Presentation is loading. Please wait.

Presentation is loading. Please wait.

Can We Continue to Exclude Small Single-Establishment Businesses from Data Collection in the Annual Retail Trade Survey and the Service Annual Survey?

Similar presentations


Presentation on theme: "Can We Continue to Exclude Small Single-Establishment Businesses from Data Collection in the Annual Retail Trade Survey and the Service Annual Survey?"— Presentation transcript:

1 Can We Continue to Exclude Small Single-Establishment Businesses from Data Collection in the Annual Retail Trade Survey and the Service Annual Survey? Authors: Kari K. Clark and David L. Kinyon Presenter: Ruth E. Detlefsen Service Sector Statistics Division U.S. Census Bureau

2 2 The Surveys Annual Retail Trade Survey (ARTS) Sales Sales tax Inventories Purchases E-commerce sales Service Annual Survey (SAS) Total and detailed revenue Total and detailed expenses E-commerce revenue

3 3 Why Exclude Small Single- Establishment Businesses? To Reduce: Respondent Burden Data Collection Costs

4 4 Restricted Data Collection - Industries –Couriers & Messengers (NAICS 492) –Warehousing & Storage (NAICS 493) –Rental & Leasing Services (NAICS 532) –Professional, Scientific, & Technical Services (NAICS 54) except Computer Systems Design & Related Services (NAICS 5415) –Administrative & Support Services (NAICS 561) –Waste Management & Remediation Services (NAICS 562) –Social Assistance (NAICS 624) –Arts, Entertainment, & Recreation (NAICS 71) –Accommodation & Food Services (NAICS 72)* –Other Services (except Public Administration) (NAICS 81)

5 5 Methodology for 1999 SAS and 2000 ARTS Cutoff was calculated by: Summing payroll of all multi-establishments Adding the payroll of single-establishments: –One at a time, beginning with the largest –Stop when sum is at least 80% of the total payroll Cutoff = payroll of last single-establishment added

6 6 Units withheld from mailing contributed: –9% of the SAS revenue estimate for all restricted data collection industries –17% of the sales estimate for NAICS 72 within ARTS. Evaluation Results

7 7 2005 ARTS and SAS Re-examined methodology due to: –Addition of detailed items –More current data

8 8 Changes to Cutoff Methodology –Source of Data –Input Data –Level of Industry Classification –Treatment of New Businesses –Evaluation Method

9 9 Source of Data Establishment List for Sample Revision –Used to develop the new sample introduced for the 2005 surveys –Includes establishments not active at the time of the Economic Census –2003 administrative data available for use

10 10 Input Data Measure of Size (MOS) instead of Payroll –MOS ≈ Annual Revenue, based on 2003 data –MOS is expected to be well-correlated with variables measured in the surveys

11 11 Level of Industry Classification Levels Used for Sampling instead of NAICS –Generally 4-, 5-, 6-digit NAICS

12 12 Treatment of New Businesses Start date between 04/01/03 and 01/01/04 Treated similarly to multi-establishments

13 13 Evaluation Method Used 70%, 80%, 90% of Total MOS –Considered varying by industry level

14 14 Calculating the Cutoffs within Each Industry Total MOS for Multi-establishments and Births Add non-birth single establishments until percent of total was met or exceeded Cutoff is MOS of last unit added

15 15 Example of Calculating the Cutoffs Sum of Multi-estabs= 500,000 Sum of Births = 25,000 Total (Starting Cumulation)= 525,000 Sum of Non-births= 350,000 Grand Total = 875,000

16 16 Example of Calculating the Cutoffs (cont) MOS - 100,000 50,000 25,000 20,000 F 5,000 ΣMOSPercent 525,00060.0% 625,00071.4% 675,00077.1% 700,00080.0% Cutoff 100,000 25,000 5,000 720,00082.2% F 790,00090.2%

17 17 Choosing the Cutoffs for 2005 Computed cutoffs for the 3 percentages Applied cutoffs to 2003 Annual Surveys Compared sales/revenue to the 3 cutoffs for small single establishments If below cutoff, data was cleared and imputed Recomputed estimates

18 18 Evaluation Criteria Coefficient of Variation <= 20% Imputation Rate <= 40% Large Changes?

19 19 Evaluation of SAS Estimates Cutoffs based on 70% had a considerable impact on estimates Only cutoffs based on 80% and 90% were examined in depth: –revenue or expenses would not be published for some industries –majority of the detail items would not be published

20 20 Final Recommendation for SAS Mail all units for SAS –The cutoffs have a considerable impact on the estimates –New detail items were to be added to the 2005 SAS

21 21 Evaluation of ARTS Estimates Cutoffs based on 80% and 90% Criteria met for previously published industries

22 22 Final Recommendation for ARTS Restricted data collection will continue for industries within NAICS 72. Applied cutoffs based on 90% 1,339 not mailed out of 5,677 total sampling units

23 23 Can We Continue to Exclude Small Single-Establishment Businesses from Data Collection in the Annual Retail Trade Survey and the Service Annual Survey? For ARTS – Yes For SAS – No

24 24 Thank You Contact Information: Kari K. Clark kari.k.clark@census.gov David L. Kinyon david.l.kinyon@census.gov david.l.kinyon@census.gov Ruth E. Detlefsen ruth.e.detlefsen@census.gov


Download ppt "Can We Continue to Exclude Small Single-Establishment Businesses from Data Collection in the Annual Retail Trade Survey and the Service Annual Survey?"

Similar presentations


Ads by Google