Evaluating Nonprofit Databases
Kirsten A. Grønbjerg Indiana University
ARNOVA Annual Conference, 2001
Miami, November 29 - December 1
Why Evaluate Nonprofit Databases?
Sector’s Increasing size and economic force
Doubling in size over 25-30 years
Sector’s growing visibility and policy relevance
Long-term “devolution” from federal to state/local levels Long-term efforts to privatize government - nonprofits Increased political/policy role of religious institutions
“Charitable Choice” – “Armies of Compassion”
Sector’s visibility to social scientists
Focus on building community capacity Concern about state of social capital, civic engagement
Need data on sector’s scope, characteristics, trends
Methodological Developments
National data
IRS registered tax-exempt nonprofits
NCCS, IS, CNP: analysis of charitable filers Guidestar: web site of registered charities Foundation Center: analysis of foundation grants Taxonomies (NTEE) and data quality efforts
Census of service industry, population
Nonprofit Almanac – IRS and Census data: national, state
State level data
State charity officials Employment registration data Incorporation listings Department of Revenues, Tax
Local level data
Variety of approaches, use of available lists
Statement of the Problem
Main database (IRS) is problematic
Registration requirements (churches, small) Filing requirements (small, affiliates) Geographic biases (headquarter/backwater; fiscal agents) Datedness (young, defunct) Incomplete enforcement, quality control Vary by field, religion, age, size, revenues, region No common identifiers to link with most other databases Knowledge about sector may be incomplete and/or biased
Non-registration/filing rates are substantial Implications:
Indiana Nonprofit Sector Project
Phase I – in process
Comprehensive database of Indiana nonprofits
Examine gaps and overlaps in databases
Survey nonprofits statewide and in 12 communities
Profile sector and databases/approaches
Phase II – in planning stage
Compare communities and their nonprofit profiles
Phases III and IV - tentative
Linkages/networks – focus on role of congregations Curriculum additions & adaptations (COP)
Building Comprehensive Nonprofit Database
Institutional Database Approach
Combine three databases: IRS, SOS, CL Then draw sample
Informant/Community Based Approach
Supplement institutional database approach Then draw sample
Hyper-network Approach
Alternative sampling strategy
Survey individuals on nonprofit affiliations, involvement Use listed nonprofits as sampling frame, avoid database
Institutional Approach
Combine three databases
All IRS BMF 501(c) and NCCS Core filers w/ Indiana addresses
32,600 (ca. 1999, 2000) – exclude affiliates located in other states, those with end-dates, internal duplicates
SOS incorporated as nonprofits in Indiana
29,400 (2000, 2001) – exclude out-of-state addresses, internal duplicates
Congregational listings from yellow pages
9,000 (2000, 2001) – exclude internal duplicates
Total of 71,000 – including cross-list duplicates
Assessing Institutional Approach
Combined database – 54,100 (from 71,000)
Major gaps – 71 percent on only one list
IRS – broadest coverage (60 percent) SOS – substantial coverage (55 percent) CL – narrowest coverage (17 percent)
Small IRS/SOS overlap – 23 percent
Diverse institutional purposes Differ in combinations of active/defunct entities Differ in timing of listings Differ in nature of geographic biases Hidden duplicates (process, name/address differences)
But broad consistency across communities
Informant/Community Based Approach
Use of all available local listings
11 communities (size, region, characteristics) 8 local field associates Extensive field guide
Added 5,000 for 11 communities
Smallest additions: 8-15 percent (fewest lists) Most communities: 20-28 percent Extreme: 64 percent (Scott County)
If done for entire state – perhaps 65,000 total
Personal Affiliation Approach
Hyper-network approach
Survey of 526 Indiana residents (May 2001)
If work – In Indiana? For nonprofit? Which (up to 2)? If attend church – In Indiana? Which? If attend membership events (21 types) – In Indiana? Which (up to 5 for each type)? If do volunteer work (10 types) – In Indiana? Which (up to 5 for each type)?
Identified 1,290 organizations
Deleted 184 (15%): 113 public/for-profit, 71 duplicates Remaining 1,106: 57% on database, 13% incomplete ID, 30% added to database
Database Assessment
Assessment of work
Complicated, difficult process – chart Special database review and source fields Tentative coding of nonprofit field, type
Preliminary findings
Incomplete nonprofit databases and sector definitions Differences in sector composition by database Community variations in database size and composition
Extrapolating to the U.S. nonprofit sector
IRS registered nonprofits ~ 50% 2.5 million IRS-registered charitable filers ~ 10% 2.7 million Nonprofit density ~ 100 per 10,000 residents 2.8 million
Next Steps – Nonprofit Survey
Obtain more detailed information through survey
Stratified sampling design Survey design
Assess relevance of preliminary findings Address policy and substantive questions about sector
Allow methodological comparisons of lists and approaches Allow Community comparisons – expanded samples Format: Mail questionnaire (20 minutes), extensive follow-up All: Age, legal status, programs/services, impact of community/policy changes, relations with other organizations, human resources, finances, management Special: Membership associations, congregations, advocacy nonprofits