Evaluation of the Implementation of the Child Support Guidelines


USER'S GUIDE TO THE PUBLIC USE DATABASE



In October 1994, CSR, Incorporated, and its subcontractor, the American Bar Association's Center on Children and the Law, conducted an evaluation of the implementation of child support guidelines under a contract with the Office of Child Support Enforcement. The evaluation was conducted from October 1994 to March 1996. The major components of the evaluation included the following: collection and analysis of child support cases from 21 counties in 11 States; interviews with key participants in the child support system in the study counties; collection and analysis of State guideline reviews; and analysis of data from the Current Population Survey.

This User's Guide to the Public Use Database is Volume III of the final report on this evaluation. Volume I presents the main study findings, integrating findings from all components. A comprehensive description of the research questions, methodology, and history of child support guidelines implementation are in Volume I, Chapter 1 of this report, which also describes study methodology and limitations on the data. Additional details on the methodology are contained in Appendix A of Volume I, and Appendix B of Volume I contains typical cases received from each of the study States. Volume II of the report analyses the State guideline reviews, deviations in those reviews, and interviews conducted with stakeholders in the study counties.

This guide is divided into four parts. These sections describe the nature of the worksheet and supplemental case record data collected, the data extraction and coding, and the database structure and provides guidance for accessing and interpreting the database. Appendix A contains the codebook for the supplemental data used for all States and counties, Appendix B replicates the data entry screens used for each county, and Appendix C provides detailed instructions for installing and using the databases.

NATURE OF THE DATA COLLECTED FROM THE 21 STUDY COUNTIES

The collected data were of two types: (1) worksheet information containing data on the income and expenses of the parties, adjustments, and the calculations of award amounts and (2) supplemental information from the case records on deviations, the size and composition of families, the type of award, case type, and other descriptive data. Although strenuous efforts were made by local court and IV-D agency personnel and by local personnel engaged by CSR to obtain complete case records containing both worksheet and supplemental data, some case files were incomplete.

Methodology and Data Limitations

The sampling plan for the evaluation imposes important limitations on use of the data. The data contained in the databases described in this user's guide were coded from the case records collected from the 21 study counties. Two counties from each of 11 States participated in the case record collection component of the study, with the exception of the State of Minnesota where data from only one county was included in the study. It is important to emphasize that the States and counties were selected purposively to represent the major guideline models (income shares, percent of income, and Melson) and for their willingness to participate in the evaluation. Therefore, these data cannot be used to generalize to any population outside these cases.

The case records were coded into two types of databases corresponding to the collected information: (1) worksheets used to calculate support awards and (2) supplemental data describing the characteristics of each case. It is important to note that most of the analyses reported in Volume I of the report were conducted using the supplemental databases, which were uniform and consistently coded across all cases and all counties. The worksheet databases were designed specifically to replicate each State's worksheet and the specific provisions of each State's guidelines. The worksheet databases are not uniform or consistent across the States, significant effort is required to interpret each worksheet database, and combining data across States and counties is extremely challenging. This guide is designed to allow the user to access and interpret the database. It will explain the nature of the data collected from each study county, the database structure, and the steps needed to use the data for analysis.

Sources of Data

The goal of the data collection was to obtain 200 case records from each county. The case records were intended to include all written documentation about the case. A variety of documents were received such as:

Standardized child support order forms;

Child support worksheets;

Financial affidavits;

Employment security records printouts;

Letters to employers;

Case notes from decisionmakers; and

Pay stubs and tax forms.

Table 1 following this page lists the 21 counties from which data were collected and the number of case records of each type (i.e., worksheet, supplemental data, or both) included in the research database. One county in New Jersey provided no usable worksheet data. It is important to note that no oral records or transcripts were collected, so that any facts or decisions not in the written record of the case were not available for analysis.

The study methodology was designed to collect data on the full range of support cases decided in the counties, including both AFDC cases and non-AFDC cases handled by the IV-D agency, as well as non-IV-D cases. Cases across all counties were nearly equally divided among three categories: (1) AFDC cases, (2) IV-D non-AFDC cases, and (3) non-IV-D cases. In Pennsylvania all child support cases are handled by the IV-D agency, and New Jersey does not make an administrative distinction between IV-D and non-IV-D cases. Child support orders applicable to the study included the following:

Temporary orders (10.5 percent);

Permanent orders (63.0 percent);

Interim orders (5.6 percent); and

Award modifications (20.9 percent).

The awards could be made in any type of case, such as separations, dissolutions, or paternity establishments. Paternity establishments were most frequent (29.3 percent of cases), followed by dissolutions (26.2 percent). Modifications to existing awards did not report a separate case type, and 18.6 percent of the records had missing case types.

Cases were collected in all relevant forums where cases are heard in the county, including courts, quasi-judicial settings, and administrative bodies, as follows: 38.9 percent of the cases were heard in courts; 51.0 percent were heard in quasi-judicial settings; and 6.7 percent were heard before administrative bodies. There was less uniformity than expected within States--only two States (Arkansas and Massachusetts) heard all cases in judicial forums, while in most States, many forums were used. Different case types were heard in different forums in some counties. Only 14.4 percent of non-IV-D cases were heard in judicial forums, compared with 38.9 percent of all cases, while 62 percent of non-IV-D cases were heard in quasi-judicial forums, and 22.4 percent were heard in administrative forums.

DATA EXTRACTION AND CODING

The following sections discuss data extraction of the supplemental and worksheet information from each case record. The supplemental data are discussed first because the supplemental information contained in the databases is consistent and uniform across all States and counties. In contrast, the worksheet data are specific to each county.

Supplemental Information

In order to address the research questions of the study, a wealth of information was extracted from each case record by CSR staff using the Supplemental Information Codebook (reproduced as Appendix A of this volume). The data were extracted onto coding sheets and then entered into the database for each county. The supplemental data were organized into the following topics in the Codebook:

A. Case Characteristics [e.g., interstate, case type (paternity, separation, dissolution), case status (IV-D AFDC, IV-D other, non-IV-D), contested or negotiated, type of award (permanent, temporary, modification), jurisdiction, forum (judicial, administrative), obligor (father, mother), arrears, obligee];

B. Characteristics of the children [i.e., number of children in the order, age of youngest and oldest, and any special needs children (physically or emotionally disabled, gifted)];

C. Characteristics of the order (e.g., amount of order, wage attachment, termination of support, retroactivity);

D. Income information on the parents [e.g., employment status of the parties, presence of income verification (pay stubs, affidavit), reason for imputing income (unemployment, failure to appear), method used for imputing income (minimum wage, past earnings)];

E. Deviations [e.g., presence of explicit statement in record of deviation, reason for deviation (second family, custody expenses), amount and direction of deviation];

F. Health care needs [e.g., presence of order for one or more parent to maintain or obtain health insurance, how costs incorporated into order (added to basic need, income deduction), provision for payment of insurance (parties split proportionally, noncustodial parent pays 100 percent), provision in order for extraordinary medical costs (one parent pays 100 percent, parties split 50/50 or proportionally)];

G. Day care costs [e.g., day care costs a part of the order, amount of day care costs, how costs incorporated into order (added to basic need, income deduction), provision for payment of day care costs (parties split proportionally, noncustodial parent pays 100 percent), consideration of federal tax credit];

H. Postsecondary education provisions [e.g., provision for postsecondary costs (immediate costs for college or vocational education, provision for future college and/or vocational education costs), provision for payment of costs (parties split proportionally or 50/50)];

I. Other expenses [e.g., provision for additional expenses not included in the child support amount (tutoring expenses, lessons), provision for payment of additional costs (parties split proportionally or 50/50, noncustodial parent pays 100 percent)];

J. Tax exemptions [i.e., provision for dependency tax exemptions (custodial parent always receives exemptions, parties alternate the exemptions)];

K. Custody and visitation [e.g., type of custody arrangement (joint legal custody with primary physical custody to custodial parent and liberal visitation rights granted to the noncustodial parent, split custody, joint/shared custody), effect of custody arrangement on support order (no effect, orders offset), extended visitation considerations (support abatement);

L. Presence of parents and/or attorneys (e.g., obligor and/or obligee represented by a private attorney or legal services, attorney present for the State); and

M. Multiple families [e.g., noncustodial parent and/or custodial parent having a previous child support order or current support obligation not subject to a court order, effect of other obligation on current action (income adjustment, deviation)].

Information related to the above topics was coded and entered into a database for each case record. The codebook associated with this task is in Appendix A of this volume.

Child Support Worksheet Information

Eight of the 11 States participating in this study use a child support guideline worksheet and incorporate it into most child support actions.(1) Many cases received for the study contained some form of worksheet. In most cases it was readily apparent if the worksheet (or one of the worksheets) in the case record was, in fact, the worksheet used to determine the award. For those cases that did not contain a worksheet or a determination could not be made as to which worksheet was used to determine the award, the case was kept in the sample of supplemental data but no worksheet data was entered.(2) In a relatively small number of cases a worksheet could be reconstructed from case notes, income statements, statements in the order itself, etc. These cases were coded as having a reconstructed worksheet.

DATABASE STRUCTURE

For each county there are 2 databases, a supplemental database and a worksheet database.(3) The databases were named using the two-letter State postal abbreviation, a county number (1 or 2), and an extension (SUP or WS). For example, the naming convention for the second county in New Hampshire is as follows:

"NH2-SUP" is the supplemental database for county 2 in New Hampshire; and

"NH2-WS" is the worksheet database for county 2 in New Hampshire.

As a part of case identification, each case was assigned the appropriate county code and a sequential number. For example, the one-hundredth case received from county 2 in New Hampshire was identified as "NH2100". In each data file this identifier is split in 2 fields: "ST_CO" contains the county code (NH2) and "CSRNO" contains the sequential number (100). This identification number links the supplemental and worksheet data for each case. The supplemental information for this case is in the "NH2-SUP" database and the worksheet information in the "NH2-WS" database. Table 1 identifies the county codes and number of records in each data file.

Paradox and DBase-IV Data Files

Data files in both Paradox and Dbf-IV formats are available for public use. The databases were originally designed using Paradox 4.5 for DOS, and this is the most user friendly format with respect to the worksheet files. A data entry screen was designed to resemble each State's child support worksheet. These are extremely useful when looking at the data on-screen. These specialized screens do not accompany the file when exported to DBase-IV format. However, the screens have been reproduced and are in Appendix B of this volume.

The supplemental files were standardized across all counties and did not require specialized data entry screens.

Data Fields

The following sections discuss the data fields for the supplemental and worksheet databases.

Supplemental Database

The standardized supplemental database is comprised of numeric fields, with explanatory text in memo fields for selected fields.

Supplemental information was coded using pre-defined numeric codes. Fields containing dollar amounts are also numeric fields.

Many supplemental questions have corresponding memo fields (e.g., C2 has a corresponding memo field called C2T). These fields were used to describe situations that could not be captured by any of the pre-defined codes.

The supplemental files were standardized across all counties. The field names are identical in every "-SUP" database. The supplemental codebook, in Appendix A of this volume, is the roadmap to locating the needed information from each case record.

Worksheet Database

The majority of data fields in the worksheet databases are numeric fields. A small number of worksheet databases contain character (memo) fields that describe income sources, debts, assets, etc. These character fields were rarely used. Additionally, a memo field was added to a few databases to allow for comments on the case to be entered by the coder. These are useful for interpreting difficult cases.

The worksheet files are more complex than the supplemental files, and are designed with formats and fields specific to each State in the study. Child support worksheets vary considerably from State to State and similar terms have quite different meanings. Following are some examples of the variability:

The cost of health insurance premiums is a deduction from income in some States, in others they are an addition to the support need, and in other States they are a credit against the support obligation (similar differences exist in the treatment of other costs such as those for day care).

The "income available for support" may be a gross income figure, a net income figure, or not described by either of these terms. States vary considerably in how this available income is arrived at and many have intermediate income figures. For example, in one State this figure may be gross income minus federal, State, and local taxes, while in another State this figure is defined as gross minus other support obligations.

These variations made standardizing field names across worksheets a difficult task. However, some standardization was possible in the areas of income, deductions from income, obligation terms, and child care. Table 2 following this page summarizes some common field names and their definitions across counties. Consulting the specialized data entry screen is still necessary, and highly recommended, to be aware of county specific fields and nuances. In addition, at the bottom of each specialized screen in Appendix B are notes that highlight areas that require special attention and provide general information about the database. The structure of each worksheet database and field descriptions follows each custom data entry screen.

Another difference across States is the labeling convention used to identify the relative parties. Three conventions were used by the States in this study: father and mother, obligor and obligee, and noncustodial and custodial parent. To determine which parent is the obligor in those States where the parties are labeled mother and father, the supplemental file for the case must be referenced (field "A8" in the supplemental file).

For purposes of quality control over the dataentry process, all computed fields in each worksheet were recalculated by Paradox after the data were entered into the database. For each computational step on the worksheet there is a field in the database preceded by "C_". This field represents the Paradox-generated value for that computational step. If a discrepancy existed between an original and a recomputed value, a "1" appears in the "ERROR" field. It is important to note, however, that this field cannot be used to identify math errors in general. Because incomplete worksheets were the rule, not the exception, many worksheets are marked as having a computational error as the result of blank fields. For example, it was common for deductions from income not to be itemized. This creates a discrepancy in recomputing the award. Therefore, these fields should be interpreted cautiously.



ACCESSING AND INTERPRETING THE DATABASE

The supplemental and worksheet records for each case can be linked through the unique case identifier. However, not every supplemental record has a corresponding worksheet file. Some records were incomplete or did not contain a worksheet for other reasons unique to a given case. For example, in some counties if a noncustodial defaults, an order is automatically entered for a specific amount and a worksheet is not completed. The following sections discuss issues related to using both databases.

Worksheet Information

Because the goal of the supplemental coding was to capture the process of incorporating various factors into the child support calculation, certain elements can only be found in the worksheet database, some of which include the following:

Income figures;

Actual dollar amounts for deductions, health care, prior orders, subsequent obligations, etc.;

Basic child support amounts derived from look-up tables or percent multipliers; and

The presumed or "guideline" amount of child support. This last element is discussed in greater detail in the next section.

Presumed or Guideline Child Support Amount

Two important fields in the worksheet file are "NSUPPORD" and "CSUPPORD". These represent each parent's guideline amount of support.(4) In those cases that do not indicate a deviation, this amount should match the amount coded in "C2" (amount of support order) in the supplemental file. In those cases that do indicate a deviation, the amount in "C2" is after the deviation. Therefore, the worksheet fields are the pre-deviation amount. For deviation cases, when the amount in "C2" is combined with the amount in "E6" (adding or subtracting the two, depending on the direction of the deviation) the result should equal the worksheet amounts.

Some worksheets allow the decisionmaker to select a payment schedule (e.g., weekly, biweekly, monthly, etc.). The user should be careful to ensure that the worksheet schedule and schedule coded in the supplemental file match. The general rule to calculating monthly amounts is to multiply weekly amounts by 4.33, biweekly amounts by 2.17, and semi-monthly amounts by 2. The amount in "E6" of the supplemental is in the same schedule as that for "C2".

1. 0 The States of Arkansas, Minnesota, and Wisconsin do not have a child support worksheet. In order to collect the information needed for this study, a worksheet was developed for these States with input from State and local personnel.

2. 0 Cases were not included in the study if the information in the case record did not fully explain how the award was arrived at.

3. 0 One county in the State of New Jersey (NJ2) does not have a worksheet database. None of the cases received contained a worksheet.

4. The "NSUPPORD" field in the State of Arkansas is the NCP's obligation after any deviation. See the data entry screen in Appendix B.