e-News Vol 1 # 5, 2004
|
|
||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||
 |
The principles outlined below contain a general guide to help you when you are designing your surveys. These seven rules should help you to improve response rates and ensure the quality of the data collected. Failing to Start the Survey With an Introduction Don't underestimate the importance of the introduction. The introduction should both ease the respondent into the survey and motivate the person to complete the survey. It should also instruct the respondent on how to get to the first question in the survey. For example, you could create an introductory page that contains a motivational message, reminds respondents of any incentive, and prompts respondents to enter the password they received through email. Give respondents confidence that they can complete the survey without any trouble. Many times problems can be avoided by using the right tool for the job. For example, do not use alpha text boxes if you are trying to create a ranking or percentage allocation question. Often, one row question is better than merging a lot of similar questions onto the same page.
Making the Questions/Survey Too Long Brevity
remains "the soul of wit." Shorter surveys promote higher response
rates.
To promote clarity and
understanding, try to keep questions under 25 words.
Respondents read more accurately and tend not to jump over words when lines are kept short. SurveyWriter wraps the text of questions, question headers, and project headers. However, if you feel that your question text still extends too far across the respondent’s screen, you can increase readability by inserting hard-carriage returns in the text as needed. While SurveyWriter can easily handle it, making a survey over 200 questions long will certainly test your respondent's ability to handle it. Using Nonstandard Question Layout You are safe if you copy standard paper questionnaire layouts. Make sure that the respondent knows where to start reading each screen. Respondents expect to see question content displayed in the upper left-hand corner of the screen. Use bold or larger fonts and spacing to clearly identify the beginning of individual questions. Two good fonts to use include Arial and Verdana, which are simple and easy to read. Also, avoid overusing color. Limit color to one primary and one or two secondary colors unless you are trained in design. If you can avoid it, limit the use of animation, sound, and embedded programs. Some respondents don’t have the necessary set up, including computer hardware, latest-and-greatest browser, and communication lines, to receive or respond to surveys with complex programming features. Complicated skip patterns are a great feature of online surveys, but be aware of the memory demands that such a survey will place on a respondent’s system.Requiring Answers to all Questions Requiring answers to all questions tends to frustrate some respondents, offends their sense of "ownership" over the process, and ultimately increases terminate rates. Use SurveyWriter's “Required Question” feature selectively. One possible solution is to provide answers such as “prefer not to answer” or “don’t know” for required questions.Setting Up Large, Multi-Question Pages By placing many questions onto a single page, you
limit your ability to incorporate skip patterns and piped responses.
Track Respondents Without Password Security If you send out your own email survey
invitations, SurveyWriter provides a new RID number function that enables you to
track which respondents have completed the survey and which require a reminder
email, without the need to set up password-based security. Of course, this
function is only intended if you want to track respondents, but you don't need
any security features. http://www.surveywriter.com/in/survey/survey999/Trackstudy.asp?
= survey link When a respondent completes a survey, this ID is placed in the data file in the Password field. You can download data to retrieve this information. Two other variable names can also be substituted
for the RID number to accommodate specific panel providers: Both variable names work similarly to the RID number described above.
 Use the Sort Formula Question to Branch in Randomized Answer Options A Sort List formula question enables you to take answers from other questions and put them into a new list. This list is then sorted in either ascending or descending order according to your preference. This formula question also allows you to select the top n answers (e.g., top 3 answers numerically) bottom n answers (e.g., bottom 3 answers numerically), the rank order number (e.g., answer ranked third numerically), or random numbers (e.g., 3 random answers). This example uses the Sort question to display 3 random answers to the respondent in a follow-up branching question. Create Sort Formula Question
Create Follow-Up Branching Question The example below is an open end question that branches in three random answers options from the Sort formula question.
For example: What do you like about |?
Respondent Classifications Expanded SurveyWriter has expanded its classifications for respondents to include two new designations. Previously, you could only download data for "completes" and "terminates." Now, classifications for "over quota" and "incompletes" have been added to the Data Download program. Each classification is described below. Completes: Any respondent who
completes the survey and gets to the "Closing Page." "Sounds like" Prequestion Filter now available The new SOUNDX filter is designed for awareness studies that require respondents to enter the desired responses themselves. If a respondent enters something close to the intended word, the respondent will be allowed to see the next question. For example, you may want to show a question only to those who entered the words, "Federal Express" in the previous question. However, you don't want to exclude them if they've spelled their response slightly incorrectly such as: Fedral Express The SOUNDX filter is used to ensure that all spellings close to the correct one will be accepted. Use the following syntax in prequestion filters. {Q1-C1-R1 [SOUNDX2] Federal Express} You will have to experiment with the similarity levels to ensure that you get the desired result from this filter. To help you test your filters, we've included a link to the test program online: http://www.surveywriter.net/in/DM_Soundx.aspEnter your desired word in the first space and your comparison word in the second. Press Submit to determine level of similarity. You will need to set your similarity levels in the filter to the number displayed to have that filter work as desired in your survey. SurveyWriter.net Enhancement List Online Enhancement Form Available Click Enhancement Form to enter your requests.
You are receiving this newsletter because you either are a client of SurveyWriter or registered at the SurveyWriter.com, Web site. Thank you for your continued support of SurveyWriter. If you have any topics that you would like see addressed in future editions, please email me. Sincerely, |
|
|
|
|
|||||||||||||||||||||||||||||
Seven
Survey Design Mistakes to Avoid
SurveyWriter is an application service provider to the Market Research Industry specializing in web-based surveys and online market research. If you received this email in error, click here to have your name removed from our list. Mail
requests: |