Professional Documents
Culture Documents
Definition of Parametric Test: Predictor Variable Outcome Variable Research Question Example Paired T-Test
Definition of Parametric Test: Predictor Variable Outcome Variable Research Question Example Paired T-Test
Definition of Parametric Test: Predictor Variable Outcome Variable Research Question Example Paired T-Test
The parametric test is the hypothesis test which provides generalisations for making
statements about the mean of the parent population. A t-test based on Student’s t-statistic,
which is often used in this regard.
The t-statistic rests on the underlying assumption that there is the normal distribution of
variable and the mean is known or assumed to be known. The population variance is
calculated for the sample. It is assumed that the variables of interest, in the population are
measured on an interval scale. In this assignment two types are given t-test i.e. for finding
difference among group mean and regression test i.e. to test cause-and-effect relationships.
Examples are given for the types below respectively.
The nonparametric test is defined as the hypothesis test which is not based on underlying
assumptions, i.e. it does not require population’s distribution to be denoted by specific
parameters.
The test is mainly based on differences in medians. Hence, it is alternately known as the
distribution-free test. The test assumes that the variables are measured on a nominal or
ordinal level. It is typically used when the independent variables are non-metric. A type of
non parameter test is the chi-square. An example of a chi square is suppose we have a cross-
sectional study in which we ask a random sample of people whether they think their general
practitioner is doing a good job, on a five point scale, and we wish to ascertain whether
women have a higher opinion of general practitioners than men have. The input variable is
gender, which is nominal. The outcome variable is the five point ordinal scale. Each person's
opinion is independent of the others, so we have independent data.
1
Predictor variable Outcome variable Use in place of
Non-parametric tests are valid for both non-Normally distributed data and normally
distributed data, so why not use them all the time?
It would seem prudent to use non-parametric tests in all cases, which would save one the
bother of testing for normality. Parametric tests are preferred, however, for the following
reasons:
1. We are rarely interested in a significance test alone; we would like to say something about
the population from which the samples came, and this is best done with estimates of
parameters and confidence intervals.
2. It is difficult to do flexible modelling with non-parametric tests, for example allowing for
confounding factors using multiple regression.
3. Parametric tests usually have more statistical power than their non-parametric equivalents.
In other words, one is more likely to detect significant differences when they truly exist.
The fundamental differences between parametric and nonparametric test are discussed in the
following points:
1. A statistical test, in which specific assumptions are made about the population
parameter is known as the parametric test. A statistical test used in the case of non-
metric independent variables is called nonparametric test.
2. In the parametric test, the test statistic is based on distribution. On the other hand, the
test statistic is arbitrary in the case of the nonparametric test.
3. In the parametric test, it is assumed that the measurement of variables of interest is
done on interval or ratio level. As opposed to the nonparametric test, wherein the
variable of interest are measured on nominal or ordinal scale.
4. In general, the measure of central tendency in the parametric test is mean, while in the
case of the nonparametric test is median.
5. In the parametric test, there is complete information about the population. Conversely,
in the nonparametric test, there is no information about the population.
6. The applicability of parametric test is for variables only, whereas nonparametric test
applies to both variables and attributes.
7. For measuring the degree of association between two quantitative variables, Pearson’s
coefficient of correlation is used in the parametric test, while spearman’s rank
correlation is used in the nonparametric test.
2
Hypothesis Tests Hierarchy
3
CONSTRUCTS
The examples above highlight a desire to capture what we mean about something through the
use of just a few words (often only one or two words). Take the following examples:
4
We often refer to constructs as mental abstractions because seldom are constructs directly
observable (e.g., we cannot directly observe depression, even though we may associate
depression with signs such as a person that often cries, engages in self-harm, has mood
swings, and so forth). Since constructs are very broad and abstract, conceptual clarity has
become one of the cornerstones of good research.
Easy to understand and measure: We know that the construct, sex, has just
two attributes in humans: male and female. If we choose to include the construct, sex,
in our research, this would be relatively easy to justify. After all, most people would
associate the construct, sex, as referring to males or females.
This difference in complexity raises two issues: (1) the need to explain to the reader what you
mean when you use a particular construct; and (2) the fact that a construct can have more
than one meaning, highlighting the importance of explaining what the construct means to
you.
5
Imagine a debate about famine, without knowing the meaning of other constructs such
as starvation, drought, poverty, disaster relief, food supply, survival, nutrition, aid, and so
forth. Without a clear and precise way of explaining what these constructs mean, we would
struggle to communicate to our audience.
Clarity is also required when setting out the constructs you intend to study
because (a) the meaning/understanding that people get from a construct can be different,
and (b) constructs can be viewed from a number of different perspectives, even when there is
6
agreement of the meaning of constructs (e.g., intelligence may be viewed in terms of IQ, but
also EI - emotional intelligence - or another perspective).
Ultimately, we have to make a choice between possible (i.e., multiple) theoretical or nominal
definitions of constructs; we have to settle on a particular definition, but explain: (a) why we
made this choice over definitions that adopt the same/similar approach, in terms of (i) how it
fits with your research agenda and (ii) support for a definition in the literature amongst other
researchers; and (b) how the conceptual definitions relates to other constructs, whether (i) in
the same area or (ii) another conceptual area we are interested in. It is important to note that
despite the need to make choices between conceptual definitions, there are
no right or wrong answers.
Since a conceptual definition only provides the platform for the operational definition that is
used to empirically measure constructs, we discuss about translating constructs
into operational definitions in the next section.
7
Translating constructs into operational definitions
Whilst constructs are sometimes mistaken for variables, they are not variables. Instead, we
use variables to operationalize (i.e., measure) the constructs we are interested in. Constructs
can be mistaken for variables because some constructs may only be represented
by one variable, such that the construct name and the variable name are the same (e.g., the
construct and variable, sex).
Ultimately, the operational definition is seldom perfect; that is, the choice of operational
definition may be constrained by factors such as a lack of access to operational/measurement
data. Also, how we construct/formulate an operational definition will impact on the
complexity of the measurement process.
8
CATEGORICAL VARIABLES
The first question highlights the use of categories and the second question levels. For
example:
Question 2: On a scale of 1 to 5, how important are the following factors in influencing
your career choice [1 = least important; 5 = most important]?
Variables with levels
Career prospects
Nature of the work
Physical working conditions
Salary and benefits
ETC...
9
Each of these types of categorical variable (i.e., nominal, dichotomous and ordinal) are
described below with associated examples:
Nominal variables
The following are examples of nominal variables. These nominal variables could address
questions like:
When we talk about nominal variables not having an intrinsic order, we mean that they can
only have categories (e.g., black, blond, brown and red hair); not levels (e.g., a Likert scale
from 1 to 5).
10
Dichotomous variables
The following are examples of dichotomous variables. These dichotomous variables could
address questions like:
Dichotomous variables are nominal variables that have just two categories. They have a
number of characteristics:
11
For example, you are either male or female. You either like watching television (i.e.,
you answer YES) or you don't (i.e., you answer NO).
For example, some variables (e.g., your sex) can only be dichotomous (i.e., you can
only be male or female). They are therefore fixed. In other cases, dichotomous
variables are designed by the researcher. For example, take the question: Do you like
watching television? We have determined that the respondent can only
select YES (i.e., I like watching television) or NO (i.e., I don't like watching
television). However, another researcher could provide the respondent with more than
two categories to this question (e.g., most of the time, sometimes, hardly ever). Where
more than two categories are used, these variables become known
as nominal variables rather than dichotomous ones.
Ordinal variables
Just like nominal variables, ordinal variables have two or more categories. However, unlike
nominal variables, ordinal variables can also be ordered or ranked (i.e., they have levels). For
example, take the following example of an ordinal variable:
So if you asked someone if they liked the policies of the Democratic Party and you presented
them with the following three categories: Not very much, They are OK, or Yes, a lot; you
have an ordinal variable. Why because you have 3 categories namely Not very much, They
are OK, and Yes, a lot and you can rank them from the most positive (Yes, a lot), to
the middle response (They are OK), to the least positive (Not very much). However, whilst we
can rank the three categories, we cannot place a value to them. For example, we cannot say
that the response, They are OK, is twice as positive as the response, Not very much.
12
Ordinal variable: Year of university entry
Level: 2006, 2007, 2008, 2009, 2010
Question: To what extent do you agree or disagree with the following statement:
Going to university is important to get a good job
[based on a 5-point Likert scale of 1 = strongly agree, 2 = agree, 3 =
neither agree nor disagree, 4 = disagree, 5 = strongly disagree]
Answer: 2 = I agree (or 1, 3, 4 or 5 on the 5-point Likert scale)
Ordinal
The importance of university to getting a good job
variable:
Level: 1 = strongly agree, 2 = agree, 3 = neither agree nor disagree, 4 =
disagree, 5 = strongly disagree
CONTINUOUS VARIABLES
Interval variables
Interval variables have a numerical value and can be measured along a continuum. Some
examples of interval variables are:
13
However, temperature measured in degrees Celsius or Fahrenheit is NOT a ratio variable.
This is because temperature measured in degrees Celsius or Fahrenheit is not a ratio variable
because 00C does not mean there is no temperature.
Ratio variables
Sometimes, the measurement scale for data is ordinal, but the variable is treated as though it
were continuous. This is more often the case when using Likert scales. When a Likert scale
has five values (e.g., strongly agree, agree, neither agree nor disagree, disagree, and strongly
disagree), it is treated as an ordinal variable. However, when a Likert scale has seven or
more values (e.g., strongly agree, moderately agree, agree, neither agree nor disagree,
disagree, moderately disagree, and strongly disagree), the variable is sometimes treated as
a continuous variable. Nonetheless, this is a matter of dispute. Some researchers would argue
that a Likert scale should never be treated as a continuous variable, even with seven
levels/values.
14
Since you are responsible for setting the measurement scale for a variable, you will need to
think carefully about how you characterise a variable. For example, social scientists may be
more likely to consider the variable gender to be a nominal variable. This is because they
view gender as having a number of categories, including male, female, bisexual and
transsexual. By contrast, other researchers may simply view gender as
a dichotomous variable, having just two categories: male and female. In such cases, it may be
better to refer to the variable gender as sex.
What is the effect of diet and The type of soda Your blood
regular soda on blood sugar you drink (diet or sugar levels
levels? regular)
15
Experiment example
You are studying the impact of a new medication on the blood pressure of patients with
hypertension.
To test whether the medication is effective, you divide your patients into two groups. One
group takes the medication, while the other group takes a sugar pill.
Your independent variable is the treatment that you vary between groups: which type
of pill the patient receives.
Your dependent variable is the outcome that you measure: the blood pressure of the
patients.
Below is chart showing the effect of the two medications on the blood pressure of the patient
16
The independent variable is usually applied at different levels to see how the outcome differs.
You can apply just two levels (e.g. the new medication and the place) in order to find
out if the independent variable has an effect at all.
You can also apply multiple levels (e.g. three different doses of the new medication) to find
out how the independent variable affects the dependent variable.
Instead, they must find already-existing examples of the independent variable, and investigate
how changes in this variable affect the dependent variable.
Research example:
You are interested in whether a higher minimum wage impacts employment rates.
You can’t control the minimum wage yourself. Instead, you look at a state that raised its
minimum wage last year, and compare it to a neighbouring state that did not.
By comparing the difference in outcomes between the two states (and accounting for other
factors), you can investigate whether the change in minimum wage had an effect on
employment rates.
17
In non-experimental research, it’s more difficult to establish a definite cause-and-effect
relationship, because other variables that you haven’t measured might be influencing the
changes. These are known as confounding variables.
In types of research where the exact relationship between variables is less certain, you might
use different terms for independent and dependent variables.
18
REFERENCES
Surbhi, S. (2017 September 1). Difference Between Parametric and Nonparametric Test.
https://keydifferences.com/difference-between-parametric-and-nonparametric-test.html
Bevans, R (2021 August 2). Statistical tests: which one should you use?
https://www.scribbr.com/statistics/statistical-tests/
Campbell MJ and Swinscow TDV. 2009. Statistics at Square One 11th ed. Wiley-Blackwell:
BMJ Books. https://www.healthknowledge.org.uk/public-health-textbook/research-
methods/1b-statistical-methods/parametric-nonparametric-tests
Butler, R. N. (1975). Why Survive? Being Old in America. San Francisco: Harper & Row.
Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational
advantage. Academy of Management Review, 23(2): 242-266.
Thomas L. (2020, May 20) Revised on April 19, 2021. Independent and dependent variables.
https://www.scribbr.com/methodology/independent-and-dependent-variables/
19
20