Survey Said SE/EE Reference Manual
Copyright © Marketing Masters 1991-2001 All Rights Reserved
|
Return to the Home Page
The Survey Creator supports the following nine
question and answer types:
- Block Text - For Instructions - No Answer is required
- Single Choice of Multiple Answers
- Multiple Choice of Multiple Answers
- Rating Choice 1 to Interval (Definable Scale)
- Relative Impression 1 to Interval (Definable Scale)
- Ranking in Order of Importance
- Multi-field (Alpha Only, Numeric Only, or Both)
- Verbatim (Limited or Unlimited Response Length)
- Dual Scale - Comparative Selections
This type of question is designed to elicit responses
which require each answer choice to have an explicit label. The
Single Choice and Multiple Choice types are identical except that
the Single Choice answer type allows only one answer to be selected
from all the possible answers defined. The Multiple Choice answer
type allows any or all answers to be selected. The last answer
for these question types can be set to "Other". This
allows the respondent to type in a verbatim answer. The Other
field can be analyzed for both frequency and content (verbatim
analysis).
Example Of Single Choice Of Multiple Answers
Example Of Multiple Choice Of Multiple Answers
Rating Choice allows the respondent to select a single
numeric answer ranging from one to the interval defined for the
question. Relative impression allows the respondent to select
any position between the extremes by using a scroll bar. There
are only two answer labels for these answer types. The answer
labels are considered extremes. The interval should be kept to
a minimum (ie. 5, 7 or 9). By defining a smaller interval, banner
and cross-table analysis yield better results and produce a smaller
grid of possible combinations.
Example Of Rating Choice
Example Of Relative Impression
The Ranking question type allows up to 24 items to
be ranked in order of importance. This allows the respondent
to assign a discrete value to each possible answer. The value
for each answer will range from one to the number of possible
choices. Answers are selected by clicking on the gray square
in front of each of the possible answers. An answer can be changed
by clicking on it again to deselect.
Example Of Ranking In Order Of Importance
The Multi-Field question type is an extremely flexible
method of obtaining respondent input. It can be used for multiple
verbatim, constant sum or variable sum answers. Multi-Field requires
setting Types and Limits for how the question is to be answered.
Types offers settings of alpha only (letters), numeric only (numbers)
or both (letters and numbers). There is also a must = field which
is used only for constant sum answers when the type is set to
numeric only. Limits allows setting the field size, minimum and
maximum values for numeric only answers. Field size controls
the number of characters that can be typed into each of the fields.
Field size can range from 1 to 62 characters while
the number of fields can range from 1 to 24. The minimum and maximum
values are valid for numeric only fields. If a Multi-Field style
question is set to numeric only, then a variety of frequency type
analysis can be performed on the respondent survey data.
A demographic type question requires a field type of both since a street
address for example usually contains both letters and numbers.
Example Of Multi-Field Demographic
To create a question asking for percent responses,
Type and Limits parameters would be set to:
- Field Type = numeric only
- Field Size = 3
- Minimum = 1
- Maximum = 100
- Must = 100
Example Of Multi-Field Percent Constant Sum
An example constant sum question might be to determine
how many business hours a week is spent on various business tasks
or activities. For a 40 hour week, the Type and Limits parameters
would be set to the following:
- Field Type = numeric only
- Field Size = 2
- Minimum = 0 or 1
- Maximum = 40
- Must = 40
Example Of Multi-Field Numeric Constant Sum
An example variable sum question might be to determine
how many leisure hours a month are spent on various activities.
For a 31 day month the Type and Limits parameters would be set
to:
- Field Type = numeric only
- Field Size = 3
- Minimum = 0 or 1
- Maximum = 744 (31 days = 744 hours)
- Must = 0
Example Of Multi-Field Numeric Variable Sum
This type of question is designed to allow the respondents
to enter text as the answer in an open ended environment. Open
ended verbatim questions have a maximum length for responses in
terms of the number of characters allowed. Questions can be sized
to allow the respondents to enter from 5 to 115 characters, or
the interval can be set to 32767 characters providing virtually
unlimited response length.
Example Of Verbatim
The Dual Scale - Comparative Selections question type allows
respondents to select 2 choices that are usually related in
some respect. This question type is used heavily in 360 - Degree
surveys for collecting information concerning what a respondent
thinks something should be vs. what he thinks it is. Also referred
to or known as GAP analysis. This is just one of the many uses for
this question type. The Dual Scale choice question type also allows
for an "Other" field for each column.
Example of Dual Scale
The Survey Analyzer offers the following analysis
options:
- Frequency
- Banner
- Cross Table
- Ranking (as frequency)
- Multi-Field (as frequency or verbatim)
- Verbatim
Frequency analysis counts the number of times each
answer was chosen on a per question basis. The analysis shows
counts and percents for each question independent of how other
questions within the survey were answered. When this option is
chosen, all relevant questions from the survey will appear in
the question selection list box. The frequency analysis automatically
selects all pertinent question types. The resultant analysis may
then be previewed in tabular text form or graphed.
Banner analysis determines the relationship between
how a primary and secondary question were answered. The analysis
displays how the secondary question was answered based on how
the primary question was answered. Four such relationships can
be defined per analysis. That is; one primary question with up
to four secondary relationships. The primary and secondary questions
are chosen from the screen list box of all questions within the
survey and may be changed quickly and easily to view different
relationships. The resultant analysis may then be previewed in
tabular text form or graphed. Banner analysis question selection
allows up to four secondary questions to be analyzed against a
primary per analysis request. Banner analysis can be performed
on Single Choice of Multiple Answers, Multiple Choice of Multiple
Answers, Rating (1 to Interval) and Relative Impression question
types.
Cross Table analysis determines the relationship
between how a primary, secondary and tertiary question were answered.
The tertiary question answers are grouped based on how the secondary
question was answered. The secondary question's answers are grouped
based on how the primary question was answered. Cross Table responses
are thus dependent on the preceding question's responses. If
only one question is selected the analysis behaves similar to
a frequency analysis. If only two questions are selected the
analysis behaves similar to a banner analysis. The resultant
analysis may then be previewed in tabular text form or graphed.
Cross-table analysis can be performed on Single Choice of Multiple
Answers, Rating (1 to Interval) and Relative Impression question
types. The Multiple Choice of Multiple Answers question type is
NOT supported for cross-table analysis.
Ranking analysis identifies those questions within
a survey that are of the ranking type and displays them in the
question selection list box. The ranking type questions may then
be chosen one at a time and subjected to frequency analysis.
The resultant analysis may then be previewed in tabular text form
or graphed.
Multi-Field analysis identifies all those questions
within a survey that are of the numeric type. If the numeric
only option is set during question definition within the Survey
Creator, the questions are displayed in the question selection
list box. Before running a frequency analysis, a number for grouping
may be selected to change the ranges that answers will fall into.
For example, a question is defined to accept numbers from 0 to
100 (i.e. Percent). The analysis, by default will report each
1 percent as a bar on a graph. If however, it is desired to group
the responses into four ranges from (1 to 25), (26 to 50), (51
to 75), and (76 to 100) then change the number in the multi-field
analysis frame (just right of question number selected box) to 4.
Multi-field question types can also be analyzed for
verbatim content. The resultant analysis may then be previewed
in tabular text form, graphed or printed in table format.
Verbatim analysis consists of searching open-ended
questions or respondent's comments for selected words (keywords)
or phrases. Questions may be subjected to frequency analysis
based on the counts of keywords and key phrases found. The resultant
frequency of keywords matched, may then be previewed in tabular
text form or graphed. Verbatim responses can be categorized by
the survey administrator and coded with a unique keyword tag or
phrase. A keyword file of the unique tags can be built and the
coded responses matched (analyzed) against it.
Survey data may be displayed and printed in either
a tabular text form showing counts and percentages or graphed.
The tabular text form below shows a Cross Table analysis
of questions 35, 36 and 6 in that order. Question 35 has two possible
answers, male or female. In this case, 102 of the respondents
said they were male or 51% of those polled, and 99 were female
or 49%. Question 36 asked if the respondent was married or single.
Of the 102 males, 55 were married and 47 were single yielding
53.9% and 46.1% respectively. The third question, number 6, has
five possible answers ranging from Completely Satisfied to Very
Dissatisfied. Notice the breakdown for each answer.
EXAMPLE PREVIEW OF CROSS TABLE ANALYSIS OF QUESTIONS 35,36 AND 6
35-1 (102) 51% Male Sex
36-1 (55) 53.9% Married Marital Status
6-1 (8) 14.5% Completely Satisfied Overall Sales
6-2 (14) 25.5% Very Satisfied
6-3 (9) 16.4% Fairly Well Satisfied
6-4 (15) 27.3% Somewhat Dissatisfied
6-5 (9) 16.4% Very Dissatisfied
36-2 (47) 46.1% Single Marital Status
6-1 (7) 14.9% Completely Satisfied Overall Sales
6-2 (10) 21.3% Very Satisfied
6-3 (10) 21.3% Fairly Well Satisfied
6-4 (8) 17.0% Somewhat Dissatisfied
6-5 (12) 25.5% Very Dissatisfied
35-2 (98) 49% Female Sex
36-1 (57) 58.2% Married Marital Status
6-1 (14) 24.6% Completely Satisfied Overall Sales
6-2 (14) 24.6% Very Satisfied
6-3 (10) 17.5% Fairly Well Satisfied
6-4 (11) 19.3 % Somewhat Dissatisfied
6-5 (8) 14.0% Very Dissatisfied
36-2 (41) 41.8% Single Marital Status
6-1 (7) 17.1% Completely Satisfied Overall Sales
6-2 (8) 19.5% Very Satisfied
6-3 (9) 22.0% Fairly Well Satisfied
6-4 (10) 24.4% Somewhat Dissatisfied
6-5 (7) 17.1% Very Dissatisfied
The following graph types and styles may be viewed
on screen and then printed in color or black and white:
Types of Graphs
. Type . . Style .
Pie 2D, 3D Percent Labels
Count Labels
Area 2D, 3D Percent Labels
Count Labels
Bar 2D Vertical and Horizontal
Vertical Stacked
Horizontal Stacked
Vertical Stacked Percent
Horizontal Stacked Percent
Bar 3D X Axis Clustered Bars
Z Axis Clustered Bars
Vertical Stacked
Horizontal Stacked
Vertical Stacked Percent
Horizontal Stacked Percent
Line 2D Lines Only
Symbols Only
Line Plus Symbol
Mean
Best Fit
Line 3D Area
T a p e
Bubble Hot Bubble
The tabular text form of analysis tables may be copied
by selecting all or a part of the textual report. This is accomplished
by putting the text selected in the clipboard. This text may
then be pasted into a word processor or other application. The
analysis graphs may be sent directly to the clipboard from the
Print Graph form. They also may then be pasted into a word processor
or other application.
Question keywords are abbreviations or short summaries
of questions which are used to recognize them at a glance. For
the question "How did you select company XYZ?" the keywords
might be "Select-How". Every question is required to
have an associated keyword which facilitates question recognition
for table, graph and printed reports.
Question keywords are also very useful when viewing
frequency, banner and cross table analysis. Question keywords
may be up to 32 characters in length, but because of the limited
space on the graphics screen, we recommend you keep keywords short
and concise.
Skip patterns allow "jumping" to a specific question depending on how
the current question is answered. An example might be a medical survey
where a respondent is asked a different set of questions based on their
gender. Another example might be an automotive service survey where a
respondent is asked if they still own the vehicle to determine if they
should finish answering the survey. Specific answers allow skipping to
a different question if they are single choice answers. Multiple choice,
verbatim, ranking, multi-field type and dual scale questions do not allow
question skipping by specific answers. They do allow conditional skipping
based on whether or not they are answered. The Block Text question type
allows unconditional skipping to a question after being read. This is
useful as a "jump point" for a series of questions that has just been
finished and the administrator needs to branch the respondent to an
entirely new section of the survey.
Forcing an Answer to be Required
Skip patterns can be used to force a respondent to answer a question
before going to the next question. In most cases, the answer can not
be validated, but it can at least require some type of answer be given.
The way to accomplish this is to set the skip pattern so that if the
question is "not answered" it skips to itself. For example, suppose
question 12 is a Yes/No question that you would like answered. Then
set a skip pattern for question 12 that defines question 12 as the
question to skip to if question 12 is not answered. Survey Said will
interpret the skip as meaning an answer is required. This is true
for all the versions of the Survey Respondent program and Java Applet,
as well as paper surveys and HTML surveys generated by Survey Said.
Copyright © Marketing Masters 1991-2001