°°°° °°° °°° °°° °°° °°° °°° °°° °°° °°° ELECTRONIC QUESTIONNAIRE MANAGER °°°° °°°° version Autumn 1994 °°° °°° °°° °°°° °°° °°° °° °° °°° °°°° °°°° °°°° °° °° °°° Authors: A & S Sharpe 4 Drovers Way Peebles EH45 9BN Scotland Internet: CENSCS@vaxb.hw.ac.uk ************************************************** The Authors disclaim all warranties as to this software, whether express or implied, including without limitation any implied warranties of merchantability, fitness for a particular purpose, functionality, data integrity or protection. ************************************************** 1. INTRODUCTION This is a shareware program (see later section 'About Shareware') - it is NOT freeware. Please try out the program. If you find it useful, please register and you will receive an unlimited program with full documentation. The manual included with the registered software includes hints and tips regarding the authoring of a questionnaire. The documentation included here provides a basic introduction to the program and how to configure it. 2. OVERVIEW This program resulted from a need to carry out on-line questionnaires in a structured way on large PC networks. These surveys were required to give quantitative feedback to system managers about the service they were providing. Many of the network users were not computer experts, so the program had to be easy to use and had to include reasonable on- line help. Although Windows applications predominated on these networks, in order to ensure a reasonable response, users had to be able to run the questionnaire as part of their login process - this meant using a DOS program. The program had to be capable of being run on ANY PC on the networks. This meant coping with the lowest spec. machines, which again meant the program had to be a DOS one. The program will run with any PC/DOS (it hasn't been tested below DOS v3.1 though). It is suggested that at a colour screen should be used, however. As well as allowing you to compose questionnaires with complex logic and diverse question types, this version of the Questionnaire Manager also allows you to ANALYSE the results. The results are stored in a form easily imported in to Lotus, Excel, Access etc., but failing these a quick analysis will be provided by the program in the form of a plain text report 3. HOW TO COMPOSE A QUESTIONNAIRE The program suite consists of: o one executable file - QMAN.exe o one ASCII text data file - must be specified as the first argument on the command line. (An example is include - EXAMPLE.dat) o one ASCII text results file - by default RESULTS.dat The data and results file locations and names can be changed as desired - type QMAN.exe /? to view all available options. If the results file already exists, further results will be appended at the rate of one delimited line per respondent. QMAN.exe manages the questionnaire using the definition statements in the data file. Thus the form and content of the questionnaire are entirely customisable. If you print out EXAMPLE.dat you will see that it contains all the types of questions described in the next section. You can copy this file or edit a new one to create your own datafiles. All you need is a plain text file editor, such as EDIT, as supplied with DOS 5 and 6.x. Alternatively Windows Notepad would be fine. 3.1 TYPES OF QUESTION With this version of the Manager, the following basic question types are available: o Options o Bar o Yes/no o Follow-on OPTIONS questions are ones where the respondent is shown a question and then presented with a list of options to choose from. In this version an Options question ALWAYS includes the final option 'Other'. An example might be: 1) What animal would be your favourite pet? 1) Dog 2) Cat 3) Snake 4) Penguin 5) I hate all pets 6) Other If you select 6) (Other), the program will automatically prompt for the 'Other' choice, and record this as delimited text in the results file. Otherwise the results file simply records the number of the option chosen. BAR questions can be used where the respondent must choose between extremes. The respondent is shown a bar with the extremes at each end, and must place a pointer at the position between them that best matches their response. A score out of 10 is recorded in the results file. An example might be: 1) How clever do you think Cats are? ¦ Not at all +---+---+---+---+---+---+---+---+---+---¦ Genius' clever ¦ In this example, the respondent thinks that Cats are quite clever, but not actually genius'. They would score 7.75/10. YES/NO questions are a little like Options questions, except that the only options are Yes or No. There is no 'Other' option. Results are stored as 1 for Yes, 2 for No. An example might be: 1) Are you a cat owner? 1) Yes 2) No FOLLOW-ON questions are special. They can be any of the above types, but are normally ONLY displayed if a specific option is chosen in an Options question - they conditionally follow-on from an options question. There may be a number of follow-on questions related to various options in the same question, or all following-on from one option. If the follow-on option is not chosen , follow-on questions are ignored. 3.2 DATA FILE FORMAT This is split in to two sections - the main section and one or more question sections. In general the section identifiers must be between [] brackets on a line of their own, starting at the first column. All other datafile lines must start at the first column and will consist of a keyword, '=' and a value - with no spaces or other characters inbetween. The value is generally plain text (with no inverted commas etc.). Some keywords require a question number as the value, this must be in two-digit format as in the section identifier. E.g. If a question section is numbered [05], this must be referenced as '05' elsewhere (and NOT '5'). Thus these are OK: Type=bar TYPE=BAR type=BAR or Fol=05 but these will not be recognised: T ype=bar type = bar type=bar (^^- leading spaces) or Fol=5 Basic syntax checking is carried out and the program will abort with an informative message if the format or syntax is incorrect. The main section has the following format: [Main] Title= SubTitle= End= Separator= The only obligatory part is the Title= line. End= is a line to display on the closing screen. Seperator= defines a character to be used instead of ',' as the field separator in the results file. Question sections are formatted as follows: 3.2.1 Options questions [01] Type=Opt Prompt= Opt= Fol= Fol= Opt= Opt= Fol= Opt= Help= 01 is the Question number, and this should be unique to avoid confusion when analysing the results. Note question numbers do not need to be sequential, but are always presented in the order that they are defined in the datafile. Prompt= is the question you prompt respondents with, it should be plain text without out inverted commas etc. There must be one Prompt= per question. There may be any number of Opt= lines, although too many will not fit on a screen. Fol= is optional and defines the question number (e.g. 01) of a follow-on question, should the previous option be selected. Help= allows a context-sensitive, question-specific help line to be added to the default help when the user hits F1. 3.2.2 Bar questions [02] Type=Bar Prompt= Lo= Hi= Help= Again Prompt= is the text of the question that respondents are prompted with. Here Lo= is the left extreme, and Hi= is the right extreme. Help= is as above. 3.2.3 Yes/no questions [03] Type=Yes Prompt= Help= Prompt=, Help= are as previously. The only possible options are 'Yes' or 'No', so you are given no choice in this. 3.2.4 Follow-on ONLY questions These are identical to any of the above, except that they must start with the line Ignore, e.g. [04] Ignore Type=Yes Prompt= Help= etc. These questions may be of any Type=, taking the appropriate prompts as defined in the above questions. Note a Follow-on question will only be displayed if it's related option is selected by the respondent. Otherwise it will be ignored. You can still designate a question as a follow-on, but remove the Ignore statement in the follow-on question. In these circumstances the question will potentially be called twice. The normal flow of events is best illustrated using the EXAMPLE.dat file question 02: In this case Option 1 is chosen, and so follow-on questions 05 and 06 are executed, followed by the next question, 03: Question Option Follow-on Question 02 -------> Opt 1 ---------------------> 05 : +-> 06 : 03 <-------------------------------------+ If Option 2 were chosen, the sequence would have been: Question Option Follow-on Question 02 -------> Opt 2 : 03 <---------+ It is possible to nest follow-on questions to more than one level: this is not recommended, as the logical flow of questioning will typically prove very difficult to maintain. 4. COMMAND LINE OPTIONS Command line options are as follows: QMAN.exe datafile [/A=answerfile] [/N [/U]] [/S] [/? | /H] [/D] [/T] where datafile A plain text file holding the questionnaire definition data, as defined in the previous section. Default is QMAN.dat /A=answerfile The file used to store the results of the survey. Default is RESULTS.dat /N NetWare aware - the UserID of the respondent is stored as the first field of the answerfile. /U Only available with /N, this option makes the Questionnaire Manager check to see if the user has already completed the questionnaire before allowing the questionnaire to be run. /S Survey analysis - requires the datafile at least to be specified. Also requires /A if an answerfile other than RESULTS.dat is used. Must use /N and/or /T if these were used when compiling the survey results. /?, /H Displays this help. /D Displays a quick summary of datafile options. /T Timestamp the completion of the survey in the results file - first field, or second if /N is used. EXAMPLES To run the questionnaire with the examples file enclosed on a NetWare Network, checking for prior completion and storing UserID as the first field in the answerfile. Results stored in answerfile H:\RESULTS.txt: QMAN.exe EXAMPLE.DAT /N /U /A=H:\RESULTS.TXT To analyse the resulting file (defaulted to answerfile RESULTS.dat): QMAN.exe EXAMPLE.DAT /N /U /S /A=H:\RESULTS.TXT ************************************************ 5. ABOUT SHAREWARE Shareware is a try-before-you-buy process in which you are allowed to use the program free for a specific period. This time will vary from program to program. The grace period for this program is 3 weeks. If after that time you find the program useful, you are expected to register with the author (see registration form below). If however after the specified time you do not find the program useful you are required to remove it from use. Remember that the shareware system will only work if you support it by registering your program with the author. This does a few things like encourage the authors of shareware to continue to write quality programs, upgrade existing programs and supply support for those programs. If you choose not to register, we would still like to hear about any comments or suggestions you may have regarding the Electronic Questionnaire Manager. 6. PROBLEMS & TECHNICAL SUPPORT If you encounter any problems with the program, write to us stating the problem along with the version number of the program. We will be glad to do everything we can to get your problems solved quickly. Also on the Internet : shirley@hw.edu.uk The only known problem currently is one that can occur if you use numeric only NetWare account IDs with numbers greater than 32,767. If you then use the /N option to record results, but do not use /N when analysing with /S, it is possible for the Questionnaire Manager to assume the User ID's are in fact valid option numbers. The IDs will then overflow the variable used to store option results. The fix is to correctly use /N when both recording results and analysing them. 7. PLANNED UPGRADES The planned upgrades to the program include: o A progress indicator (bar graph and percentage) to reassure respondents when completing long surveys o Better context-sensitive help o More question types o Follow-on questions in Yes/no type questions