Skip to Page Content
Home  |  Contact Us  |  Press Room  |  Site Overview  |  Help  |  Login  |  Register
Add to MyNCSL

2002 NLPES-Fall Training Conference Notes
Database Workshop Summary

 DATABASE WORKSHOP, held September 19, 2002, 10:30 a.m.-12:00 p.m.

Instructors:
Maureen Tonn, CPA, Information Technology Division, Arizona
John Curran, MPA, Kansas Division of Post-Audit

Description: This course gave the process for requesting and using agency data from the initial assessment to data analysis.

Both Mr. Curran and Ms. Tonn stated similar needs for data collecting, validating, and analysis; however, Mr. Curran emphasized Ms. Tonn's points and specifically addressed data analysis.

Maureen's key point in data processing and analysis was "Don't put the cart before the horse." First know what you want from the data. There is a lot of data available but knowing what you want the data to tell you will provide direction for data collection. Validate the data and determine its value in relation to the importance of the analysis it will support. And, importantly, document the processes.

The process steps are: 1) initial assessment of data; 2) data request; 3) receipt of data; and 4) data analysis.

Initial Assessment of Data

Will the data answer the question?

    • What do you need to know
    • may need to adjust your scope
    • How will you answer the questions?
    • specific variables
    • evidence level needed
    • How long will it take to get needed data?
    • limits are budget, staff, programming, etc.
    • What has to be done to make data useable?
    • cleaning, restructuring

Are there alternative methods to supplement or replace the data? Keep this in mind, just in case the data turns out to be unusable.

    • What limitations do these data have regarding objective questions?
    • incomplete answer
    • not definitive enough for needed evidence level
    • missing data
    • not representative
    • biased data , i.e., certain complaints may be omitted from an agency tracking database

Assessment tools

Maureen provided a sample interview instrument for determining the usability of data

    • User interview
    • what? how? when? what do you think? observation of how data is captured
    • Program manager interview
    • how is data used
      • data use, i.e., is unused data reliable
      • accuracy
    • IT interview
    • synthesize interview information from other interviews to ask specific data questions
    • find out how data is kept and how it can be downloaded

Document-very important

    • Understand the data and get the following--
    • data dictionary
    • field list
    • code book
    • procedure manual
    • Decide data usability
    • is data available
    • does it need more work, e.g., cleanup
    • what risks are there in using data-talk to staff
    • what are data's limitations, e.g., do all complaints get put into database
    • can you do what is needed to make data good in the time available

Data Request

Ask for all the data you will need but don't waste time formatting, etc., too much data that won't be used

    • Make data request in writing
    • Consider informing agency upper management
      • agency IT resource will prioritize your request
    • Provide adequate time to gather data

Receipt of Data

Use your IT resource to convert or reformat data but keep the original data as backup. Maureen provided a sample data checklist

    • Packing slip check
      • did you receive everything?
      • are contents what you ordered?
      • any obvious defects?
    • Programs such as Data Junction can be used to convert data, and programs like Cambio can strip out headers and such from reports, and save time

Validate the data

    • Consider validation throughout the process
    • Consider alternatives if you can't use the data
    • IT questionnaire helps check data reliability
      • an IT survey instrument was provided

Clean and restructure the data

    • Are there any changes or missing data that can be cleaned or fixed to make the data usable?
      • convert data, e.g., change to new agency codes
      • interpolation of missing data
      • data restructuring to answer the questions

Data Analysis

Mr. Curran said to question:

Will the data answer the questions being asked in the project?

    • Determining the value of data is a continuous process and runs throughout the audit.
      • data may need to change to reflect the audit scope
    • Allow enough time to get the data
      • may need to go to agency head
      • consider sending a letter establishing data receipt deadlines

Contact person:
Maureen Tonn
Information Technology Services
Arizona Office of the Auditor General
mtonn@auditorgen.state.az.us
602-553-0333

Back2002 Fall Training Conference Notes

Denver Office: Tel: 303-364-7700 | Fax: 303-364-7800 | 7700 East First Place | Denver, CO 80230 | Map
Washington Office: Tel: 202-624-5400 | Fax: 202-737-1069 | 444 North Capitol Street, N.W., Suite 515 | Washington, D.C. 20001