muhamad jantan: [email protected] to data analysis1 the research process observation broad...

Muhamad Jantan: [email protected]

Introduction to Data Analysis 1

The Research The Research ProcessProcess

OBSERVATION

Broad Area of Research Interest

PROBLEM DEFINITION

Research problem

delineatedPRELIMINARY

DATA GATHERING

Interviews & Literature

review

THEORETICAL FRAMEWORK

Variables clearly identified

GENERATION OF

HYPOTHESES

DATA COLLECTION, ANALYSIS &

INTERPRETATION

DEDUCTION

Hypotheses substantiated?

Research Questions answered?



Data Analysis - IssuesData Analysis - Issues

What is data analysis?

Types of techniques

Identifying the right technique

Using computers to analyse dataReporting the analysis



The Data Analysis The Data Analysis ProcessProcess

Questionnaire Editing

Generating Composite Measures

Editing the SPSS Output

Transfer Edited SPSS output to

Word File

Data Entry & Data Definition

Data Validation

Testing Goodness of

Measure

Identifying the Right

Technique

Syntax Command to

SPSS

Interpreting the SPSS

Output

Report Writing

In most cases, the process is not sequential and linear, but more iterative where you

may need to go back to some steps to explore further



Data Analysis - TypesData Analysis - Types

Purpose of StudyExploratory; Test of Differences; Establishing Relationships

Number of VariablesUnivariate; Bivariate; Multivariate

Level of MeasurementsNominal, Ordinal, Interval & Ratio



Purpose of Study - Purpose of Study - ExploratoryExploratory Basically to address the issue of “What is the

current situation – vis-à-vis the phenomenon of interest?”

Example:1. What is the current level of innovation of Malaysian

Companies?2. What is the preferences of Malaysian consumers of

organic food?3. What percentage of the Malaysian consumers will

use internet banking?Etc…..



Purpose of Study – Test Purpose of Study – Test of Differencesof Differences To address the question of “Is there a difference

between Type A and Type B in terms the phenomenon of interest, Y?”

Example:1. Are companies in the electronics sector more

innovative than companies in the food sector?2. Are consumers in the professional line more likely to

use internet banking compared to those in other occupation?

3. Is the preference for organic food higher for urban consumers compared to consumers living in the rural areas?



Purpose of Study – Purpose of Study – Establishing Establishing RelationshipsRelationships To address the question of “Is there a

relationship between phenomenon X with phenomenon, Y?”

Example:1. The larger the company, the more innovative it is?2. The more business a company undertakes the more likely is the

company to use internet banking?3. The more the consumer believe that organic food is good for the

health, the more likely he/she is to consume organic food.4. The more secure is internet banking, the more user-friendly is

the system, and the easier to access the system, the more likely a company is to use internet banking.



Number of Variables - Number of Variables - UnivariateUnivariate Analysis involves only one variable Exploratory studies typically involve univariate

analysis Example:

– What is the current level of innovation of Malaysian Companies?

– What is the preferences of Malaysian consumers of organic food?

– What percentage of the Malaysian consumers will use internet banking?



Number of Variables - Number of Variables - BivariateBivariate Analysis involves two variables Test of differences involves two variables Establishing relationships may involve bivariate

analysis (correlation) Example:

1. Are companies in the electronics sector more innovative than companies in the food sector? – sector & level of innovation

2. The more business a company undertakes the more likely is the company to use internet banking? – size (in terms of business transactions) & likelihood of using internet banking



Number of Variables - Number of Variables - MultivariateMultivariate Analysis involves three or more variables Analysis typical when trying to establish

relationships Example:

1. The more secure is internet banking, the more user-friendly is the system, and the easier to access the system, the more likely a company is to use internet banking. – security, user-friendliness, ease of access, and likelihood of using internet banking



Data Types – Levels of Data Types – Levels of Measurement Measurement

Purpose of StudyExploratory; Test of Differences; Establishing Relationships

Number of VariablesUnivariate; Bivariate; Multivariate

Level of MeasurementsNominal, Ordinal, Interval & Ratio



Data Types - Levels of Data Types - Levels of MeasurementsMeasurements

Empirical Scale

Basic Operations

Measures of Typical use

Averages

Nominal Determination of equality

Classification Male-Female Occupations

Mode

Ordinal Determination of greater or less

Ranking Preference Attitude

Median

Interval Determination of equality of intervals

Index numbers Temperature

Mean

Ratio Determination of equality of ratios

Sales Unit produced No. of customers

Mean Geometric



Nominal ScaleNominal Scale

Classificatory or Categorical

e.g. Sex – Male/Female

Colour

Mutually exclusive & collectively exhaustive



Ordinal ScaleOrdinal ScaleCategorize and rank

e.g. Preference in job attributes

Please rank from 1 most important to 5 least important the following attributes:

Interacting with others

Using multiple skills

Completing a task from beginning to end

Serving others

Work independently



Interval ScaleInterval ScaleAllows for measurement of distance between two points on the scalee.g. Preference in job attributes

Using a scale of 1 (strongly disagree), 2 (disagree), 3 (neither agree nor disagree), 4 (agree) and 5 (strongly agree), please indicate the extent of your agreement by circling the appropriate number.

The following are very important to me

Interacting with others 1 2 3 4 5

Using multiple skills 1 2 3 4 5

Complete a task from beginning to end 1 2 3 4 5

Serving others 1 2 3 4 5

Working independently 1 2 3 4 5



Ratio ScaleRatio Scale

Has absolute zero; thus allowing for not only differences but also proportions in the differences

e.g. Number of years in the organization



The Right Technique?The Right Technique?Research QuestionResearch QuestionConcern for Central Tendency; Concern for Central Tendency;

Comparing groups; RelationshipsComparing groups; Relationships

Number of VariablesNumber of VariablesUnivariate; Bivariate; Univariate; Bivariate; MultivariateMultivariate

Level of MeasurementsLevel of MeasurementsParametric and Non-parametricParametric and Non-parametric



The Right Technique?The Right Technique?What is the purpose of the analysis?What is the purpose of the analysis?

What is the level of measurement?What is the level of measurement?

How many variables are involved?How many variables are involved?



Descriptive AnalysisDescriptive AnalysisPurpose: To describe the distribution of the variables of interest

Answers the question of “What is ...?”

TechniquesFrequencies Distribution - if 1 ordinal or nominal variable, Cross-tabulation - if 2 ordinal or nominal variables Means - if 1 interval or ratio level variable Means of subgroups - if 1 interval or ratio level variable by subgroups



Test of DifferencesTest of DifferencesPurposePurpose: : To evaluate the differences between 2 or more groups To evaluate the differences between 2 or more groups

with respect to a variable of interestwith respect to a variable of interest

TechniquesTechniques depends on depends on

Levels of Measurement of the Levels of Measurement of the VariableVariableNumber of GroupsNumber of GroupsIndependence of the GroupsIndependence of the Groups



Test of DifferencesTest of DifferencesMore than 2

group?

Are they independent

?

Are they independent

?

Nominal: 2-test

Ordinal: Mann-

Whitney

Continuous: t-test

Yes

Nominal- McNemar

Ordinal - Wilcoxon

Signed Rank

Continuous: Paired t-test

No

No Yes

Nominal: 2 test

Ordinal: Kruskal-Wallis ANOVA

Continuous: 1-way ANOVA

Nominal: 2 - test

Ordinal: Friedman 2-way

ANOVA

Continuous: Factorial 2-way

ANOVA

Yes No



RelationshipRelationship

Purpose:Purpose: To establish relationship between To establish relationship between variablesvariables

TechniquesTechniques depends on depends on

Whether or not there exist Whether or not there exist dependent variable(s)dependent variable(s)Number of dependent and Number of dependent and independent variablesindependent variablesLevels of Measurement of the Levels of Measurement of the VariableVariable



Dependence Dependence RelationshipsRelationships

How many dependent variables?

Scale of Dependent

Multiple Regression

ANOVA Discriminant Analysis

Canonical Correlatio

n

Scale of Independe

nt

Scale of Dependent

Scale of Independe

nt

Multivariate ANOVA

More than 1

One

Conjoint Analysis

Scale of independe

nt

Interval

Nominal

Interval Nominal

Scale of Independe

ntIntervalNominal

Interval

Interval

Nominal

muhamad jantan: [email protected] to data analysis1 the research process observation broad...

Documents