- EpiData-list - University of Manitoba Mailing Lists

Error in reading csv file
by epidata-list＠lists.umanitoba.ca 16 Nov '06

16 Nov '06

Keith had an error on reading csv file: "I am trying to read a .csv file using EpiData Analysis. The .csv file has 128 fields in it. The first row contains column headings (8 Chars Long). None of the fields contain data over 64 chars long." Please submit the file in a zip archive to the system mailbox at info at epidata.dk and we will use the file for further testing. Regards Jens Lauritsen EpiData Assocation

1 0

Question
by epidata-list＠lists.umanitoba.ca 16 Nov '06

16 Nov '06

EPIDATA ANALYSIS 1.1 (Build 68) This is the first time I have attempted to use this product. Any insight into the cause of this problem is much appreciated! I am trying to read a .csv file using EpiData Analysis. The .csv file has 128 fields in it. The first row contains column headings (8 Chars Long). None of the fields contain data over 64 chars long. When I click "Read Data" and select the .csv file to read, I get this error message... Data in field Hospital, record 206 exceed maximum length Access violation at address 00732A30 in module 'EpiDataStat.exe'. Read of address 0000000C Operation aborted The data in field Hospital, record 206 is only 36 characters. I have found that if I truncate that field to 18 characters, then EpiData Analysis will read that record successfully. I get the same error on other fields within my .csv file and again, if I truncate the data within those fields to 18 characters, then EpiData Analysis will read the entire file successfully. Thank you in advance for any assistance you can provide! Keith _____________________ Keith Higginbotham Computer Systems Center Alabama Dept of Public Health 334-206-7040

1 0

Error in export to Stata v8 from EpiC june 2006 (v3.1.1)
by epidata-list＠lists.umanitoba.ca 14 Nov '06

14 Nov '06

An error has been found in writing Stata files with EpiC from June 2006, v.3.1.1 It was the version on the testing page. An updated version of EpiC 3.1.2 has been placed on the testing page (see link at top of www.epidata.dk/download.php). Regards Jens Lauritsen EpiData Association

1 0

Double-entry problem
by epidata-list＠lists.umanitoba.ca 09 Nov '06

09 Nov '06

Hello all - I am collecting and entering data for a project with well over 700 variables. Of course, we are double entering roughly 20% of our data records to check for consistency and mistakes. I created a .dbl file from the .rec file to prepare for double entry as instructed in the EpiData manual. But it seems like there is a next step that I am missing. We've double-entered a few records and corrected a few mistakes. I was under the impression that the fields we corrected during double-entry would overwrite those fields and save to the original record. However, when I opened the records that had been double-entered and corrected the original responses were still there. All of the corrections we made were not there. What my next step in EpiData to accept the changes that were made during double entry for each of those records? Thanks a lot! Sarah McDougal Department of Aging & Mental Health 13301 Bruce B. Downs Blvd, MHC 1400 Tampa, FL 33612-3807 Phone: (813) 974-1975 Fax: (813) 974-1968

1 1

Which estimates to have as default in OUTBREAK ANALYSIS ?
by epidata-list＠lists.umanitoba.ca 07 Nov '06

07 Nov '06

> I prefer B of the two options. Add the /o option to provide a separate > table of odds ratios, or replace the RR with OR (I prefer the > replacement, since the user should know what the study design is). Add > the /ci option to provide CI for the attack rates (not something I > routinely look for). Add the /t option for chi-squared and p, Fisher > exact p. > > Jamie I agree option B is better. I wouldn't mix OR + Attack Rate, so if /o is added it should replace RR with OR and eliminate AR. /AF, can be useful to add Attributable Fraction (both in exposed and in population) Greetings, Pedro

1 0

Which estimates to have as default in OUTBREAK ANALYSIS ?
by epidata-list＠lists.umanitoba.ca 07 Nov '06

07 Nov '06

In version 2.0 a new /OA (Outbreak analysis) parameter will be added to analysis. The question is what should be shown as default (other aspects will come as further options): e.g.: . tab ill vanilla milk coffee /oa Would make a table for ill by all other variables, and then show what: A: N total number of observations for exposed and unexposed: Ill number of persons with outcome Attack Rate + CI (percent with outcome) Odds Ratios + CI B: N total number of observations for exposed and unexposed: Ill number of persons with outcome Attack Rate (percent with outcome) RR + CI Please discuss default and desired options Regards Jens Lauritsen EpiData Association

1 2

Help: copying graphs/tables from Epidata into word
by epidata-list＠lists.umanitoba.ca 25 Oct '06

25 Oct '06

hello I am sure this has a simple answr but its quite annoying. I cannot copy my tables from epidata into word How do i do this thanks Bankole Olatosi PhD(c), MS, MPHResearch Associate, South Carolina Rural Health Research Center,Columbia, SC Tel: 803-251-6317Fax: 803-251-6399 PhD Candidate Health Services Policy and Management,Norman Arnold School of Public Health,University of South Carolina _________________________________________________________________ Stay in touch with old friends and meet new ones with Windows Live Spaces http://spaces.live.com/signup.aspx

1 2

Mode of copying from output window
by epidata-list＠lists.umanitoba.ca 24 Oct '06

24 Oct '06

Help: copying graphs/tables from Epidata into word For graphs it is quite easy: After creating a graph a copy is always placed on the clipboard, so: 1. create a graph - and make sure it is visible in the output window (close edit graph if you requested edit). 2. start wordprocessor, drawing programme, presentation programme etc. 3. press ctrl+v - and the graph is copied into the current open programme. If your pc is not too old you can have EpiData Analysis and the word processor open at the same time and switch btw. them by pressing alt+tab keys. Regarding tables: Is the development now such that default copying (ctrl+c) should be html formatted rather than text formatted ? Now: Shift+Ctrl+C: html formatted. Ctrl+C: tab delimited copying. We are experimenting with a "mouse over" effect, such that if the mouse is over a table or graph in output, then it would be nice that "right-click" would enable "copy table" or "copy graph". regards Jens lauritsen EpiData Association

1 0

Clarification - document on options
by epidata-list＠lists.umanitoba.ca 23 Oct '06

23 Oct '06

Shavinder Singh has made some comments which are welcomed, but also necessitates some clarifications. The document put on the testing.php page is NOT a list of commands. It is a somewhat internal technical document created with the intent to write an overview of which options will be in which commands. The reference to the commands is the same as users from version 1.0 of EpiData Analysis. Regarding the comments on dates I will only add that maybe we should change the name of date functions from to year(datevariable) yearval(datevariable) day(datevariable) dayval(datevariable) month(datevariable) monthval(datevariable) since now there is a conflict with variable names date, year and month, which users often will use. Also we should implement some sort of warning to users when they use for a variable name an internal name. E.g. neither of "or and if" will ever work as a variable name Regards Jens Lauritsen EpiData Association

1 0

Feedback about the document 2.1 about unification
by epidata-list＠lists.umanitoba.ca 23 Oct '06

23 Oct '06

Dear Jens, Some feedback about the document 2.1 about unification and. Here is some feed back from user of EPI6 and Epidata and Epi analysis. There is lots of good function for fine-tuning of the output from analysis. The commonly used commands like FREQ, TABLES, and MEANS would be able to produce tailored output as per the needs. Few suggestions; 1. FREQ v1 produces the frequency table only. The summary measures such as observations, mean, SD, SE, minimum, maximum, 25th 50th 75th percentile are not shown. The provision of these details is helpful in examining a quantitative variable in detail. EPI6 had the facility of shutting the frequency table (esp. if it was too big) and print only the summary measures. The syntax was, FREQ v1 /n Kindly explore the possibility of this switch with FREQ command and the following switch also is so handy (EPI6). Another function was frequency weighted counts on variable FREQ v1 /SUM=v2, where v1 is a categorical and v2 is numerical variable. Is tab v1 /w=wvar same command or different, is not clear ? 2. Describe command gives the summary measures to some extent. This command has not been activated in the ver 1.1 of Epi Analysis. 3. MEANS command needs thorough overall. It requires the conversions of both the variables as numerical. The categorical variables are not numerical variables. Hence this should be changed accordingly. The (/q or quite) quite switch closes the output totally. Actually this should omit the display of tables only. The summary measures should be displayed as was the case with EPI6 command MEANS v1 v2 /n 4. Dates remain a sour issue with EPI Analysis, though there is lot of help available from the Epidata team members especially Jamie. Is it possible to use SET option to define the date command such as SET DATE FORMAT =mm/dd/yyyy or SET DATE format=dd/mm/yyyy We use most of the time only one type of date format. Then date or date range selection would become simple like simply typing the date variable. mm/dd/yyyy (If this does not make sense from programmers point of view, kindly forgive me) 5. The help manual As EPI ANALYSIS evolves in future, a well-written command reference becomes a necessity esp. keeping in mind the beginners. For example, SET options are clear to experienced user but not to the beginners. Scatter graphs with regression line drawn or an option to draw it like /r 6. Many commands listed in the command reference 1.1 are not implemented yet. The EPI ANALYSIS would be a wonderful product is most of these options are actually implemented. Keep up this good work. With regards, Dr. Shavinder Singh

1 0