Statistical Computing and the Year 2000 (the y2k problem)

Who is likely to have a Y2k problem in Statistical Computing?

These pages discuss ways to deal with the "Y2K" problem as it relates to Statistical Computing.  (Note this does not address any other kinds of Y2K problems that you might have with your computer hardware or other kinds of computer software).  Before diving into these problems and solutions, you might ask yourself the following questions to see if you are likely to have a Y2K problem related to your statistical computing.

1. Do you analyze data that contains dates?  If the answer to this is no, then you probably won't have a Y2k problem for Statistical Computing.

2. If you analyze data that contains dates, are they stored using 4 digits to represent the year (e.g., 1999, not 99).  If yes, then you probably won't have a Y2K problem for Statistical Computing.

3. If your dates are stored using only 2 digits, are the dates contained in your data going to start crossing the year 2000 any time soon.  For example, you might be analyzing birth dates and you will soon be getting data for those born after the year 2000.  Or, you may analyze dates of only adults, and it will be years before any of the people in your data files will be born in the year 2000.  If your dates will not contain the year 2000 for a long time, then you probably won't have a Y2K problem in Statistical Computing.  

If you do analyze dates, and your dates are stored only using 2 digits for the year, and you will soon be encountering dates in your data that are in the year 2000, then you will soon need to deal with Y2K problems in your data.  The following web pages discuss how you can address these problems in SAS, in SPSS and in Stata.  

SAS

SAS stores dates internally so all Gregorian calendar dates between the years 1582 and 20,000 are handled correctly by SAS System date and time functions.  Even so, the data you read into SAS and your SAS programs may implicity assume that dates are all from the 1900s (1900 to 1999).  The following links help identify Year 2000 problems in your data and SAS programs, and how to solve them.

• ATS has created a White Paper Using Dates in SAS in the Year 2000 and Beyond which shows example programs to illustrate the problems which can arise when handing dates in the Year 2000 and beyond in SAS, and some examples of solutions to those problems.  Even if you are not using SAS, these examples may be useful for detecting Year 2000 problems in other applications and strategies for solving them.
• The SAS site has useful information, including...
• Be sure to see their information describing the YEARCUTOFF= OPTION.
• They also have a fantastic White Paper entitled  The Year 2000: Preparing for the Inevitable .

SPSS

SPSS stores dates internally so all Gregorian calendar dates between the years 1582 and 9999 are handled correctly by SPSS.  Even so, the data you read into SPSS and your SPSS programs may implicity assume that dates are all from the 1900s (1900 to 1999).  The following links help identify Year 2000 problems in your data and SPSS programs, and how to solve them.

• ATS has created a White Paper Using Dates in SPSS in the Year 2000 and Beyond which shows example programs to illustrate the problems which can arise when handing dates in the Year 2000 and beyond in SPSS, and some examples of solutions to those problems.  Even if you are not using SPSS, these examples may be useful for detecting Year 2000 problems in other applications and strategies for solving them.

STATA

Stata stores dates internally so it can deal with dates from Jan 1, 100 to Dec 31, 9999 (but one should be cautious dealing with dates before October 15, 1582 when the Gregorian calendar was put in effect).  Even so, the data you read into Stata and your Stata programs may implicity assume that dates are all from the 1900s (1900 to 1999).  The following links help identify Year 2000 problems in your data and SPSS programs, and how to solve them.

• ATS has created a White Paper Using Dates in Stata in the Year 2000 and Beyond which shows example programs to illustrate the problems which can arise when handing dates in the Year 2000 and beyond in Stata, and some examples of solutions to those problems.  Even if you are not using Stata, these examples may be useful for detecting Year 2000 problems in other applications and strategies for solving them.
• Be sure to see the Stata Year 2000 Compliance Page and check for useful links at the Stata home page.

For More Information

For more information about the year 2000 and statistical computing, please feel free to use ATS Statistical Consulting.

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services