Dear Niamh:
I started with SAS in 1985, then went to SYSTAT, then briefly to SPSS, but unfortunately never learned STATA, but this is what I learned over the past ten years: I have no need whatsoever for any of the proprietary analysis packages at the low level of my sophistication. Whenever I felt is was the fad of the day to do some fancy logistic regression, I always got hung up in the end that the dataset was so full of heterogeneity that made it entirely inappropriate, and that doing more sophisticated stuff ended up in an incomprehensible black box that nobody could check. Since EpiData Analysis became available, I have used only that and we have managed to publish papers with intermediate sophistication (like stratified survival analysis and the like) in high impact journals (IF>10) and it was always transparent to everybody what we did.
As for the process itself, whether you use any of the proprietary software packages or not, everybody needs a quality-assured dataset, and none of the proprietary software packages offers that, they are for analysis, not for data capture. Indeed, as Jens could further elaborate, the initial intent for the development of EpiData software was precisely to have software that would offer the possibility of obtaining validated, quality-assured data commensurate with good clinical practice, then have good export functions, so that one would have a format that would be readable for import into the analysis of choice software, and that is precisely what EpiData Entry does better than any other software, and above all it is free. We then asked the EpiData Association to please develop an Analysis module as well that would cover more than 90% of the needs of researchers (but would abstain from offering complex regression analysis) because the sorry state of affairs was that researchers from low-income countries would get their Master's of PhD degree in some industrialized country, were taught there to use SPSS, SAS, or STATA - and and some universities I know of - even all of them. Then they would return back home, the student-loaned software license would expire, and they had no other avenue left than pirating a copy of the only thing they knew. Now, forcing students to ultimately steal software cannot possibly be the solution to the problem. What I have seen in my career as a teacher is that the vast majority of students who play around with outrageously expensive proprietary software are actually doing stuff at a very, very low level of sophistication, and I ask myself what for one needs a one-thousand-dollar package to do a stratified analysis if one can do that with user-friendly, entirely transparent and most importantly free EpiData Analysis?
Apologies for the lengthy advertisement for EpiData software.
Good luck with your research,
Hans
On 20:59, epidata-list@lists.umanitoba.ca wrote:
Hi all,
I have a quick question. I used Epidata to put all my data from a survey I conducted. I did not export all my inputted data in SPSS but I am using SPSS for analysis. I was wondering is it possible to export all of my data in Epidata into SPSS. Or is it better to manually put it into SPSS. Also, I am new to research and statistics and was advised to use Epidata first and then analyse using SPSS. I was wondering why people use Epidata first and then export to SPSS. I mean, what are the advantages of doing this? Is it better to do this first and then export to SPSS?
Any advise would be greatly appreciated,
Kind Regards,
Niamh