA	NA
25 Case Closure	43	54	13	15
26 Medical Support	41	31	10	10
27 Non-Cooperation	38	35	9	9
28 Administrative Enforcement (liens, levy, tax intercept, credit reporting, license suspension)	Included in Case Management
29 Prepare for Hearing	122	111	106	115
30 Contempt	151	175	160	160
31 Initiating Interstate Enforcement	132	120	19	21
FISCAL
32-33 Fiscal Activities (total hours per day – sum of all staff)	9	4	7	7
Ordered Cases per Fiscal FTE	2040	2159	4390	3050
CUSTOMER SERVICE
34-36 Phone & Walk-ins (total hours per day – sum of all staff)	31	11	10	14
Total Cases per Customer Service FTE	1869	1461	3751	3789
NC-none completed

There are several areas of observations from the information displayed in Exhibit 9 that deserve special mention.

3.6.1. Differences between Focus Groups and Time Logs

The first observation to be made is that generally the time estimates from the focus groups are higher than those from the time logs. One problem with the time logs is that they were only kept for a week period while several child support processes (e.g., order establishment, intake) may occur over a longer period due to notification, scheduling and waiting for returned information.

3.6.2. Amount of time to complete general functions

Intake. The amount of time required to complete an intake did not vary consistently by the type of intake (i.e., IVA, NIVA, Foster Care, Interstate, and Medical); that is, no type took consistently less time or more time than the other types.
Locate. With the exception of two sites, most of the participants reported locating NCP address and employer in tandem.
Establishment. The time required to complete several types of establishments was captured (e.g., paternity administratively, order administratively, judicial, responding interstate; and initiating interstate). With the exception of judicial established orders and initiating interstate, there are not large time differences required to complete these types of establishments. However, judicial establishments took more time and initiating interstates took less time than the other types of establishment.
Enforcement. A large proportion of time for enforcement activities is spent on case management. As is evident in Exhibit 9, the focus groups found that this averages around 3 hours (167 minutes) per day. We also estimated a monthly amount to review monthly reports (e.g., caseload listings and drivers' license suspension). As a result most administrative enforcement is included in the monthly enforcement case management.
Fiscal. The number of ordered cases to hours spent on fiscal activities did not vary significantly between the Counties participating in the focus groups and time logs (this is evidenced in Appendix E). Most were in the range of 2000 ordered cases per fiscal FTE. The exceptions are two counties, one that participated in the time logs has a high ratio of ordered cases per fiscal FTE and dragged up the average amount shown in Exhibit 9.
Customer Service. As discussed earlier, it was difficult to separate case management from customer service particularly in the smaller counties where staff are more likely to perform both functions. In part, this may explain why the amount of time spent on customer service relative to caseload is larger for the smaller counties. Yet, it was also difficult to separate customer service from case management in Adams County, where there is not a specialized customer service unit.

In visiting the County offices, it was evident that Counties have different philosophies about customer services that also may affect how much time they send on activities. For example, one County requires non-IVA applicants to have all of their forms completed and copies of required documents prior to any assistance. Conversely, another County reported that they assist non-IVA applicants with completing applications and finding required documents (e.g., birth certificates).

3.6.3. Amount of time to complete specific processes

As is evident in the more detailed time estimates presented in Appendix E, there was considerable variation in time estimates for some processes by County. For example, one County reported a good cause exemption took 5 minutes whereas another county reported that it took 158 minutes on average. In situations such as this, where there are large discrepancies, the discrepancy usually results from the difference between the County’s approach to child support. For example, in the County reporting 5 minutes to complete a good cause exemption perceived it as a piece of paper that needed to be filled out and chronicled. On the other hand, the county that reported 158 minutes investigated whether the criteria for good cause were truly met. This investigation resulted in more time spent.

3.7. Alternative Methodology

As discussed earlier, we explored an alternative approach to developing staffing standards that does not rely on time estimates. Using County statistical data, we conducted analyses to determine whether staffing levels are correlated with the following State and Federal performance indicators:

Proportion of non-marital children with paternity established;
Proportion of IV-D cases with child support orders established;
Proportion of current child support obligations collected; and
Proportion of cases with arrearages that have a collection.

We did not include cost effectiveness ratios, which are also a Federal performance indicator, because of their extreme variability, particularly in small counties.

County data were available for calendar year 1999 and for the first four months of 2000. These time periods were estimated separately and compared for consistency in the estimates. National data are available from 1998 and are incomplete for some states.

3.7.1. Overview of Findings from Statistical Analysis

This section presents the results from the multiple linear regressions used to statistically estimate the impact of staffing on performance. (The specific results are displayed in Appendix F.)

3.7.1.1. PEP

The relationship between staffing and PEP is statistically insignificant in Colorado. (This was also statistically insignificant in the analysis of all States.) We are not surprised at this finding based on our conversations with County and State IV-D administrators who report that a larger share of their intakes already have paternity established. Many credit the increase to in-hospital paternity acknowledgment programs. Others suggested that in-hospital paternity acknowledgment programs along with other factors (e.g., the courts) have a greater impact on PEP than staffing levels.
The relationship between PEP and unemployment is statistically significant in Colorado. The findings indicate that a 1 percentage point increase in the unemployment rate will lower PEP by 3 percentage points. This is consistent with other research indicating that unemployment rates are negatively correlated with paternity acknowledgment rates.^[17]

3.7.1.2. Percent of cases under order

PEP and the percent of cases under order are statistically positively correlated. In Colorado, a 3 percentage point increase in PEP increases the percent of cases with orders by 1 percent. Nationally, a 2 percentage point increase in PEP increases the percent of cases with orders by 1 percent.
We suspect that this is not a one-to-one relationship because the two variables are measured in different units. PEP measures the percent of children born out-of-wedlock with paternity established, while the other indicator is measured as a percentage of cases. We do not believe much should be inferred about the difference in the relationship between Colorado and the nation. States vary in how they define PEP. We would want to control for these differences before making any inferences from the data.
The relationship between staffing levels and the percent of cases with orders yields statistically anomalous results. Specifically, there is a positive correlation between the two. We believe this an artifact of the existing Colorado staffing standards which recommend more staff per ordered case than non-ordered case.
The statistical analysis of all States shows a negative but statistically insignificant correlation between staffing and the percent of cases with orders.

3.7.1.3. Percent of current support paid

The number of cases per FTE is negatively correlated with the percent of current support paid. This correlation is statistically significant. Specifically, increasing the number of cases per FTE by 100 decreases the percent of current support paid by 3-4 percentage points. However, there are some diminishing returns, so the ratio of 100 to 3-4 does not hold precisely for all staffing levels.
The percentage of cases under order is positively correlated with the percent of current support paid - a 3-5 percentage point increase in the cases under order increases the payment of current support by 1 percentage point.
The proportion of cases located also affects the percent of current support paid. A 5 percentage point increase in located cases increases current support paid by 1 percentage point.

3.7.1.4. Percent of arrears cases with payment

Payment of current support and the percentage of arrears cases with payments are highly and positively correlated. The analysis indicates that a 1.6-1.8 percentage point increase in the payment of current support increases the proportion of arrears cases with payments by 1 percentage point.

^[17]Freya Lund Sonenstein, Pamela Holcomb, and Kristin Seefeldt (1994) Promising Approaches to Improving Paternity Establishment Rates at the Local Leve in Child Support and Child Well-Being, Irwin Garfinel, Sara S. Mclanahan and Philop K. Robins, eds., The Urban Institute Press, Washington, D.C.