New Technologies for Survey Research II
Survey | Data | Exercise | Evaluation
Creation date: 11/13/99
Author: Karl Ho
This workshop is to focus post-processing and administering of the data collected from web survey. We also introduce the TELEform system that enhances questionnaire design, automated data entry and web surveys using PDF forms.
Topics include:
I. Web survey using Frontpage
This Survey Example collects data about the transportation choices and dessert preferances. Frontpage forms collect the web survey in a form as follows:
"gender" "age" "class" "trans_d" "trans_b" "trans_t" "dessert" "name" "Date" "Time" "Remote Name"
"2" "3" "5" "" "" "1" "4" "Mary Poe" "16 Nov 1999" "12:25:50" "rss0p133.acs.unt.edu"
"1" "3" "6" "1" "" "1" "2" "Karl Ho" "16 Nov 1999" "12:26:46" "rss0p133.acs.unt.edu"
"1" "5" "6" "1" "" "" "5" "Steven Hawkins" "16 Nov 1999" "12:27:31" "rss0p133.acs.unt.edu"
"2" "6" "6" "" "" "1" "1" "Marilyn Monroe" "16 Nov 1999" "12:27:56" "rss0p133.acs.unt.edu"
"1" "6" "6" "1" "1" "1" "2" "George Bush" "16 Nov 1999" "12:28:30" "rss0p133.acs.unt.edu"
"2" "3" "4" "" "1" "" "2" "Jane Doe" "16 Nov 1999" "12:29:06" "rss0p133.acs.unt.edu"
"1" "5" "1" "1" "1" "1" "4" "Henry Ford" "16 Nov 1999" "12:29:24" "rss0p133.acs.unt.edu"
"2" "4" "5" "1" "" "" "3" "Helen Hunt" "16 Nov 1999" "12:29:43" "rss0p133.acs.unt.edu"
"1" "2" "2" "1" "" "1" "5" "Peter Smith" "16 Nov 1999" "12:30:25" "rss0p133.acs.unt.edu"
"2" "6" "6" "" "" "1" "4" "" "16 Nov 1999" "12:31:07" "rss0p133.acs.unt.edu"
The data have the first line as the field names. It is recommendable to apply the eight-character rule, which saves us from renaming or letting SPSS rename the variable names. All data are wrapped by double quotes and delimited by tabs, as was specified. We may use any editor to remove the unneeded double quotes but leave the tabs as follows:
gender age class trans_d trans_b trans_t dessert name Date Time Remote Name
2 3 5 1 4 Mary Poe 16 Nov 1999 12:25:50 rss0p133.acs.unt.edu
1 3 6 1 1 2 Karl Ho 16 Nov 1999 12:26:46 rss0p133.acs.unt.edu
1 5 6 1 5 Steven Hawkins 16 Nov 1999 12:27:31 rss0p133.acs.unt.edu
2 6 6 1 1 Marilyn Monroe 16 Nov 1999 12:27:56 rss0p133.acs.unt.edu
1 6 6 1 1 1 2 George Bush 16 Nov 1999 12:28:30 rss0p133.acs.unt.edu
2 3 4 1 2 Jane Doe 16 Nov 1999 12:29:06 rss0p133.acs.unt.edu
1 5 1 1 1 1 4 Henry Ford 16 Nov 1999 12:29:24 rss0p133.acs.unt.edu
2 4 5 1 3 Helen Hunt 16 Nov 1999 12:29:43 rss0p133.acs.unt.edu
1 2 2 1 1 5 Peter Smith 16 Nov 1999 12:30:25 rss0p133.acs.unt.edu
2 6 6 1 4 16 Nov 1999 12:31:07 rss0p133.acs.unt.edu
Using SPSS Text Import Wizard, we can read in the data and save an SPSS program or the "predefined format" to read in the data next time. Web survey is distinguished from regular questionnaire survey that researchers can get updated data as often as they want. You can read in the data every two weeks, every week or even every day! However, it is highly recommended that you keep a log on data management and record different waves of the dataset. It will take up some disk space for saving different versions of the data set but it will safeguard your data in case the web survey is hacked or corrupt or one of your data set gets contaminated.
2. Importing Web Survey Data into SPSS
We focus on SPSS to read in web data. The process is very similar in SAS and SPSS data can be easily converted in to SAS formats. It is recommended using SPSS 8.0 onwards (the newest version is 10.0). The Text Import Wizard in SPSS (8.0 onwards) facilitates importing raw, ASCII text data. Certain procedures, however, are warranted:
1. Export the result file (text format) to local disk.
2. Remove the double quotes and save the file under a different file name.
Convention using numbers and characters to represent different versions (e.g.
survey01a.txt) is advised.
3. Read in the data using SPSS Text Import Wizard (In SPSS Data Editor, File -- >
Read Text Data)
In Step 2, make sure to specify Delimited option and read in the variable names at the top of the file.
In Step 4, make sure only the tab option is chosen. IF you select other or space, SPSS will be confused and skip the multiple-response answers.
In Step 6, which is the final and most important step, save the format for future use and CHECK THE RADIO BUTTON TO PASTE THE SYNTAX. It is not recommended you skip the syntax and just run the import procedure. It will end up you have to do it every time with different setup. To standardize the import procedure, it is highly recommended to import the data using Syntax or an SPSS program. You can also edit the program to apply other procedures like labels, renaming variables, etc.
The following is the program SPSS creates using the Text Import Wizard:
GET DATA /TYPE = TXT
/FILE = 'C:\ACS\Acs-class\survey1\form_result.txt'
/DELCASE = LINE
/DELIMITERS = "\t"
/ARRANGEMENT = DELIMITED
/FIRSTCASE = 2
/IMPORTCASE = ALL
/VARIABLES =
gender F1.0
age F1.0
class F1.0
trans_d F1.0
trans_b F1.0
trans_t F1.0
dessert F1.0
name A14
Date A11
Time A8
V10 A20
.
Note that since the Frontpage generated name for the "Remote Name" field does not follow the eight-character rule, SPSS converts it automatically into V10. You can apply variable and/or value labels to the program:
Var label gender "Gender" age "Age" Class
"Classification" name "Respondent's name (Optional)".
value label class 1 "Freshman" 2 "Sophomore" 3 "Junior" 4
"Senior" 5 "Graduate" 6 "Faculty or Staff".
value label dessert 1 "Chocolate mousse" 2 "New York style cheese
cake" 3 "Napoleơn" 4 "Brownies" 5 "Lime Sherbet".
var label trans_d "Driving" trans_b "Bus" trans_t "Taxi".
Since the data by default are in string format, some variables such as the numeric values will be converted as numbers. For data and time variables, however, they will be treated as string even though they appear as numbers. Double click the tab and change the type into date and time formats respectively. Note that they should be scale measurement. It will be useful to plot the time and date of the responses:
Frequencies
Multiple Responses
Group $TRANSP Transportation
(Value tabulated = 1)
Pct of Pct of
Dichotomy label Name Count Responses Cases
Driving TRANS_D 6 37.5 60.0
Bus TRANS_B 3 18.8 30.0
Taxi TRANS_T 7 43.8 70.0
------- ----- -----
Total responses 16 100.0 160.0
0 missing cases; 10 valid cases
Pie Chart
Examples:
Static report
Dynamic charts using Java (point your cursor to the bars and get instant data)
Dynamic
charts using ActiveX (You will need to use MS Explorer to see this link.)
The TELEform system is composed of a data capture server, a questionnaire designer, data verifiers and a PDF module. The data capture server is responsible for capturing the data via scannable questionnaires and PDF files on the web that feeds data directly to the database on the TELEform data capture server. The Designer is a point-and-click application that helps set up a questionnaire with features of automated forms processing and document capture. Designer includes layout tools for creating new forms and rending existing forms automated. The scanning system, TELEform Reader, is equipped with recognition engines that identify handwriting, barcodes, and check boxes.
1. Select Features:
System Administration & Control¡@
Last updated: 01/18/06 by Karl Ho