Definitions
Here are some of the commonly needed terms
for understanding surveys.

Survey - a series of questions designed to
find out useful information from a particular
population.

Population - a group of people having views,
opinions, and attitudes you want to measure.

Respondent - one person who responds to the
survey.

Response - a complete set of answers to the
survey given by one respondent.

Subpopulation - a subgroup within a survey
population, or within a group of respondents,
who have certain distinct answer patterns.

Identifier - a name, phrase, or other alpha-
numeric description that identifies an
individual respondent.  It can be up to 25
characters long.  You can leave it blank if
you like.

Multiple-choice Question - one of three types
of questions you can use in a SurveyMaker
survey; it offers the respondent a series of
predefined choices, from which he or she
merely selects one by number.  The number of
the option becomes the answer value used for
statistical calculation.

Numeric Question - one of three types of
questions you can use in a SurveyMaker
survey; the respondent supplies a number,
such as age, number of children, or length of
service with the company.  A numeric answer
can be up to 10 digits long, including
decimal point (and minus-sign if it happens
to be a negative number).

Comment Question (narrative) - one of three
types of questions you can use in a
SurveyMaker survey; the respondent supplies a
comment, which can be up to 250 characters
long.  You can enter comment responses into a
data file right along with multiple-choice
and numeric values.

Frequency - a survey statistic; shows the
number of respondents who gave a particular
answer to a multiple-choice question; may be
expressed as a count or as a percentage of
the people who responded to that item.




Non-response - a survey statistic; for each
item, there will probably be a few people who
do not supply an answer, for any of a variety
of reasons.  SurveyMaker keeps track of these
non-response counts separately, and adjusts
the statistical calculations accordingly.
When entering response data, you just hit the
asterisk key ( * ) to show a non-response.

Range - the spectrum of answer values to a
given question; identified by the lowest
observed answer value and the highest.

Mean Value - a survey statistic; the average
of the responses to a given question.

Median Value - a survey statistic; the value
in a group of responses that is the midpoint
of the range of values; half of the answers
exceed the median and half fall below it.
Mode - (modal value) a survey statistic; the
most frequently occurring answer value in a
set of responses to a particular item.

Standard Deviation (SD) - a survey statistic;
a number that indicates the amount of
variability in the responses to a particular
item.  When SD = 0, there is no variation in
the data; all responses are the same.  A high
value of SD indicates a wide dispersion of
the answer values around the mean value.

Histogram - an advanced SurveyMaker feature;
a vertical bar-chart that shows the relative
frequency of occurrence of the various
possible answer values across the range of
values.  Also known as a bell curve or a
frequency distribution.


Percentile - an advanced SurveyMaker feature;
a group of response values that make up one
per cent of the total respondent population,
with the answers arranged in order of size.
Each of the 100 percentiles has a count, i.e.
the number of responses in that group, a
lowest answer value and a highest value.
Percentiles are useful for evaluating
performance data, e.g. test scores, sales
levels, etc.

Decile - an advanced SurveyMaker feature;
a group of response values that make up ten
per cent of the total respondent population,
with the answers arranged in order of size.
Deciles are simpler to work with than
percentiles, but give less detail in
evaluating the data.  Each decile contains 10
percentiles.  You can look at quintiles (20-
per cent bands), quartiles (25-per cent
bands), etc. by identifying the proper ranges
on the percentile scale.

Correlation coefficient - an advanced survey
statistic; a calculated number that shows the
degree to which the responses to two
different question items vary together.  This
is often useful for identifying possible
patterns among the respondents.  [Note to
advanced users: SurveyMaker uses the Pearson
product-moment formula for calculating the
correlation coefficient.]

Cross-plot - an advanced SurveyMaker feature;
a chart showing the relationship between the
responses to two question items, with one
item on the horizontal (X) axis and the other
on the vertical (Y) axis.  [Advanced users:
SurveyMaker gives you a scatterplot, a cross-
tab matrix, and a regression plot.]
Data Filter - a workscreen that comes up when
you tell SurveyMaker to scan a data file and
build a report.  The filter allows you to
screen out responses that don't match the
criteria you specify on the workscreen.  This
enables you to pull out reports for specific
subpopulations of respondents based on their
answer values to particular questions.