data capture
Opscans may be used to capture data from surveys and other data gathering efforts.
The information is translated to an electronic file with one line in
the output per opscan. Only one mark per item is accepted. Multiple marks for
any item produce an asterisk for that item in the data file. Omitted
items produce blanks.
You may request your data in either of two forms:
.csv File (Excel-ready, comma-separated values format)
Raw Data File (Fixed columns)
.csv File
This file will open in Excel. You may choose which areas of the opscan
you wish in your output file -- just let us know. In contrast to a Raw
Data File (see below), the form, group number, and all items are exactly
as coded on the opscan.
For example, an Orange form marked with an ID # of 123456789 and item 1 marked as A and item 2 marked as C would create a line in the file as follows: 123456789, 1, 3,
Raw Data File
Data is placed in the columns shown below,
depending on which form is used.
|
Form
|
ID
|
Form
|
Seat #
|
Group
|
Data
|
Identifiers
|
|
Orange
|
1-9
|
10*
|
11-13
|
14*
|
15-74*
|
|
|
Purple
|
1-9
|
10*
|
11-13
|
14*
|
15-174*
|
|
Green
|
1-9
|
10*
|
11-13
|
14*
|
15-62*
|
|
Gray
|
|
1-60*
|
61-63
|
*Although the forms offer choices marked 1-10,
the data file created during scanning translates the responses to 0-9
in order to allow one column per response. Form and group fields are
also "bumped down". You will need to "bump up" these columns before
running an analysis. This is easy to do in Excel, SAS, SPSS, etc.
There are no delimiters in the file. All items are one column, fixed width.
Coding In Numbers
If you need to code in a number, such as 4907, use the ID
area on the orange, purple or green sheets if you can.
Otherwise you will need to use 4 items to code this number
and you have the problem of how to code in a "0". If necessary, we
generally tell people to use the "10" choice for the zero.
After "bumping up" each column and changing all "10"s to
"0"s you can put the number together. (Do not try to read
the field as a 4 column variable and then add 1. If you add
1 to the "bumped down" number of 3896 you will not get 4907.
Adding 1111 will also produce incorrect results.)
Analysis of Results
On request, Test Scoring will provide frequencies and means. More in-depth analysis is generally not available.
For Item Means, please provide a list of the
survey questions used to generate the data so
that choices that shouldn't be used in averaging can be eliminated, e.g., choice 4 in the following question:
How did you like the
movie?
- I liked it
- It was okay
- I didn't like it
- I didn't see it
Questionaire Design
For more information on creating good questionnaires, see A Brief Guide to Questionnaire Development by Dr. Robert Frary.
Learning Technologies also offers information on evaluation models, methods, and tools at:
http://www.edtech.vt.edu/edtech/id/eval/eval.html
Page last updated
August 7, 2008
.