Reload full frameset of the WebQ Page

WebQ

Outline

WebQ is a JavaScript application for Q-sorting questionnaire items online. Q-sorting can be characterized as a process whereby a subject models his or her point of view by rank-ordering items into 'piles' along a continuum defined by a certain instruction (McKeown & Thomas, 1988, p. 30). A basic principle of the Q-sort technique consists in that items are evaluated relative to each other. This is usually accomplished by providing the items on cards which the subject lays out and sorts into horizontally ordered category piles on a desk. This layout cannot easily be implemented on a computer screen.

In WebQ, therefore, items are displayed in questionnaire format with radio buttons alongside each item for choosing ranking categories. Everytime the user clicks the Update-function button, statements will be reordered on screen into their momentarily selected category 'piles' ('piles' of items are placed beneath each other).

When grouping and re-grouping of items into their final rank-ordered categories is accomplished, the user is ready to click the Send button. Upon that, the data will be pasted into a ready-to-send email window, where the user can add some additional comments, and send the email to the researcher's address.

A researcher's guide section describes how to set up WebQ for your own Q-sort project.

WebQ Demo

If you may want to learn by doing before continuing this documentation on Q-sorting with WebQ, then go to the Sample WebQ or directly to the WebQ Tutor. This will give you a quick idea of how to sort statements. Don't hesitate to also check the Send button ­­in these demo WebQ examples no information will be sent away to an existing email address.

Q-Sort Technique

The technique of Q-sorting was first introduced by William Stephenson, the creator of Q Methodology (see: Stephenson, 1953; the textbook by Brown, 1980; or the Sage Booklet primer by McKeown & Thomas, 1988). Whereas Q Methodology takes quite peculiar an epistemological stance as a privileged pathway to human subjectivity, the Q-sort ­­as well as Q-correlation and Q-factor analysis­­ are techniques that are applied within other schools of thought also: Carl Rogers (1954) who popularized the self sort with ideal sort correlation as a measure of adjustment, and Jack Block (1961) who developed the California Q-set (CQS) as a standard psychometric instrument.

A Q-sample (or Q-set) consists of a set of stimuli each printed on a separate card. Typically, the stimuli are statements expressing different opinions on a certain issue, and the number of statements is somewhere between 30 and 60, though as much as 100 statements are not quite unusual. The process of Q-sorting a set of stimuli amounts to having the subject model his or her subjective point of view on the issue at hand by rank-ordering the stimuli along a continuum defined by a condition of instruction (McKeown & Thomas, 1988, p. 30). A condition of instruction is a guide for sorting Q-sample items. This can be a simple request, like:

• Sort the items according to those with which you most agree (+5) to those with which you most disagree (-5).
• Sort the items according to those that are most like object / person X (+5) to those most unlike that object / person (-5).

Sometimes, the subject is asked to sort the same Q-sample under two or more variations on the same basic condition of instruction. For instance, the target object / person may vary, like in Roger's assessment of real self vs. ideal self descriptions. Or the ranking continuum may be applied to differing theoretical constructs. For example, in a study of political perceptions, the respondent could be asked to sort the sample items according to "what is most like / unlike a conservative point of view", and "what is most like / unlike a liberal point of view."

The usual technique involves a forced sort, i.e. putting under each point on the continuum a prescribed number of cards. The distribution of the pile sizes usually follows a modification of a (flattened) normal curve as displayed in the following example of a finished Q-sort:

Note that this example serves demonstration purposes only, and that the very small number of very simple items should not be misrepresented as in any way typical for Q-samples in general. The same example will be shown later in WebQ's layout, and it is also used for the Sample WebQ.

Strategy for rank-ordering a Q-sample, procedural steps. How was this example Q-sort (provided by an anonymous observer of Bill's character) arrived at? In addition to the well-shuffled pack of cards, the subject in a Q-sort session is provided with either a single long ruler or a set of separate distribution marker cards which would represent the ranking continuum with its pile categories (-2 through +2) under which the cards will be layed out in the prescribed distribution (1 - 2 - 3 - 2 - 1), and a detailed step-by-step instruction. The general idea of the sorting strategy is to begin with presorting items into three piles (left: disagree - middle: neutral - right: agree), then pick out the most significant representatives of both, the extreme right (+2), and for the extreme left pile (-2), and then continue to work towards the center of the ranking continuum. A prototypical set of instructions is given by McKeown & Thomas (1988, p. 31f.). However, note that the following quotation refers to a more typical Q-sort design with 11 piles (-5 through +5, with frequencies 3-4-4-7-7-10-7-7-4-4-3):

1. The subject is asked to read through the items to become familiar with them. As this is done, the subject sorts them into three piles: Placed to the right are those with which the subject agrees, to the left those with which he or she disagrees, and in the middle those about which he or she is either neutral, ambivalent or uncertain.
2. During Q-sorting the subject spreads the items out under the distribution markers, while maintaining the general left-center-right relationship. This facilitates the reading of the items contextually and the making of comparisons.
3. Studying the items to the right, and in conformity with the distribution, the subject selects the three items that are most like his or her position (or, the number of items called for) and places them vertically, under the +5 marker. The order of the items under the markers is not important; all three items beneath the +5 marker will receive the same score when the data are recorded.
4. Turning now to the left side, the subject studies the items, and selects three from among those on the left that are most unlike his or her position. These are placed under the -5 marker. Again, the specific order does not matter.
5. Returning to the right side, the subject now picks the four items that are more like his or her position than the remaining ones among the grouping but which are not as significant as the four already selected (located under +5), and places them under the +4 marker. On second thought the respondent might decide that an item selected for +4 is more importatnt than one uder +5. He or she is perfectly free to switch it with another at this or any other time.
6. Attention reverts to the left side and the process is repeated, with the subject working toward the middle 0 position, until all of the Q-sort statements are positioned from left to right. Items placed under the middle marker (0) often are the ones left over after all of the positive and negative positions have been filled. The reason for having subjects work back and forth is to help them think anew the significance of each item in relation to the others. Once completed, the Q-sort should be reviewed, the subject making adjustments among items that, upon rearrangement, more accurately portray his or her personal point of view.
7. Finally, statement scores for the completed Q-sort are recorded by writing the item numbers on a score sheet that reproduces the Q-sort distribution (cf., the figure above).

Q-Sorting with WebQ

Overview

WebQ is a computer implementation of the Q-sort technique which requires the respondent to rank-order a set of stimuli (typically, statements of opinion) according to a certain condition of instruction (e.g., on a continuum ranging from -5: least agree to +5: most agree). Initially, all items, ordered randomly, are placed within the neutral category pile, with the 0 radio buttons checked. Piles of items (initially empty except for the neutral pile) are ordered vertically with the most positive category on top, and the most negative category on bottom. Items are moved to and from piles by first changing checked radio buttons and then clicking on an 'Update button.' For each pile, the number of box icons signifies how many items are to be sorted into that pile. Different colors of the box icons show whether there are too few, too many or the correct number of items in the pile. When Q-sorting is finished, and the 'Send' button clicked, the response will be delivered to the researcher's email address.

Layout

On startup all items are presented in random order beneath the central ('0') pile bar, with all radio buttons correspondingly checked for '0'. When the Q-sort is finished, it differs from the typical table-top layout (cf. Q-Sort Example: "My Personal View of Bill" above) only in that the item piles are distributed from top to bottom instead of from right to left.

Procedure

WebQ's upper frame gives a description of the meaning of the symbols which are used to inform how many statements are required for each of the categories, and how many of these places or 'slots' are still empty, occupied, or 'overcrowded.' In the center of the upper frame the title of the Q-sort is displayed ("My Personal View of Bill"; this space may contain a short instruction text also). At the right-hand side there are three action buttons:

Help

Opens WebQ help page in a separate window.

Update

By clicking on the Update button all items are sorted into piles according to radio buttons checked.

Send

After checking the distribution of statements across categories, the Q-sort data are prepared for submitting by email to the researcher's address. The respondent is asked to provide a code word, and then the data and a specific prompt for additional comments are pasted into the body of a ready-to-send email.

This snapshot of the ranking frame shows the state immediately before the Q-sort is finished. Upon the preceding Update the items were sorted as displayed. In the meantime the respondent has changed response categories of two items (3 and 9): Therefore, clicking Update again, will move '3. Persistent' from pile +2 to pile +1, and '9. Distant' from pile 0 to pile -1 --and the Q-sort will be complete, with all boxes green.

Additional details on using WebQ can be found on the WebQ Help Page.

Researcher's Guide to WebQ

Using WebQ for collecting Q-sorts is very easy and efficient, especially if you just want to run it on your PC or for group administration in the computer lab (planning and preparing a full-blown online-survey is more demanding, of course). For the respondent, sorting cards on a table may be more fun than clicking and scrolling on the computer screen - for the researcher, however, the computer reduces work load involved in preparing and administering the Q-sorts, and in collecting and inputting data to a minimum.

Setting up a WebQ-sort is a matter of minutes, and does not require knowledge in HTML.

Quick Start

• Download webq.zip and unzip files to a new directory (e.g., c:\webq).
• Prepare your Q-sample (the set of statements) and its design: Number of statements; number of categories (e.g., 9 categories: from -4 thru +4); number of statements for each category (e.g., 2-3-3-4-6-4-3-3-2, would be one possible distribution for 30 statements), and the definition of extremest categories (e.g., 'Most agree' and 'Most disagree').
• There are two files to edit. Under Windows, use the program notepad - do not use an HTML-Editor for that!
  1. samplejs.htm - this is where the details of your Q-study are defined (described below). Save this file with a different name, e.g., study1js.htm.
  2. samplewq.htm - this is the start-file ('frame-set'), in which you only have to specify the name of the html-file with the study definitions: In the line <FRAME SRC="samplejs.htm" name="controlframe"> replace samplejs.htm with that file name (e.g., study1js.htm). Save this file also with a different name, e.g., study1wq.htm. When everything is finished, open this file with the browser to run your WebQ-sort.

When you edit the samplejs.htm you'll find comments (preceded by '//') that inform you what to do:

// Edit the following lines for your purposes

//*** study_title will be displayed at the center of the top frame
//  (may contain also short instruction. Use <br> for line break)
study_title = 'Sample WebQ-Sort<br>My Personal View of Bill';

//*** address that will accept results - use 'Nobody@localhost' 
//    for testing
addressee = 'Nobody@localhost';  

//*** email subject line:
subject = 'Q-Sort data for Sample WebQ'; 

//*** URL of the 'help page' and name of button for 'help page';
help_url="webqhelp.htm";      
help_button="Help";

forcedchoice = false ; //set 'false' to permit sending sorts that
                      //do not match the prescribed response distribution
                      //(usually set 'true')

randomsort = true;  //set 'false' if statements are to be ordered
                    //in original sequence (usually set 'true')

//*** commentprompt will be pasted at bottom of the email-message:
//   use '\r\n' for line break
commentprompt = "You can add comments here: -- \r\n\r\nmore comments here:   " 

//*** Name of the extremist categories:
rightanchor = 'Most like me';
leftanchor = 'Least like me';

//*** No. of statements:
numstatements = 9;

//*** No. of categories:
categories = 5;

cat = new Array(categories); //don't change this line!!

//*** Distribution of the categories. If you have more categories add it.
//    (and adjust the 'categories=' above)
cat[1] = 1;
cat[2] = 2;
cat[3] = 3;
cat[4] = 2;
cat[5] = 1;

//*** Your statements. You can adjust it to the no. of your statements.
statement = new Array(numstatements); //don't change this line
statement[1] =  "Boastless";
statement[2] =  "Unsparkling";
statement[3] =  "Persistent";
statement[4] =  "Coldhearted";
statement[5] =  "Timid";
statement[6] =  "Tricky";
statement[7] =  "Charitable";
statement[8] =  "Jovial";
statement[9] =  "Distant";

//study definitions part ends here (do not delete next line)

For instance, in the line starting with 'addressee =', replace nobody@localhost with your own email address. Note: In this sample file, the variable forcedchoice is not set to true which would be recommended in general - you should try both versions to see how that determines what happens when clicking Send before the Q-sort is finished alright.

Submitting and Receiving the Q-Sort Data

After clicking the Send button and providing a code word (PeterS in this case), the respondent's email client (here: Netscape Communicator) pops up with:

(Note that all user-defined email options like signature, and Bcc are in effect as usual).

The respondent is invited to add comments beneath the prompt lines that are defined by the variable commentprompt (cf., samplejs.htm above) before sending this email off. When the researcher receives this email (provided that the variable addressee has correctly been set) it looks quite the same, of course.

The Q-sort data are enclosed in "> <", and are formatted as follows: cols. 1-8 the code word (or 'SortId'), and beginning at col. 11, the checked response categories for all statements ordered by statement number.

Importing WebQ-Sort Data in PQMethod Software

For creating a data file (e.g., study1.dat) into which to cut & paste WebQ-sort data records, run PQMethod, specify a 'Project Name' (e.g., study1), and select QENTER where you are asked to provide study title and the design specs of your study.

After closing QENTER and PQMethod, open the data file (study1.dat) with a text editor. Again, under Windows, the best choice is notepad. The two header lines required would look as follows:

  0  0  9 Study1 using WebQ-sort data  
 -2  2  0  0  0  0  1  2  3  2  1  0  0  0  0  0  0  0  0  0  0  0

Do not change these lines except for the number of sorts entered (cols 4-6 in first line). Having inserted the first sort, the study1.dat would contain these three lines:

  0  1  9 Study1 using WebQ-sort data 
 -2 2 0 0 0 0 1 2 3 2 1 0 0 0 0 0 0 0 0 0 0 0
PeterS     0-1 1 0-2 2 0 1-1 

Make sure that everything is at the correct position (all numeric data must be right-justified), and that you press the ENTER-key after the last data line.

You will probably notice that some browsers will break the data record into two or more lines. In this case you must delete the line breaks after inserting the data to the .dat file (make sure that columns are correctly justified, insert an empty space where necessary).

Do not forget to adjust the no of sorts value (cols 4-6 in first header line) before saving the file and running it with PQMethod!

WebQ and Browser Versions

WebQ runs best with Netscape Navigator (NN) 4.x or MS Internet Explorer (MSIE) 4.x. MSIE 5.0 unfortunately does not allow to insert line breaks in the email body, the email will be formatted less nicely, and the respondent may therefore overlook the prompt for comments.

NN 3.0 and MSI 3.0 are supported, however, in a restricted fashion with respect to submitting the data, and adding comments (described in the WebQ Help Page).

Another quite specific restriction which only applies when running WebQ from the WWW (but not if run from a local or network drive), refers to the usage of external javascript-code files. For the sake of ease of maintenance, the main portion of WebQ's javascript code is stored in the file webqcode.js, and included from the study-definitions part of the code (e.g., samplejs.htm) at run-time. For NN 3.0 (NN 3.0 only) the including of an external file requires that your provider's web server is configured for MIME-type 'application/x-javascript .js'. You can use Delorie's HTTP Header Viewer after uploading WebQ files to your web site for checking whether the webqcode.js has got the correct MIME-type. However, to be on the sure side, you can simply include the main source code (content of webqcode.js) into the file with the study definitions (as described therein).

Additional Tips and Recommendations

• Since Aug-2002 there exists a modified version of WebQ, which does not rely on the browser's facility to correctly handle the 'transparent email-interface' for submitting the data, and which provides the possibility for simple post-sort queries. There is no extensive documentation for this version available, though example and commentaries within code should be self-explanatory.
  • Download webqnew.zip
• Check my list of other implementations of computerized Q-Sorting.

References and Links

• Block, J. (1961/1978): The Q-sort method in personality assessment and psychiatric research. Springfield, IL: Charles C. Thomas. (reprinted in 1978 by Consulting Psychologists Press, Palo Alto, CA).
• Stephenson, W. (1953): The study of behavior: Q-technique and its methodology. Chicago: University of Chicago Press.
• Brown, S. R. (1980): Political subjectivity. New Haven, CT: Yale University Press.
• McKeown, B. & Thomas, D. (1988): Q Methodology. Newbury Park: Sage Publications.
• Rogers, C. R. (1954): Psychotherapy and Personality Change. Chicago: University of Chicago Press.

The QMethod Page (http://www.rz.unibw-muenchen.de/~p41bsmk/qmethod)

Maintained by Peter Schmolck. The home of WebQ, and the place from where the free PQMethod software for the analysis of Q-sort data can be downloaded.

Credits

• WebQ in its present form is based on the original version of WebQ programmed by Rick Watson,
• Layout and functionality was revised in two major revisions by Peter Schmolck with the help of Eshref Januzaj