RBUS2900 

– Business Research Methods

WHAT ARE WE ACHIEVING TODAY?

 Research Design - Which research design will best answer the

question I am asking?

 Measurement scaling – What are the properties of different

measurement scales?

 Hypothesis Testing – How do we statistically test a hypothesis?

RBUS2900 Lecture 4
BUSINESS RESEARCH METHODS Research Design

THE RESEARCH PROCESS

Research Interpret and

Problem Research
Questions / Data Analysis Report
Identification Design
Hypotheses Findings

Research
Method

Lecture 4
RESEARCH DESIGN
Measurement
Data
Sampling Ethics Research Design
Collection

WHAT IS A RESEARCH DESIGN? WHICH IS THE BEST DESIGN?

 It is a blueprint for outlining all of the procedures from the  Research design is not a discrete choice between exploratory
hypotheses to the data analysis. (qualitative), and descriptive and explanatory (quantitative)
studies.
 It is a framework for specifying the relationships among the study’s
variables.  Researchers combine design strategies in ways that provided a
broader perspective than might be possible from one single design
 It is a plan for selecting the sources and types of information approach, ultimately leading to new and interesting insights.
relevant to the research question.

1
 RBUS2900 – Business Research Methods

PROBLEM CRYSTALISATION THE METHODOLOGICAL PYRAMID

 The research design chosen should reflect the degree of problem

crystallisation.
 A study may be viewed as exploratory, descriptive or explanatory.
o Exploratory studies tend to be loosely structured with an objective
of understanding phenomena.
o The goal of a descriptive or explanatory research designs is to test
hypotheses or answer specific research questions.

EXPLORATORY DESIGN DESCRIPTIVE DESIGN

 Yields qualitative data.  Describes, characteristics of objects, people, groups,

 Usually small-scale research undertaken to define the exact organisations, or environments and tries to “paint a picture”
nature of the problem and to gain a better understanding of of a given situation
the environment within which the problem has occurred.  Descriptive studies are quantitative (i.e., the analysis is “in
 The research is interested in observing, listening and numbers”, not words).
interpreting. As such the researcher is intimately involved in  A descriptive study can be as demanding as explanatory
the research process and in constructing, with participants, studies, involving careful thought about measurement and
the results. sampling.
 There are many methods (e.g. interviews, focus groups,
ethnography etc.) associated with an exploratory design.

Descriptive Diagnostic Studies

EXPLANATORY DESIGN
 Descriptive studies can look at two types of problems:
descriptions and diagnostic.
 The purpose of description is typically to study the “level” of a  Explanatory or “causal” research deals with the study of how
variable (i.e. segmentation, census). one variable (IV) affects, or is responsible for, changes in
another variable (DV).
 The purpose of diagnostic is to diagnose reasons for business
outcomes / problems.  Many relationships of interest are those which are
dependence or causal relationships.
 The survey is the method associated with descriptive designs.
 The experiment is the method associated with the
explanatory design.
 Explanatory research is not always possible.

2
 RBUS2900 – Business Research Methods

EVIDENCE REQUIRED FOR CAUSAL CLAIMS

 While no one can ever “prove" that variable A causes variable

B, one can gather evidence that supports the belief that A does,
in fact, lead to B or that a dependence relationship exists.

 We seek four types of evidence in support of causation.

RESEARCH DESIGN ACTIVITY
o Concomitant variation - is there a predicted covariation between
A and B?
o Temporal sequence - is the time order of events moving in the
hypothesized direction?
o Nonspurious association - is it possible to eliminate other possible
causes of B?
o Theory - is there theoretical support for the predicted relationship?

AN EXAMPLE DESIGN DETERMINES METHOD

Research Design

Interview / Focus Group /

Exploratory
Ethnography etc.

Descriptive/ Descriptive
Survey Method
Diagnostic

Explanatory Experimental Method

3
 RBUS2900 – Business Research Methods

MEASUREMENT SCALES

 The type of statistical analysis you can use is dependent on the

scale of measurement.

 An understanding of the four types of measurement scales, and

whether they are non-metric/metric, is fundamental to mastering
measurement and data analysis.

Lecture 4
MEASUREMENT SCALING
Research Design

Nominal Scales Ordinal Scales

 Represents the most elementary level of measurement in which  Objects or individuals are categorized, and the categories form a
objects or individuals are assigned to categories for identification rank order along a continuum.
and classification purposes only.
 Ranking scales allowing things to be arranged based on how
 The values representing the categories (categorical variables) that much of a concept they possess.
have no numerical properties.
 Ordinal data is sometimes referred to as ranked data.
 Properties: identity
 Properties: identity and ranking.

Interval Scales Ratio Scales

 Scales that capture information about differences in quantities of a  Represents the highest form of measurement.
concept from one observation to the next.
 All the properties of interval scales with the additional attribute of
 Intervals between the numbers on the scale are all equal in size. an absolute zero, indicating an absence of the variable.

 Properties: identity, ranking and magnitude.  Properties: identity, ranking, magnitude and an absolute zero.

4
 RBUS2900 – Business Research Methods

Examples
MEASUREMENT SCALES – DISCRETE OR CONTINUOUS
 Nominal scales: Ethnicity, gender, political affiliation Discrete Measures (non-metric)
 Ordinal scales: Ranking students based on exam results  Measures that can take on only one of a finite number of values.
 Interval scales: Temperature (C)

 Ratio scales: Profit, time Continuous Measures (metric)

 Measures that reflect the intensity of a concept by assigning values

that can take on any value along some scale range.

Discrete
Measures

Continuous
Measures

Lecture 4
HYPOTHESIS TESTING
Research Design

HYPOTHESIS TESTING FORMULATING HYPOTHESIS FOR ANALYSIS

1. Formulating the hypothesis for analysis

 A null hypothesis is a statement of the status quo, one of no
2. Select an appropriate test difference or no effect.

3. Choose level of significance H0: Tutorial attendance is not associated with grades
4. Data collection
 An alternative hypothesis is one in which some difference or effect
5. Data analysis is expected.
6. Business research conclusion Ha: Tutorial attendance is associated with grades

5
 RBUS2900 – Business Research Methods

Non-Parametric Test Parametric Test

SELECT AN APPROPRIATE TEST
Assume that the variables are measured Assume that the variables are
on a nominal/ordinal scale. measured on interval/ratio scale.
 The researcher should take into consideration how the test statistic
is computed. The test statistic measures how close the sample has Tests are available for testing variables Popular parametric tests are the t-test
come to the null hypothesis. from one sample, two independent and the Z-test conducted for
samples, or two related samples. examining hypotheses about means.
 Hypothesis testing procedures can be broadly classified as
A popular example is the chi-square test The t-test can be conducted on the
parametric or nonparametric, based on the measurement scale of conducted for examining hypotheses mean of one or two samples
the variables involved. about proportions. (independent/related samples) of
observations.
Although normality is assumed, the t-
statistic is quite robust to departures
from normality.

Type I Error () Type 2 Error ()

CHOOSE A LEVEL OF SIGNIFICANCE  Type I error occurs when the  Type II error, or beta error, occurs
sample results lead to the when, based on the sample
 Whenever we draw inferences about a population, there is a risk rejection of the null hypothesis results, the null hypothesis is not
that an incorrect conclusion will be reached. Two types of error when it is in fact true. rejected when it is in fact false.
can occur:  The Type I error is controlled by  Unlike , which is specified by the
establishing the tolerable level of researcher, the magnitude of β
1. Type I error risk of rejecting a true null error depends on the actual value
hypothesis of the population parameter.
2. Type II error
 The risk level  is specified by the
researcher.

 Generally in business research

we use 

The Power of A Test

 The Power of A test is the probability (1 – ), of rejecting the null

hypothesis when it is false and should be rejected.

 Although  is unknown, it is related to . An extremely low value of

 (e.g.,  = 0.001) will result in intolerably high  errors.

6
 RBUS2900 – Business Research Methods

 For this course we will use a significance value of 0.05.

 Probability values must be below 0.05 to be significant.

THE END