RKS Findings

Odyssey Respondent Research
Record Keeping Study

Odyssey Program Seminar

February 26, 2020

Diane Willimack

Erica Marquette

Demi Hanna

ESMD

<number>

<number>

 

Outline

• •Current Collection Strategy 

• •Research Goals and Methods 

• •Respondent Profile 

• •Preliminary Findings and Conclusions 

• •Next Step Recommendations 

<number>

 

CURRENT Collection Strategy: Industry-based

<number>

<number>

<number>

 

Research Goals

• •DEFINITION: How do companies define data items based upon their charts of accounts and financial reporting requirements? Can we determine a harmonized definition that aligns with company records? 

• •UNIT: What data are available at what level?  (e.g. establishment, company, industry, state) 

• •TIMING: When are the data available? Are different data items available at different times? If so, what and when?  

• •BURDEN: How readily available is the information we are asking? Are some items easier? Harder? Why? 

<number>

 

Research Methods

• •Created interview guide with exploratory questions, for example:  

• • •Definitions: In what industry do you classify your business? 

  • •Unit: At what level is the data available? 

  • •Timing & Burden: How much time does it take and how many people or data sources are involved? 

  • •Burden: How much manipulation of data in business records is involved in order to provide data that meets Census Bureau requirements 

• •Interviewed 28 “medium size” companies August 5 –  November 14, 2019, in 4 cities in the Northeast and Southern regions 

<number>

 

Respondent Profile

• •All companies were sampled in at least two in-scope surveys 

• •Nearly all interviewed companies were in at least 2 sectors 

• •13 companies had over 50 establishments 

• •17 companies operated in more than 5 states 

• •4 public for-profit companies, 6 not for-profit companies 

• •Over 70 different 6-digit NAICS 

<number>

 

Preliminary Findings: Definition

• •All companies followed a general chart of accounts, the difference was in the detail  

• •At least 7 companies may have been misclassified or may not have understood our distinction among classifications 

• • •Different questionnaires = different industries (4- or 6-digit NAICS) 

<number>

 

Preliminary Findings: Unit

• •Companies do not use the term “establishment”         

• • •Region, Office, Location 

  • •Cost center, Department, Business segment 

• •“Establishment” level data: 

  • •For some companies, revenue was not easily attributed to individual locations 

  • •If location is meaningful to management decisions, records were kept to support that 

  • •Some companies track information by “establishment” for budget purposes 

• •For almost all of the companies, their product details did not align with our categories  

<number>

Management decisions were not the only reason things were kept by location

<number>

<number>

 

Preliminary Findings: Timing

• •17 of the companies had fiscal years that were approximately calendar year 

• •Almost all companies said June was a good time for survey response 

• •50/50 split among respondents between reporting survey data all at once and staggering it 

• •Want some type of “reporting calendar” so they know what is coming and when and what is the due date 

<number>

 

Preliminary Findings: Burden

• •Nearly all companies said they had to ask internal colleagues for information to complete the surveys 

• •Census surveys do not match up to internal reporting 

• • •Management reports 

  • •Financial reports 

  • •Regulatory reports 

• •Respondents do not like to make decisions on how to manipulate their data to match our requests 

<number>

 

Conclusions

• •Sources of burden: 

• • •Identifying the portion of their company to report for (on different forms) 

  • •Interpreting the meaning of our questions 

  • •Allocating their data to what (they think) we’re asking for 

• •From the respondents’ perspective, NAICS classifications tends to be artificial 

  • •They don’t recognize our distinctions across forms 

  • •Confused when we only ask for data about a piece of their company 

• Our lack of a holistic view of their company in data collection causes respondents confusion in responding to our surveys, which impacts: 

• • oData quality 

  • oReporting burden 

<number>

 

Next Step Recommendations

• •Create proof of concept instrument that: 

• • •Starts with consolidated figures (e.g., “top level numbers”) 

  • •Asks for breakdown of data by: 

• • • •Location? 

    • •Business segment?  

    • •Cost center? 

  • •Follows major categories of an Income Statement or Balance Sheet 

• •Conduct cognitive testing of multiple alternatives 

  • •Which data are readily available at what level? 

  • •How to help respondents “map” their data to our requests? 

    • •Nature and degree of discrepancies 

<number>

 

CURRENT Collection Strategy: Industry-based

<number>

<number>

<number>

 

Proof of Concept PROPOSED Collection Strategy: Topic/Account-based

<number>

<number>

<number>

 

Proof of Concept PROPOSED Collection Strategy: Topic/Account-based

<number>

<number>

<number>

 

Thank you!

Diane K. Willimack

diane.k.willimack@census.gov

x3-3538

Erica Marquette

erica.marquette@census.gov

x3-7603

Demi Hanna

demetria.v.hanna@census.gov

x3-3351

Kristin Stettler

kristin.j.stettler@census.gov

x3-7596

Melissa Cidade

melissa.cidade@census.gov

x3-8325

<number>

 

File Type	application/vnd.openxmlformats-officedocument.presentationml.presentation
File Modified	0000-00-00
File Created	0000-00-00

---

© 2024 OMB.report | Privacy Policy

---