Sample Solution Set/Extra Examples of "Identifying Variables and Their Labels
in Research Questions"

This business of identifying "what are my variables and how should they
'best' be labelled" is one of those 'gloriously messy/tricky real-world
chores' that invariably confront every researcher! For example, for the
variable 'income' (I'm just making this up!), do I as the researcher need the
information 'as finely' as "in precise dollars and cents?" Or will ranges be
OK: 1) "Under $10,000; 2) $10,001 to $20,000; 3) $20,001 to $30,000," etc.!
Oftentimes, there is 'no single best' or 'right or wrong answer' too!

I wondered if a 'sample solution set' would be helpful! I've made up some
research questions and will illustrate what the variables are and what the
labels should be - hope the following are helpful:

Descriptive Question: "What are the primary barriers to successful delivery
of courses by modem?"

The single target outcome variable here would be "barriers to successful
modem course delivery:" what we are trying to find out about - our 'goal' for
the descriptive study!

Correlational Question: "What is the relationship between prior exposure to
computer use and propensity to enroll in modem courses?"

There will be 2 main variables in this correlational question, and they are
the ones being related or associated:

1) Predictor Variable: The one that "came first:" "prior exposure to computer
use" and
2) Criterion Variable: The one that 'came after' and is being forecasted from
the first one: "propensity to enroll in modem courses"

There could be lots of 'confounding' variables for this example --
uncontrolled variables that 'get in the way' and can 'markedly influence' the
seemingly simple relationship between the 2 variables above! They could
include: whether or not you have a computer in your home; whether or not your
friends/family like and use computers; availability of modem courses; age;
gender; area of major; etc.!

Experimental Question: "What is the effect of daily practice in a computer
lab on performance in courses by modem?"

Here we would envision two major types of variables:

Independent Variable: the one that 'came first' and is used to form the
groups -- maybe an overall 'umbrella' label like "Method of Instruction in
Courses by Modem" -- and there would be the following two groups (we also
call them "levels" of the independent variable):
Group 1: Those who get the daily lab session - we call this the
'experimental' or 'treatment' group; and
Group 2: Those who don't get the daily lab session - but instead the 'usual'
method of instruction - we call this the 'control' or 'comparison' group!

Then we try and 'keep constant' as many other factors or contaminants as
possible -- admittedly messy to do in the real world we live in! or "life is
not a sterile laboratory," thank goodness! such as "what is taught; in what
order; experience of teacher; year in school of the students:' etc., and then
we compare at the end to see if the two groups differed on average:

Dependent variable: the one that "came after" and the one that we would
expect to be 'affected by' which group the subjects were in! That would be
"performance in modem courses" (however it is measured; for instance, an
aptitude or achievement test - same test given to both groups at the end)!

Evaluation Question: Could be same as for Experimental; or an evaluation
question which might take place in a non-experimental setting might be
something like this: "What is the impact of group assignments on performance
in modem courses?"

For this one, maybe the students got to pick whether or not they worked in
groups, so we don't in this case have the 'tight control' that we would in an
experimental study. Maybe it is even the case, such as for our own Intro to
Research class, that *all* of the students work in groups! So it's not
'experimental,' but still we want to 'assess the impact' or 'evaluate' how
group work impacts performance in modem courses!

Independent variable: came first: Assignment to group work;
Dependent variable: came after - the 'target outcome' which we think or
expect will be 'impacted by' the group work: which would be 'performance in
modem courses" (as for experimental question, however it is measured -- i.e.,
letter grade; score on standardized test; etc. -- lots of leeway here!)

Go on to Assignment 3-1-2
or
Go back to Understanding Variables

E-mail M. Dereshiwsky at statcatmd@aol.com
Call M. Dereshiwsky at (520) 523-1892

Copyright © 1999 Northern Arizona University
ALL RIGHTS RESERVED