Direct Entry vs Transcribing Paper

Direct Entry vs Transcribing Paper

Post by Sarr J. Bl » Tue, 09 May 2006 23:23:47

A project I work on does data collection and management for clinical
trials. We use a web based data entry system, but most of our data is
still written on paper forms and then transcribed. We do the
transcribing twice to improve the accuracy of that step.

Our funding sources are starting to push for direct entry. This
eliminates the time, cost and errors from the transcription. But does
anyone have or know of any data (opinions are welcome, too, but please
try to distinguish the two :-)) on whether data entered on line is
more reliable, less reliable or about the same as data entered on
paper. Clearly the handwriting problem goes away, but other errors are
intraduced. I mean introduced. :-)

A second order question is how any of this is affected by the kind of
machine being used: desktop vs laptop vs PDA vs...


Sarr Blumson XXXX@XXXXX.COM ~sarr/

Direct Entry vs Transcribing Paper

Post by Davi » Wed, 10 May 2006 08:55:10

ello Sarr,

On Mon, 8 May 2006 14:23:47 UTC, "Sarr J. Blumson"
< XXXX@XXXXX.COM > wrote:

So if there is a discrepancy between the two transcriptions, what
happens to that data?

From what I've heard (mostly opions) is that organizations that push
their data entry systems closer to the source get roughtly the same
data entry errors. That is, a "meter reader" may still fat finger
an entry. The system may have knowledge about the data coming in
and try to spot obviously or reasonably wrong data. Credit Card
Numbers include check digits to improve reliability, but if your
system doesn't already have that there isn't any improvement on
the quality of the data.

The biggest advantages to pushing data entry closer to their source
(some observed projects, as well as opinions leading to the decision)
is that "downstream information handlers in your organization" can
access the (now) online information and use it to aid their group's
knowledge. Thus if you have many tiers in your organization that
help gather information and process it, pushing the data online earlier
makes the data readily available to all groups for their various

There is a downside to gathering the information as it may be more
costly to gather the data or the new groups responsible for the data
may not be used to the technology. However, the projects I've worked
with were able to show later on that the cost of gathering the data
earlier was a benefit to the overall organization and usually to the
customer as well.

A couple good examples of this are credit card processing and
shipping companies (e.g. FedEx, UPS, USPS, etc). A mostly paper
trail and manual collection process with a heirachical design
was replaced by roughly the same electronic equivalent. Now
these organizations can look at their data in other ways. Fraud
detection becomes easier. Various groups can plan more efficiently
to meet customer trends. Customers may be given access to their
account data as well as the current transaction history (or the
package location for shipping customers).

I've worked for one company that chose to design a handheld unit
specifically for their needs. This was deemed more useful than
the alternative technologies. The keypads were organized specifically
for the task at hand. Have you ever noticed that the keys on a
fast food restraunt machine doesn't use the QWERTY keyboard layout.
Another company I'm familiar with supported everything from desktop
down to PDA and all the variations of very simple cell phones. You
design the entry process a bit differently when you are entering
alphanumberic data and only have a numeric phone pad. Mistakes are
different, but sometimes recognizable and recoverable. Certainly
if your proposed data entry person doesn't want to do the entry
on the type of machine that they are given, there could be some

Also consider the fallback position in your design. What happens
when parts of the electronic network aren't working? Does the
whole process still work to some degree and not shut everyone
out? Collection points can be a better design than the single
web server. Ever notice what happens at the airport when the
airline computers get even a second or two slower? When it
breaks everyone knows it. A good design trys to anticipate
problems at all levels of the system.

You're welcome,



Direct Entry vs Transcribing Paper

Post by Thad Smit » Mon, 15 May 2006 02:18:27

Accuracy is one issue. Another one, in some cases, if flexibility. I
can imagine that exceptional cases may arise in which some writes an
important note on the paper form that explains an anomaly. Would the
direct entry be as accommodating? What would the user need to do to
enter notes? Even if there is a provision, it is likely to be more
awkward to do. Can a normally required field be omitted if there is no
data for some reason?

Another consideration is that direct entry may be slower than filling
out a paper form -- its easy to do one-handed with very little working
space. In general, it's faster to fill out the appropriate section of a
paper form that navigate a computer entry form. Of course, data must be
later keyed from the paper form, but if it is important to minimize time
during data collection, paper may win.


Direct Entry vs Transcribing Paper

Post by bks » Mon, 15 May 2006 06:49:50

In article <rOdGr40LMPU3-pn2-Wrdgur279qc7@localhost>,

It's still the UPS guy who does the entry, not the customer.