440 - MAIN (LOG) PAGE 
 440 - OUTLINE
 EPI-INFO page  <=you are here



Note: this information will be useful for ENH440 students who want to access a statistical package for biostatistics.  This is an effective one, useful for most purposes, simple to use, though getting rather old.  ENH522 students can use this or any other statistical package (such as SPSS) to complete an assignment.


You should point your browser at the CDC Epi-info home page home page.   You will see TWO versions of Epi-Info. 

YOU CAN SELECT VERSION 6.04d (BUILT for the DOS OPERATING SYSTEM) The FAQ which follow have been written for version 6.04d.  It arrives in three large compressed (zipped) files that have the extension  .exe    NOTE:  Don't be afraid that it uses the older DOS operating system.  The programming is behind the scenes, and you won't have to know anything about DOS commands other than you look for the DOS Prompt which will look something like this C:\EPI6\:  and all your commands will be typed on the same line after that.   That's it!  The FAQ instructions below should help...


This was released relatively recently, and most of the bugs have been ironed out. It will take longer to download.  It carries a good tutorial, and is relatively easy to set up on the screen.  The tutorial below will not help, however. 

If you install v6.04, follow these instructions I suggest that you print these instructions and keep them handy.

1. You will need an empty temporary directory.  To create a new directory, click on windows explorer and then click on the C: directory to highlight it.   Then click on File  at the top of the screen, and select NEW and folder.   Give the new folder a name,  e.g., C:epitemp.   (Note: The temporary directory should not be called EPI6).

2.  I found it handy to download the program first onto 3.5" floppy diskettes.  That way, you can try the install again if it doesn't work first time.  When you click on each file on the CDC's website, you will be asked where it should be saved.  Just indicate the A: drive.   Have THREE formatted diskettes ready.  They should be fully formatted with no "bad sectors";  you will need the entire disk space.  There are three big compressed (zipped) files  

    epi604_1.exe           epi604_2.exe            epi604_3.exe           

You could also have downloaded these files directly into the empty temporary directory on the hard drive of your computer.   If you downloaded them onto floppies, copy these files into the new directory.  In either case, you should now have these three files in the new EPITEMP sub-directory (you can call it what you like, but do remember the name).

3.   Unpack the compressed files by going into windows explorer and double-clicking on each of the three files in turn (if you use Win 95/98/NT/2000). For each, a window will appear asking if you want to continue extraction. Respond with <Y>  Repeat this with all files.  You should then have around 70-80 new files as well as the three original compressed files.

4.  Scroll down to find  install.exe   and hit <enter>.  A blue INSTALL box will appear and you will be asked questions.  Here are the questions and the responses you should give:-

    source? = <C>,     

    destination? = <C>,     

    floppy disks?  = <N>,   

    Normal inst? = <(enter)>,     

     Install or copy?  = <I>

     select vid driver? = press <F8> for "all" ,    

           ....... then <F4>,    .... and again <F4>   and you are done.

5.    To place the windows Epi-Info icon on your desktop,  press START, then  RUN and enter in the box: C:\EPI6\setup.exe An icon will appear:  Click for Epi Info 6 Home   drag-and-drop it onto your desktop.   Use this to start Epi-Info. 


LEARNING TO USE EPI-INFO an00014_.wmf (5936 bytes)

Consult the yellow pages in the course notes.  This is a simple tutorial that gives you exactly what you need to get started in questionnaire construction, data entry, data recoding, analysis and production of printable output.  The following section includes all the common questions.  Please read through this material before shouting Help!   



The following notes and tips relate to Epi-Info version 6.04b (not release 2000) and only address the type of survey work that is used in Epidemiology.  

For more advanced use, additional assistance can be found using the excellent on-line hypertext "help" facility within Epi-Info itself, and the support unit at the Centers for Disease Control Epi-Info FAQ website       

 Info FAQ for the Epi v6.04 version ....(get the an00014_.wmf (5936 bytes)bugs out!)  

Q1 an00014_.wmf (5936 bytes)ADo I need to go through the ENTER step each time I use the program?@ Not if you are just adding more data or analysing it.  However, if you generate a new instrument (questionnaire) or adapt an old one, then Epi-Info needs to establish a new .REC file, and it does that in the ENTER step.
Q2 AOn the opening screen of the ENTER step, it tells me that the default drive is C but doesn=t accept changes.@ Don=t worry - just hit the enter key when you come to that line, and all will be well.                                                                                                an00014_.wmf (5936 bytes)an00014_.wmf (5936 bytes)
Q3 AI made the questionnaire last time, but can=t find it this time when I want to begin entering data....@ The common error here is to forget to add the drive or path every time you save or name a file. On the University=s system you only work with the A:  drive, but at home you could be creating and saving the files to the C: drive.   If you don=t specify the drive, it will be saved somewhere within EpiInfo directory on your C: drive.   Personally I always save to the A: drive.  - It's easy to find each time, and it=s portable!

Q4 AI redesigned the questionnaire in Eped but it doesn=t save it properly; it seems to confuse the old and the new versions.@   Save the new version to a different file name NOT by using the F9 key, but instead pressing F2 and selecting save file to.   Give it a new name with a .QES suffix, and from then on only use that version. You will need to go through the ENTER program again of course just once while the program sets up the new .REC file with the new name:

Q5 AI saved the questionnaire in Eped and started to enter data, but forgot to add the field corresponding to a question. How can I add another variable or field at this time?@     The short answer is that once you have started to enter data, you cannot alter the questionnaire (or data-entry screen) unless you are prepared to re-enter the data already entered.  This is not usually a major obstacle unless this means re-entering many hundreds of observations. You will need to pull up the old questionnaire using F2, alter it, and then file again using F2 and save to.
Q6 AI initially wrote the age question as a continuous variable and now realize that it is not appropriate for a TABLES command. What can I do?@   Epi Info will create a new variable for you with AGE shown as age-groups. Follow the instructions in the tutorial yellow pages. See page 61 for adding/changing variables, and page 62 for specific steps in recoding a variable in this way.
Q7 AI used the command ROUTE and gave it a   A:filename but it didn=t work in the wordprocessor@  Try using the .txt extension on whatever filename you are using. The .txt extension is recognized as a generic wordprocessing format (actually ASCII code).   When you get into your w/processor to look for the file, make sure it is searching the right drive or directory, and that it is searching for either Aall filetypes@ or A.txt@
Q8 AI entered everything, went through the ANALYSIS and produced all these tables, then routed to a file name - but when I got into my wordprocessor, there was nothing there - an empty file!@    This is a common problem. You should ROUTE to the fileneme BEFORE you carry out the analysis that you wish to save. Think of the route command as a railroad switch. Only trains travelling through AFTER the switch is made will end up where you want them to go. So- try the tables until you are producing what you want, THEN route to the file before you do the tables again (Remember the up-down arrows will remember many of the last commands, so you may not have to enter all the tables commands again.  The default ROUTE state is "SCREEN",  You can change it to "PRINTER" or to a specific (filename).txt
Q9 Everything went in to the questionnaire as planned but the variable names I gave were not used! Longer, awkward variable names appeared at the top of each column!@    What has happened here is that either you did not specify the variable names using the curly brackets { }, OR the variable name that you gave inside the curly brackets ended up on a different line to the data entry field for that question.    Make sure that after all the formatting of the questionnaire, the {variable name} and the     <Y>   <A>   or   ##     etc. are on the same line.                   an00014_.wmf (5936 bytes)
Q10 AWhy can=t I construct the questionnaire in my wordprocessor?@ You probably should create the questionnaire that you send out to the respondents using your w/processor.   But you would need to create another data-entry screen from it in any case, because the data-entry fields have to be in place to record the data.
Q11    an00014_.wmf (5936 bytes) AWhen I route the output to a file and look at it using the wordprocessor, the tables are all misaligned and cannot be read.@    That is because the wordprocessor is probably using a proportionally-spaced font (e.g. the Ai@ takes only a fraction of the space taken by a Aw@).  What you need for tables is a font that has every letter the same width.  So, when you first get into the word processor, and before you pull up the .txt file saved from your EpiInfo session, immediately place the cursor at the top left corner and go to the font-change option. Change the font from the default to any Acharacters-per-inch@ or "CPI" font such as Courier or another old typewriter font Elite for the tables. The tables will appear straight and aligned. You can change back to your favourite proportional font for the summary if you wish.   
Q12 "The percentages did not appear on the tables that were saved to disk even though I used the command SET PERCENTS=ON in the analysis."  This is probably because you  "turned on" the percentages AFTER you generated the tables.   Everything you generate AFTER using the SET PERCENTS=ON command will have percentages included until you turn them off again.
Q13 "Which percentages am I supposed to keep, and which are to be discarded?"   In general, the percentages should ADD along each level of the INPUT variable.  The input variable is usually the demographic variable (age, gender, income, etc...).  So the percentages should add for each of those levels.  Look carefully at the yellow pages and see the progression from a raw data table that should not be shown to anyone without a stress-counsellor present, all the way to a properly-edited table suitable for displaying to people with heart conditions.                                                                                            an00014_.wmf (5936 bytes)
Q14 "I cannot tell with some variables whether they are input or outcome variables!"   That's right!  Many variables can take on either role; it all depends how you use them.  For instance, take EDUCATION v KNOWLEDGE.  Here, education is the input variable, and knowledge the outcome.  Thus it makes sense to say in the results that " educated people knew most and high-school dropouts knew less... etc." .    But further on, you may want to find whether those who knew more really acted more safely.  So in the cross-tab:  KNOWLEDGE v SAFE-BEHAVIOUR, the knowledge becomes the "input" variable!   In most cases the input variable has happened chronologically before the outcome, and it has the characteristic that IF these two variables are related, then the input variable would influence the output.
Q15 an00014_.wmf (5936 bytes)"Why are there no 100% totals shown in the margins of the tables?"     This is because unlike the "counts" or "frequencies", which of course add in BOTH directions, the percentages can only add in ONE direction, - and EpiInfo does not know which way you want to do it.  So it gives you the cell percentages, and leaves you to add the appropriate marginal percentages (including the 100%) as part of your editing.
Q16  "Which of the statistics am I supposed to keep and use?".   Depends upon your task.  With most survey results, the most valuable analysis is a set of carefully-chosen cross-tables ("cross-tabs"), so you can safely erase the statistics during the editing stage, or, even switch them off during the EpiINFO analysis using the command SET STATISTICS=OFF.  an00014_.wmf (5936 bytes)

Sometimes you may want to retain a note of the odds ratio to include in the summary you will write - but to help you explain the table, rather than just to quote it to people who may have no idea what it means.

  Return to ENH522 outline page                       an00014_.wmf (5936 bytes)
  Return to Proteus home page