research methodlogy

441
Course: Research Methodology ( MGT 602)

Upload: kashif-niazi

Post on 17-Sep-2015

18 views

Category:

Documents


2 download

DESCRIPTION

Research Methodology

TRANSCRIPT

  • Course:

    Research Methodology( MGT 602)

  • Instructor Ayyaz Mahmood Assistant Professor at CIIT BS,MBA,MS, PhD(thesis under evaluation) 12 years of teaching at University and 9 years of

    Industry experience. Supervised a number of MBA and MS thesis. Published papers and attended conferences.

  • Introduction

    Overview of the course :Business research is an organized and deliberateprocess through which organization effectively learnnew knowledge and help improve performance.

  • Introduction

    Objectives of the course : To understand and develop a systematicapproach to business research

    To emphasis on the relationship betweentheory , research and practice

    To Integrate different research activities in anorderly fashion

  • Outcomes of the course are : To formulate research questions Develop theoretical framework Develop hypotheses Learn to select from different researchmethodologies

    Develop skills for data analysis andinterpretation.

  • Research Methodology

    COURSE OUTLINE:Course Intro Building blocks of science in research [1]Broad problem area , Preliminary Information Survey [ 2]Literature Review [2]Literature Review[2]Theoretical Framework [3]Theoretical Framework[3]Hypothesis Development [3]Hypothesis Development[3]Elements of Research Design (purpose, investigation type, researcher interference, study setting)[4]Elements of Research Design (unit of analysis, time horizon, Measurement of variables) [4,6]Measurement of variables (operational definition) [6]Measurement of Variable (Scales) [6]Validity and Reliability [6]Data Collection Methods (Interviews , Questionnaire) [7]Data Collection Methods (Questionnaire, observation) [7]Sampling (Probability Sampling) [8]Sampling (Non Probability Sampling) [8]Experimental Design [5]Refresher on Statistical Terms [9]Introduction to SPSSIntroduction to SPSSData Analysis and Interpretation ( Getting data ready for analysis) [10]Data Analysis and Interpretation (Feel for Data, Testing the goodness of Data) [10]Data Analysis and Interpretation (Descriptive Statistics)[10]Data Analysis and Interpretation (Inferential Statistics( Person Correlation, Hypothesis Testing t-test, ANOVA, Chi Square)[10]Data Analysis and Interpretation (Inferential Statistics( Hypothesis Testing, Multiple Regression) [10]Data Analysis and Interpretation (Inferential Statistics( Mediation, Moderation, Rank) (Hand out)Research Report contents (Sample report)

  • Business Research ScenariosA. A manager observes that the customers are not pleased

    Are my customers satisfied from my product/service ?B. It is observed that hydro construction project projects

    tend to have a low successes rate.What could be reasons behind it. ?

    C. The new product introduced is not doing so well.Have we selected the right market, features or price ?

    For all the above scenarios management needs to findreliable and creditable information to understand the issueand then take appropriate decisions in order to achieveperformance

  • InformationReducesUncertainty

    I dont knowif we

    shouldreduce our

    product prices?

  • Define Business Research

    Business research is defined as the systematic andobjective process of gathering, recording andanalyzing data for aid in making business decisions.

    Research information is neither intuitive norhaphazardly gathered.

    Literally, research (re-search) -search again Business research must be objective Detached and impersonal rather than biased It facilitates the managerial decision process for all

    aspects of a business

  • Research Methods Is the way in which research studies are

    designs procedures by which data is collected are analyzed. We would be focusing on the survey methodologyin which the research is conducted by collectingdata and analyzing them to come up with answersto various issues of interest.

    The different areas of problem could be related toFinance, Accounting, HR, Marketing etc.

  • Types of Research Two purpose of research are

    To solve a currently exiting problem in the worksetting (Applied Research )

    To add to the general body of knowledge (Basic Research)

  • Applied research is when research is done with theintention of applying the results of its findings tosolving specific problem currently being experiencedin the organization

    e.g. To improve the attendance at an X organization A transport service can be introduced, Hasflextime improve the employee performance at auniversity)

  • Basic research done mainly to improve ourunderstanding of certain problems that arecommonly occur in organizational setting and how tosolve them

    e.g. increase the productivity of clerical workers inservice industry,

    increase the effectiveness of project orientedbusiness

  • Research Philosophy and Choices

  • Important assumptions about the way in which oneviews the world.

    These assumptions effect the research strategy andthe methods you choose and practicalconsiderations.

    Researcher concerned with facts, such as theresources needed for manufacturing will havedifferent view on the way research

  • Researcher concerned with the feelings and attitudesof the workers towards their managers in that samemanufacturing process.

    Their strategies and methods probably will differconsiderably and what is important and significant

  • Philosophy of Choices

    Deductive Develop a theory and hypothesis (or hypotheses)and design a research strategy to test thehypothesis

    Inductive Collect data and develop theory as a result of yourdata analysis

  • Characteristics of Good Research

    Purposive: Definite aim (Help reduce turnover,absenteeism, complete projects on time )

    Rigor: Sound methodological design, systematic andscientific. Avoid individual biases. (Managerinterviews few employee on their preference for flexitime and device policy)

    Testability: After properly selecting thecases/respondents and collection of data logicallydeveloped hypothesis statements can be testedusing statistical tests.

  • Replicability: Applying the same method the findingfrom more than one study suggest the same results.

    Precision and Confidence: Study of the wholeuniverse of item, events or population of interest isnot possible. But we try to come close to reality aspossible (precision)and also be confident of ourfindings that they are correct (confidence).

    Objectivity: The interpretation of the results shouldbe based on facts, not on our own subjective feeling

  • Generalizability: Applicability of the finding on avariety of firms/organization

    Parsimony: Simplicity in explaining the phenomena ispreferred, rather than managing many factors andtheir effect (45% variability is explained by 4variables and 48% variability is explained by 10variables)

    Management and Behavioral science result are not100% scientific or exact. We deal with measuringsubjective feelings , attitudes, perceptions. Meetingall the characteristics of good research is difficult

  • Hypothetic-Deductive Method of Research1. Observation2. Preliminary Information gathering3. Theory formulation4. Hypothesizing5. Further data collection6. Data Analysis7. Deduction

  • Observation

    One senses certain changes are occurring New behaviors are surfacing in an environment When one considers the situation important then

    move to the next step E.g. Customers are not pleased as they used to be.Are customers at the store are grumbling orcomplaining.

  • Preliminary Information Gathering

    Know more about what has been observed Talk to more people about it( other employees,

    customers) Know what is happening is happening and why.

    E.g. Talk to customers if they are happy with theproduct or service. The customer might be happywith the products but the problem is that therequired products are out of stock and salesperson are not helpful. The salesman input on thisissues reveals that the factory does not deliver ontime so in order to satisfy the customer thesalesmen communicates different delivery dates.

  • Hypothesizing

    Some testable or educated supposition are made E.g. If sufficient inventory is made customers wouldbe less dissatisfied customers

    Accurate and timely information of the delivery tothe sales person can also reduce the dissatisfiedcustomer.

  • Further Scientific Data Collection

    Data with respect to each variable in the hypothesisneed to be obtained.

    E.g. Measure the current level of customer satisfactionand measure the satisfaction level when the stocksare made readily available.

    Measure the current level of accuracy ofinformation to sales person on the stock and thesatisfaction level of customer and then measurethem again once the level of information hasincreased.

  • Data Analysis

    Data gathered statistically is analyzed and see if thehypothesis have been supported or not. E.g.

    Do an correlation analysis of the tow factorslike level of information and satisfaction.

  • Deduction

    Arriving at a conclusion by interpreting the meaningof the results of the data analysis. E.g. If the customer satisfaction has increase by certainamount when the availability of information andthe stock.

    We could recommend that these two factorsinfluence the satisfaction of the customers

  • Recap lecture We tried to examine what research is? Research Philosophies and choices We considered the two types of research Hall Marks of Research (Purposive, rigor, testable, replicabilty,

    precision and confidence, objectivity , genralizability and parsimony) The seven steps of hypothetic deductive research

    method1. Observation2. Preliminary Information gathering3. Theory formulation4. Hypothesizing5. Further data collection6. Data Analysis7. Deduction

  • Research Methodology

    Lecture No : 4(Theoretical Framework)

  • Recap

    Literature Review involves searching anddocumenting

    There are different formats of Documenting(APA)

    There is a structure of review (importance,objectives, definitions, relationships identified,gaps)

    Theoretical framework is representation of yourbelief on how variables related and why

    Variables are of 4 different kinds

  • Theoretical Framework After conducting literature review, survey and

    defining the problem (research questions) We develop our theoretical framework Theoretical framework is a conceptual model of how

    we theorarize the relationships among severalfactors that have been identified to the problem. Problem is depleting sales Factors influencing are quality of products, price,competition etc ( based on the literature)

  • Based on the previous literature we discuss theinterrelationship between the different variableswhich are of interest to us and concerns theproblem.

    By developing this kind of conceptual frameworkwould help us claim and test certain relationships.

    i.e. From this framework we develop hypothesisstatements which are then tested to find out if ourtheory was valid or not

  • Sales

    Quality

    Price

    Competition

  • Types of Variables

    Dependent (Criterion Variable) primary interest Describe or explain the variability or predict it. We study what variables influence dependent

    variable So by studying these we might able to find a

    solution of the problem E.g. Sales are low , employee loyalty is

    dropping

  • Independent (Predictor variable) Which influences the dependent variable The influence might be positive or negative When independent variable is present the

    dependent variable is also present. With each unit of increase in independent

    variable there is an increase or decrease inthe dependent variable

    E.g. Advertising on sales, recognition onloyalty

  • Exercise : List the independent variable

    A manager believes that good supervision andtraining would increase the production level ofthe workers.

  • Moderating Variables

    Moderating Variables have strong contingent(conditional) effect on the independent dependentvariables relationship.

    i.e. in the presences of the a third variable therelationship between the independent anddependent is modified

  • Distinction between Independent andModerating Variable

    Some times one gets confused as to when a variableis to be treated as independent variable and when itbecomes a moderating variable

  • Situation A

    Willingness tolearn new ways

    Quality ofTrainingPrograms

    GrowthNeed ofemployee

  • Situation B

    Willingnessto learnnew ways

    Quality ofTraining Prog

    GrowthNeed

    High/Low

  • Both the scenarios have 3 variables First scenario training programs and growthneeds are independent variables thatinfluence the dependent variable

    Second scenario dependent variable stays thesame growth need becomes the moderatingvariables

  • i.e. only those who have high growth need willbecome more willing to learn new thingswhen quality of the trainings is increased.

    Hence the relationship between dependentand independent variable become contingent(conditional) on the existence of themoderator.

  • The linear effect of training and growth need onwillingness

  • The effect of training is contingent on high/low growthneed (slope/intensity)

  • Mediating/Intervening

    A variable which surfaces between the time theindependent variable operates to influence thedependent variable.

    Temporal /sequential quality Surfaces as a function of the independent variable

  • Exam diff ExamPerformance

  • ExamDifficulty Stress

    ExamPerformance

  • WorkforceDiversity

    OrganizationEffectiveness

  • Integrating Moderating, Mediating Variables

  • Theoretical Framework Is a conceptual model Foundation of the research Logically developed, described and elaborated

    network of association as a result of interviews,observation and literature survey. So we identify a problem Identify the important variables from literature etc. Logically developing network of associations and elaborate Generate hypotheses and later tested

  • Components of Theoretical Framework

    Identification of variables ( name and type) Discussion how and why these variables are related Direction of the relationship need to be theorized

    and discussed (positive/negative) Discussion on why these relationships exists, support

    from previous research. A schematic diagram

    Note: Must read example on page 93

  • Recap

    Types of Variables Independent, Dependent, Moderating, Mediating(Intervening)

    Examples of their relationships with eachother

    Developing of Theoretical Framework Variables, logical Relationships, Directions,Explanations

  • Research Methodology

    Lecture No : 8(Research Design-continue)

  • Recap

    We covered some of the research design elements We talked about the research purpose

    (exploratory, descriptive, hypothesis testing) Type of investigation

    (causal, correlations) Extent of researcher's interference

    (High,moderate,low)

  • Study Setting: Contrived and Non-contrived

    Organizational research can be done in the naturalenvironment where work proceeds normally (i.e., innon-contrived setting) or in artificial, contrivedsettings.

    Correlation studies are invariably conducted in non-contrived settings, whereas rigorous causal studiesare done in contrived lab setting

  • Correlation studies done in organizations are calledfield studies ( factors influencing in a call center its employeesturn over ).

  • Studies to establish cause and effect relationshipsusing the same natural environment in whichemployees normally function are called fieldexperiments

    Example: employees who have been given recognition and

    employee who have not been given recognition.

  • Cause effect studies in contrived environment inwhich The environment extraneous factors arecontrolled are termed as lab experiments.

  • Example: Select all new employees with the same scores in the

    entry test and provide one group training and theother no training and controlling that they are notexposed to any senior employee who could guidethem.)

  • Unit of Analysis: Individuals, Dyads, Groups,Organizations, Cultures

    The unit of analysis refers to the level of aggregationof the data collected during the subsequent dataanalysis stages.

  • Individuals: If the problem statement focuses on howto raise the motivational levels of employees ingeneral, then we are interested in individualemployees in the organization and would like to findout what we can do to raise their motivation.

  • Here the unit of analysis is the individual.(managersperception on the factors which influence thesuccess of the project)

  • Dyads: If the researcher is interested in studying two-person interactions, then several two-person groups,is known as dyads and will become unit of analysis.

  • For example, analysis of husband-wife(are theysatisfied with the education provided by the school)in families and mentor-mentee (perception on thebenefit of mentoring).

  • Groups: If the problem statement is related to groupeffectiveness, however, then obviously the unit ofanalysis would be at group level.

  • For example, if we wish to study group decision-making patterns, we would probably examining suchaspects as group size, group structure, cohesiveness,and the like, in trying to explain the variance in groupdecision making.

    In such cases the unit of analysis will be groups.(useof I.T by the different department)

  • Organizations: If we compare different departmentsin the organization, then the data analysis will bedone at the departmental level - that is, theindividuals in the department will be treated as oneunit and comparison made treating the departmentas the unit of analysis.

    (Conservation of energy initiatives by public andprivate organization)

  • Cultures: If we want to study cultural differencesamong nations, we will have to collect data fromdifferent countries and study the underlying patternsof culture in each country, here the unit of analysisused will be cultures.

    (Moral values of Eastern vs Western cultures)

  • Time Horizon: Cross-sectional versus Longitudinal

    Cross-Sectional Studies A study can be done in which data are gathered just

    once, perhaps over a period of days or weeks ormonths, in order to answer a research question. Suchstudies are called one-shot or cross-sectional studies.

    (data collected from project managers and theirpsychological well being between October tillDecember)

  • Longitudinal Studies In some cases, the researcher might want tostudy people or phenomena at more than onepoint in time in order to answer the researchquestion. For example, the researcher mightwant to study employees behavior before andafter a change in the top management, tolearn the effects of change.

  • Or when data on the dependent variable aregathered at two or more points in time to answer theresearch question, are called longitudinal studies.(use of electricity by a city in summers and then inwinters)

  • Scenarios

    Following are some scenarios , for each indicate howresearcher should proceed, giving reasons:

    1. Purpose of the study2. Type of investigation3. Researcher Interference4. Study setting5. Time Horizon6. Unit of analysis

  • Recap

    Research Design elements Study setting Time Horizon Unit of analysis Secnarios

  • Research Methodology

    Lecture No : 9(Measurement of Variables/Operational Definition)

    1

  • Recap

    Research Design elements Study setting Time Horizon Unit of analysis

    2

  • Measurement of Variables

    In order to find answers to our question and in orderto test our hypothesis we need measure ourvariables of concern.

    3

  • Why the need for measuring

    To test the hypothesis the variables need tomeasured.

    Finding the answers to our questions is possiblewhen we have some statistics/ numbers .

    Some variables are easily measurable e.g. Height,salary, hours worked.

    Some are not so easily measured motivation level,success level of projects, satisfaction, loyalty etc.

    4

  • Questions like1. How long have you been working in this

    organization?2. What is your marital status ?3. How much is your salary ?4. What was the cost of last project ? But some variables are abstract and subjective e.g.

    satisfaction, happiness, achievement motivation,effectiveness of the organization.

    5

  • One cannot simply ask what is the achievementmotivation level of your employees.

    But before we start measuring the variables itsabstractness needs to be addressed.

    There are ways to in which the abstractness of thenotion could be simplified into observablecharacteristics.

    6

  • For instance Thirst cannot be seen but we expectthat a thirsty person would consume lots of liquid.

    Hence the behavior of the thirsty person is that hewould drink fluids.

    If several individuals say they are thirsty we canmeasure thirst by measuring their consumption ofliquid, although the concept itself is abstract.

    7

  • Reducing abstract concepts so they are measurable iscalled operationalizing.

    Operationally defining a concept so that it becomesmeasurable is achieved by looking at the behavioraldimensions, facets , or properties represent by theconcept.

    8

  • Steps to Operationalization

    one needs the define component of theconcept.

    Under each concept possible quantitativemeasurable elements need to identified.

    Against each developed concepts specificquestions could be formulated. The questionscould be supported by secondary data,observation or self report

    9

  • Operational Definition

    10

  • The operational definition of Learning could bestated as The ability to recall the lesson, it is alsothe ability apply the lesson learned to practicalsituation and finally it is the understanding of alesson.

    Even though these dimension have to an extentreduced some of the ambiguity but we still need tofurther classify what is meant by understanding,application so that we can measure learning as awhole.

    11

  • With some effort we can define what is meant byunderstanding , i.e. the ability to answer questionscorrectly and give appropriate answers. We alsodefine what is application, which is the ability tosolve problem by applying the lesson learned andintegrate it with other relevant material.

    Now we are in better position to measure theconcept learning.

    At this stage we can develop questions whichaddress the synthesized concepts and obtain dataon them.

    12

  • 13

  • 14

  • 15

  • 16

  • 17

  • What is not operational definition

    It does not define the correlates of a concept i.e. motivation and performance are two separate

    concepts and they might be correlated we cannotsubstitute one with the other

    Motivation can lead to performance but we do notmeasure performance by motivation.

    We need to differentiate between the reasons(factors/antecedents) with dimensions.

    Dimensions are the sub components of a conceptand factors/ antecedents the causes of the concept

    18

  • We operationally define concepts and ask questionsthat are likely to measure the concept.

    So for abstract concepts we need to define thesubjective feelings and attitudes.

    For straight forward variables , objective data is usedsuch as salary, number of tee shirts.

    A number of subjective concepts have beenopertionalized by the subject experts and we can usethem for research.

    19

  • Recap

    Measurement is necessary to give answers or to theresearch question , or to test our hypotheses.

    The opeationalizing of certain subjective variablesare necessary for measurement.

    The abstract concepts are broken down todimensions and its elements.

    Questions are formulated on them Not to confuse dimensions with antecedents

    20

  • Research Methodology

    Lecture No : 10(Measurement of Variables/Scales)

    1

  • Recap

    Measurement is necessary to give answers or to theresearch question , or to test our hypotheses.

    The opeationalizing of certain subjective variablesare necessary for measurement.

    The abstract concepts are broken down todimensions and its elements.

    Questions are formulated on them Not to confuse dimensions with antecedents

    2

  • Scales and Measurement

    We have operationalized the concepts and convertedthem into dimensions and elements

    We also have attached questions with theseelements against which we would collect some data.

    Each question needs to measured

    3

  • Measurement is the process of assigning numbers orlabels to objects, persons, states of nature, or events.

    Done according to set of rules that reflect qualities orquantities of what is being measured.

    4

  • Measurement means that scales are used.

    Scales are a set of symbols or numbers, assigned byrule to individuals, their behaviors, or attributesassociated with them

    5

  • Types of Scales Four types of scales are used in research, each with

    specific applications and properties. The scales are Nominal Ordinal Interval Ratio

    6

  • Nominal Scale: Simply the Nominal scale is count of the objectsbelonging to different categories. Ordinal Scale: The ordinal scale positions objects in some order ( such as it indicates that pineapples are juicer thenapples and oranges are even more juicer thanpineapples)

    7

  • Interval Scale: It can gives us information as to what extent(level)one is juicer than the other. How much better is the pineapple than the appleand orange is better than the pine apple. Is pine apple only marginally better than the apple . Ratio Scale: It is most comprehensive scale, has all characteristicsof other scales.

    8

  • Nominal Scales Nominal scales are used to classify objects,

    individuals, groups, or even phenomena.

    Examples of nominal variables: Gender State of residence Country Ethnicity

    9

  • Nominal scales are mutually exclusive (meaning that those items being classified will fit

    into one classification).

    These scales are also collectively exhaustive,(meaning that every element being classified can fitinto the scale).

    10

  • As it might appear on a questionnaire, examples ofnominally scaled questions included:

    What is your class rank at CIIT?1.Freshman 3. Junior2.Sophomore 4. Senior

    11

  • The numbers themselves do not have meaning(we could have used letters, too),

    They are used just to identify the possibleresponses to the question.

    Thus, in evaluating responses to this you cannotuse the mean.

    Permitted statistics; frequencies (% and counts,modes )

    12

  • Nominal scale is always used for obtaining personaldata such as gender or department in which oneworks, where grouping of individuals or objects isuseful, as shown below.1. Your gender 2. Your department

    ___Male ___Production___Female ___Sales

    ___Accounting___Finance___Personnel___R & D___Other (specify)

    13

  • Ordinal Scales These scales allow for labeling (or categorization) as

    in nominal scales, but they also allow for ranking. Example: Rate these vacation destinations in terms

    of how much you would like to visit from one to fivewith one your most preferred and five your leastpreferred.1. Bermuda2. Florida3. Hawaii4. Aspen5. London

    14

  • This type of scale can provide information aboutsome item having more or less of an attribute thanothers, but no information on the degree of this.

    Permitted statistics: Frequencies, median, mode

    15

  • Ordinal scale is used to rank the preferences or usageof various brands of a product by the individuals andto ranks order individuals, objects, or events as perthe examples below.

    16

  • Rank the following personnel computers with respecttheir usage in your office, assigning the number 1 tothe most used system, 2 to the next most used, andso on. If a particular system is not used at all, in youroffice, put a 0 next to it.

    ____Apple ____Hewlett Packard____Compaq ____IBM____Comp USA ____Packard Bell____Dell Computer ____Sony____Gateway ____Toshiba

    17

  • Interval Scales

    Contains the information available in ordinal scales(ranking) but with the added benefit of magnitude ofranking.

    Interval scales have equal distances between thepoints of a scale.

    These scales can contain a zero point, but they aresubjective and are not meaningful (0 C = 32 F).Temperature is an example of a interval scale

    Permitted statistics; mean, median, mode, as well asmore advanced tests.

    18

  • 18/5/2015 19

    On a scale of one to five, with five meaning you stronglyagree, and one meaning you strongly disagree consider thisstatement I believe my college education has prepared mewell to begin my career.

    1 2 3 4 5

    Stronglydisagree

    Somewhat

    disagreeNeither

    Somewhatagree

    Stronglyagree

  • Ratio Scale

    The most comprehensive scale Has all of the characteristics of the other three with

    the additional benefit of an absolute, meaningfulzero point.

    Examples include: Weight Sales volume Income Age

    Permitted statistics same as with interval data.20

  • A ratio variable, has all the properties of an intervalvariable, and also has a clear definition of 0.0. Whenthe variable equals 0.0, there is none of that variable.Variables like height, weight, enzyme activity areratio variables.

    21

  • Temperature, expressed in F or C, is not a ratiovariable. A temperature of 0.0 on either of thosescales does not mean 'no temperature'.

    However, temperature in Kelvin is a ratio variable, as0.0 Kelvin really does mean 'no temperature'.

    22

  • 18/5/2015 23

    Ratio scales are usually used in organizationresearch when exact numbers on objective asopposed to subjective factors are called for, as inthe following question: How many other organizations did you work forbefore Date joining this system?

  • Please indicate the number of children you have ineach of the following categories?---- below 3 years---- between 3 and 6---- over 6 years but under 12---- 12 years and over

    How many retail outlets do you operate?

    24

  • Comparison between scales

    The researcher would like to know what is thepercentage of people who like Pepsi, 7up, Coke,Miranda?

    Choose the soft Drink you want to order.Pepsi7UpCokeMarinda

    25

  • The researcher would like to know among the 4 softdinks which they prefer the most ,assigning 1 to mostand 4 to the least

    Pepsi7UpCokeMarinda

    26

  • The researcher would like to know what extent the 4drinks are liked

    On a scale of one to five, with five meaning youstrongly like, and one meaning you strongly dislikeconsider this statement I like/dislike this soft drink .

    Pepsi 1 2 3 4 5 Coke 1 2 3 4 5 7up 1 2 3 4 5 Marinda 1 2 3 4 5

    27

  • The researcher would like to know how manyPepsi , Mrindia , etc you consume in a monthPepsi: _____7Up: _____Coke: _____Marinda:_____

    28

  • 29

  • 30

    Very badBadNeither good nor badGoodVery good

    PoorFairGoodVery goodExcellent

    How good a car is Honda?

    Balanced or Unbalanced

  • 31

    Very badBadNeither good nor badGoodVery good

    Very badBadNeither good nor badGoodVery goodNo opinionDont know

    Forced or Unforced Choices

    How good a car is Honda?

  • Rating Scales

    32

  • 33

    I plan to purchase a laptop in the 12 months.

    YesNo

    Simple Category (Dichotomous) Scale

  • 34

    What newspaper do you read most often for financial news?

    East City GazetteWest City TribuneRegional newspaperNational newspaperOther (specify:_____________)

    Multiple-Choice, Single Response Scale

  • 35

    What sources did you use when designing your new home?Please check all that apply.

    Online planning servicesMagazinesIndependent contractor/builderDesignerArchitectOther (specify:_____________)

    Multiple-Choice, Multiple Response Scale

  • 36

    The Internet is superior to traditional libraries forcomprehensive searches.

    Strongly disagreeDisagreeNeither agree nor disagreeAgreeStrongly agree

    Likert Scale

  • 37

    Semantic Differential

    A measure of attitudes that consists of a series of seven-point ratingscales that use bipolar adjectives to anchor the beginning and end of

    each scale.

  • 38

    Numerical Scale

    An attitude rating scale similar to a semantic differential except that it usesnumbers, instead of verbal descriptions, as response options to identify

    response positions.

  • 39

    Stapel Scales

    A measure of attitudes that consists of a single adjective in the center ofan even number of numerical values.

  • 40

    Constant-Sum Scales

    A measure of attitudes in which respondents are asked to divide a constantsum to indicate the relative importance of attributes; respondents often sort

    cards, but the task may also be a rating task.

  • 41

    Graphic Rating Scales

    A measure of attitude that allows respondents to rate an object by choosing anypoint along a graphic continuum.

  • Research Methodology

    Lecture No : 11(Goodness Of Measures)

    1

  • Recap

    Measurement is the process of assigning numbers orlabels to objects, persons, states of nature, or events.

    Scales are a set of symbols or numbers, assigned byrule to individuals, their behaviors, or attributesassociated with them

    2

  • 3

  • Using these scales we complete the development ofour instrument.

    It is to bee seen if these instruments accurately andmeasure the concept.

    4

  • Sources of Measurement Differences

    Why do scores vary? Among the reasons legitimatedifferences, are differences due to error (systematic orrandom)1. That there is a true difference in what is beingmeasured.2. That there are differences in stable characteristics ofindividual respondents

    On satisfaction measures, there are systematicdifferences in response based on the age of therespondent.

    18/5/2015 5

  • 3.Differences due to short term personal factors moodswings, fatigue, time constraints, or other transistoryfactors.Example telephone survey of same person, differencemay be due to these factors (tired versus refreshed)may cause differences in measurement.

    4.Differences due to situational factors calling whensomeone may be distracted by something versus fullattention.

    18/5/2015 6

  • 5.Differences resulting from variations inadministering the survey voice inflection, nonverbal communication, etc.

    Differences due to the sampling of items included inthe questionnaire.

    7

  • 7. Differences due to a lack of clarity in measurementinstrument(measurement instrument error).Example; unclear or ambiguous questions.

    8. Differences due to mechanical or instrument factors blurred questionnaires, bad phone connections.

    18/5/2015 8

  • Goodness of Measure

    Once we have operationalized, and assigned scaleswe want to make sure that these instrumentsdeveloped measure the concept accurately andappropriately.

    Measure what is suppose to be measured Measure as well as possible

    18/5/2015 9

  • Validity : checks as to how well an instrument that isdeveloped measured the concept

    Reliability: checks how consistently an instrumentmeasures

    10

  • 11

  • Ways to Check for ReliabilityHow to check for reliability of measurement instruments

    or the stability of measures and internal consistencyof measures?

    Two methods are discussed to check the stability .1. Stability

    (a) Test Retest Use the same instrument, administer the test

    shortly after the first time, taking measurement inas close to the original conditions as possible, tothe same participants.

    18/5/2015 12

  • If there are few differences in scores between thetwo tests, then the instrument is stable. Theinstrument has shown test-retest reliability.

    Problems with this approach. Difficult to get cooperation a second time Respondents may have learned from the first

    test, and thus responses are altered Other factors may be present to alter results

    (environment, etc.)

    13

  • (b)Equivalent Form Reliability This approach attempts to overcome some of the

    problems associated with the test-retestmeasurement of reliability.

    Two questionnaires, designed to measure the samething, are administered to the same group on twoseparate occasions (recommended interval is twoweeks).

    18/5/2015 14

  • If the scores obtained from these tests arecorrelated, then the instruments have equivalentform reliability.

    Tough to create two distinct forms that areequivalent.

    An impractical method (as with test-retest) andnot used often in applied research.

    15

  • (2)Internal Consistency Reliability

    This is a test of the consistency of respondentsanswer to all the items in a measure . The itemsshould hang together as a set.

    i.e. the items are independent measures of thesame concept, they will correlated with one another

    18/5/2015 16

  • Developing questions on the Concept Enriched Job

  • Validity

    Definition: Whether what was intended to bemeasured was actually measured?

    18/5/2015 18

  • Face Validity The weakest form of validity Researcher simply looks at the measurement

    instrument and concludes that it will measure whatis intended.

    Thus it is by definition subjective.

    18/5/2015 19

  • Content Validity

    The degree to which the instrument items representthe universe of the concepts under study.

    In English: did the measurement instrument cover allaspects of the topic at hand?

    18/5/2015 20

  • Criterion Related Validity The degree to which the measurement instrument

    can predict a variable known as the criterionvariable.

    18/5/2015 21

  • Two subcategories of criterion related validity Predictive Validity

    Is the ability of the test or measure to differentiateamong individuals with reference to a futurecriterion.

    E.g. an instrument which is suppose to measurethe aptitude of an individual, when used can becompared with the future job performance of adifferent individual. Good performance (Actual)should also have scored high in the aptitude testand vise versa 22

  • Concurrent Validity Is established when the scale discriminatesindividuals who are known to be different that isthey should score differently on the test.

    E.g. individuals who are happy at availing welfareand individuals who prefer to do job must scoredifferently on a scale/ instrument which measureswork ethics.

  • Construct Validity Does the measurement conform to some underlying

    theoretical expectations. If so then the measure hasconstruct validity.

    i.e. If we are measuring consumer attitudes aboutproduct purchases then do the measure adhere tothe constructs of consumer behavior theory.

    This is the territory of academic researchers

    18/5/2015 24

  • Two approaches are used to measure constructvalidity

    Convergent Validity A high degree of correlation among 2 differentmeasures intended to measure same construct

    Discriminant Validity The degree of low correlation among varaiblesthat are assumed to be different.

    18/5/2015 25

  • To check validity through Correlation analysis, FactorAnalysis, Multi trait , Multi matrix correlation etc

    26

  • Reflective vs Formative measure scales: In some multi item measure where it is measuring

    different dimensions of a concept do not hangtogether

    Such is the case of Job Description Index measurewhich measures job satisfaction from 5 differentdimension i.e Regular Promotions, Fairly goodchance for promotion, Income adequate, HighlyPaid, good opportunity for accomplishment.

    27

  • In this case some items of dimensions Incomeadequate and Highly paid to be correlated butdimension items of Opportunity for Advancementand Highly Paid might not correlated.

    In this measure not all the items would related toeach other as its dimensions address differentaspect of job satisfaction.

    This measure /scale is termed as Formative scale

    28

  • In some cases the measure dimensions and itemscorrelate.

    In this kind of measure/scale the differentdimensions share a common basis ( commoninterest)

    An example is of a scale on Attitude towards theOffer scale.

    Since the items are all focused on the price of anitem, all the items are related hence this scale istermed as Reflective Scale.

    29

  • Recap

    30

  • Research Methodology

    Lecture No : 12(Data Collection-Interview)

    1

  • Recap

    2

  • Primary Data Primary Data = information obtained exclusively for

    current research Personal Interview Focus Groups Panels Delphi Technique Telephone Interview Computer assisted telephone

    interviewing and Computer administered telephonesurvey

    Self-Administered Surveys

  • Secondary Data

    Company Archives Gov Publications Industry Analysis

  • Primary Data Collection Methods

    Focus Group Panels Interviews (face to face, telephone, electronic media) Questionnaires (personally, mail, electronic) Observation Other (projective tests)

  • Focus Group: Usually consist of 8 to 10 members , with a

    moderator leading the discussion for 2 hours on aparticular topic, concept or product.

    Member are chosen on the bases of their expertiseon the topic.

    E.g Discussion on computers and computing , orwomen mothers , social networking etc

    Less expensive and usually done for exploratoryinformation. Cannot be generalized

    6

  • Panels: Similar to focus group but meets more than once in

    order to study the change or interventions need tobe studies over a period of time.

    Members are randomly chosen E.g effect of advertisement of a certain brand need

    to be assessed quickly, panel members could beexposed to the advertisement and intention ofpurchase could be assessed.

    When the product is modified then the response ofthe panel can be observed 7

  • Observation measures: Methods through which primary data is collected

    without the involving people. E.g: Wear and tear of books , section of an office,

    seating area of railway station which indicate thepopularity, frequency of use etc.

    E.g: The number of cans in the dust bin and theirbrands, the number of motor cycles vs cars parked inthe university parking lot

    8

  • Interviewing: Collect data from the respondent on an issue of

    interest. Usually administered at the exploratory stage of

    the research. In case large set of respondents are needed then

    more than one interviewer are used , hence theyneed to be trained so that biases , voiceinflections, difference in wording are avoided

    Structured and Unstructured

  • Un Structured: No planned sequence of questions, help in exploring

    preliminary issues.

    e.g. Tell me something about your unit and department, and perhaps even the organization as a whole interms of work, employee and whatever else youthink is importantCompared to other departments, what are thestrengths and weakness of your department

    10

  • In case they identify a difference you can ask How can you improve the situation ?

    Encouraging the respondent to reflect on the positiveand negative aspects of it.

    Try to pleasant and see if the respondent is notcomfortable.

    11

  • Through unstructured the different major areasmight be exposed. It from these the researcher canpick some areas as focus variables which needfurther probing.

    Now the researcher can device a more focusedapproach and develop a more structured interviewemphasizing on some particular issues.

    12

  • Structured: Know at the outset what information is needed.

    Focusing on factors relevant to the problem. The focus is on the factors which have surfaced

    during the un structured interview.

    13

  • E.g: During the previous unstructured interview itwas identified that the department needsimprovement.

    Now you can focus on questions which addresseshow to improve the department, i.e. the factorswhich can improve the department

    14

  • This can be done through face to face, over thetelephone or through the computers via internet.

    Specific same questions are asked from differentrespondents.

    The information collected is tabulated and then thedata is analyzed.

    15

  • The result could highlight the important factorsinfluencing the issues.

    This information is of qualitative in nature whichcould be then empirically tested and verified usingother methods like questionnaires.

    16

  • Guideline for Interviews

    Listen carefully Motivate the respondents How to take notes Built proper trust and rapport with interviewee Clarification of complex issues Physical setting Explaining the reasons for research and criteria of

    selection

  • Face to Face Adv :Clarify doubts, repeating, rephrasing, gettingnon verbal cues Dis : vast resources required, cost, anonymity

  • Telephone: Adv : Wider reach in short time, some time easy todiscuss personal information over the phone Dis: Can be terminated without warning, cannot havea prolonged interview, non verbal cue.

    19

  • Closed vs. Open Questions Easy. Cost of coding is reduced. Quicker, standardized interviews. Can be answered without thinking. Pre-testing is a must. Limit the richness of data.

  • Recap

    The Data is collected from primary and secondarysources

    The primary data collect via Observation, panels, interviews, questionnaires etc Interview are structure and unstructured While interviewing there are certain guidelines There are structured and unstructed interviews There are some advantages / disadvantages of face

    to face vs telephone interviews .21

  • Research Methodology

    Lecture No : 13(Data Collection-Questionnaire)

    1

  • Recap

    The Data is collected from primary and secondarysources

    The primary data collect via Observation, panels, interviews, questionnaires etc Interview are structure and unstructured While interviewing there are certain guidelines There are structured and unstructed interviews There are some advantages / disadvantages of face

    to face vs telephone interviews .2

  • Questionnaires

    Data Collection is mechanism when theresearcher knows exactly what is required andhow to measure the variables of interest. Types of Questionnaire:

    Personally administered questionnaire Mail Questionnaire

  • Personally Administered Questionnaires Mostly local area based, org is willing to have a groupof employee respond to it. It is Cheaper then interviews, helps remove doubts,motivating respondents

  • Mail Questionnaires: Wide geographical area can be reached, respondentshave flexibility of time , It is more cost effective butthe response rate is low, Can improve by giving some incentives and doubtscannot be clarified.

  • Guidelines for Principles

    Content and purpose of question (Subjective/Objective) Language and wording( Jargon/ Technical) Type and Form (open ended, closed ended) Positively and Negatively Worded Biases ( loaded, leading, social desirable, double barreled) Sequencing of Questions

  • Content and purpose of Question: If the variables tapped are subjective feeling we

    need to measure the dimension and elements ..Use interval scales

    If the variables are objectives/ facts a singledirect question may be asked.

  • Language and Wording of question:

    The level of respondents have to be considered.Slang and Technical jargon has to be avoided

    e.g. Work is a drag, she is a compulsive worker. TechJargon like organizational structure , 360 degreeappraisal

  • Type and Form of Questions: Open ended vs Closed Ended Positively vs Negatively Worded

  • Open ended vs Closed Ended

    In open ended the respondent chooses any way theylike. E.g. any five things which interest him at his job. In close ended the respondent have to make achoice among the given alternatives e.g. out of thelist of 10 job characteristics rank any 5

  • Positively vs Negatively Worded :

    Have some positive and some negative wordedquestions to break the monotony. E.g. Coming to work is great fun or coming to work isno great fun

  • Biases in Questions:

    Double Barreled:Questions has more than one question within it.

    E.g. Do you think that the course content isadequate and it applicable at your work?

  • Ambiguous Question:Respondent does not know what it means. E.g. To whatextent would you say you are happy?Do you discuss you work with your boss regularly? Doyou go to movies frequently?Frequently may mean once in a week, or once in amonth. Regularly may mean every day, or every week ,or every month.

  • Recall Dependent: Questions based on past experiences and rely onmemory. E.g. After 30 years of work one would notremember the first job details such as name ofthe boss/ years worked in a department

  • Leading Questions: Are worded in such a way that it would lead the

    respondent to answer in a way that theresearcher would like to or want to give.

    E.g. Dont you think that in these days ofescalating costs of living employee should begiven good pay raise?

    Better.. To what extent do you agree thatemployee should be given higher pay raise.

  • Example: Dont you think that more women should be

    promoted to decision making line positions inorganization

  • Loaded Questions: Are when they are phrased in an emotionally

    charged manner.

    E.g. To what extent do you think management islikely to be vindictive/(cruel) if the union decides togo on strike.

    Better. To what extent you favor strike To whatextent you fear that there would be a adversereaction from the management.

  • Did P.T.I Lose the elections in Punjab

    Better P.T.I was not chosen in Punjab

  • Social Desirability:Is when questions are worded such that theyelicit(draw out) socially desirable responsee.g. Do you think that older people should be laid off?..better There are advantages and disadvantages to retainingsenior citizens in the workforce. To what extent do youthink companies should continue to keep the elderlyon their payroll.

  • Exercise

    If you have been in the company for fifteen yearsplease indicate the year of joining or the name of youcolleague.

    Bad question as it is recall dependent

  • My colleague is good and efficient .

    Bad Question: Double Barreled

  • Working Women should not have children.

    Bad: Loaded question an emotional issue for women

  • Investment in children's future should be animportant goal of the administration.

    Bad Question: Socially desirability

  • This job uses a lot of skills I have.

    Okay no problem with the wordings

  • For this country to keep on remaining competitiveshould we not spend more on research.

    Bad: Leading question

  • Other Guide lines

    Length (20 words) Sequencing (funneling , same positive andnegative question)

    Classification Data or Personal Data

  • Recap Questionnaires Personally Administered Questionnaires Mail questionnaires Guide line for wordings

    Content and purpose (Subjective vs Objective) Language and wording ( slang/technical) Types of formats (open / closed ended) Positively worded and Negatively worded Bias/ Favoritism(Leading, loaded, ambiguous, doublebarrel, socially desirable)

    Length of the question Funneling

  • Research Methodology

    Lecture No :14(Sampling Design)

  • Recap Data collection Interviews and Questionnaires Personally Administered Questionnaires Mail questionnaires Guide line for wordings

    Content and purpose (Subjective vs Objective) Language and wording ( slang/technical) Types of formats (open / closed ended) Positively worded and Negatively worded Bias/ Favoritism(Leading, loaded, ambiguous, doublebarrel, socially desirable)

    Length of the question Funneling

  • Lecture Objectives

    Define sampling, reasons for sampling, sample,population, element, sampling unit and subject

    Sampling process Different sampling design

  • Sampling

    The process of selecting the right individuals,objects, or events as representative of entirepopulation is known as sampling.

    PopulationPopulation

    SampleSample

  • Relationship between sample andpopulation

  • Reasons for Sampling

    Budget and time Constraints (in case of largepopulations)

    High degree of accuracy and reliability (if sampleis representative of population)

    Sampling may sometimes produce moreaccurate results than taking a census as in thelatter, there are more risks for makinginterviewer and other errors due to the highvolume of persons contacted and the number ofcensus takers, some of whom may not be well-trained

  • Population

    It refers to the entire group of people, events orthings of interest that the researcher wishes toinvestigate.Example: If regulators want to know how patientsin nursing homes run by a company in France arecared for, then all the patients in all the nursinghomes run by them will from the population.

  • Element

    An element is a single member of a populationExample: If 1000 blue collar workers(laborworkers) in a particular organization happen to bethe population of interest to a researcher, eachblue collar worker therein is an element.

  • Sample

    A sample is a subset or subgroup of thepopulation. By studying the sample, the researchershould be able to draw conclusions that aregeneralizable to the population of interest.

    Example: If there are 145 in-patients in a hospitaland 40 of them are to be surveyed by the hospitaladministrator to access their level of satisfactionwith the treatment received, then these 40members will be the sample.

  • Sampling unit

    It is the element or set of elements that is availablefor selection in some stage of sampling process.

    Example: Sampling units in a multistage sampleare city blocks, households, and individuals withinthe households.

  • Subject

    It is a single member of the sample, just as anelement is a single member of the population.

    Example: If a sample of 50 machines from a totalof 500 machines is to inspected, then everyone ofthe 50 machines is a subject, just as every singlemachine in the population of total population of500 machines is an element

  • Parameters

    The characteristics of the population such as thepopulation mean, the population standarddeviation, and the population variance are referredto as its parameters.

    Example: Average weight, , of all 30 year oldwomen in Australia, % of voters, p, in N.S.W whothink the Government is doing a good job tocontrol inflation.

  • The Sampling Process

    Sampling is the process of selecting a sufficientnumber of right elements from the population so,the major steps in the sampling include.1. Defining the population2. Determine the sample process3. Determine the sampling design4. Determine the appropriate sample size5. Execute the sampling process

  • The Sampling Process

  • Defining the populationSampling begins with precisely defining the targetpopulation. The target population must be definedin terms of elements, geographical boundaries andtime.

    Example: A target population may be, for example,all faculty members in the Department ofManagement Sciences in the V-COMSATSnetwork,All housewives in Islamabad,All pre-college students in Rawalpindi,

  • The target group should be clearly defined ifpossible, for example, do all pre-collegestudents include only primary and secondarystudents or also students in other specializededucational institutions?

  • Determining the sample frameThe sampling frame is a (physical) representationof all the elements in the population from which hesample is drawn. Also termed as a List.

    Often, the list does not include the entirepopulation. The discrepancy is often a source oferror associated with the selection of the sample(sampling frame error)

    Information relating to sampling frames can beobtained from commercial organizations

  • Example: Student telephone directory (for thestudent population), the list of companies on thestock exchange, the directory of medical doctorsand specialists, the yellow pages (for businesses)

  • Determining the sample design

    Two major types of sampling Probability samplingThe elements in the population have some known,non zero chances or probability of being selectedas sample subjects. Non probability samplingThe elements do not have a known orpredetermined chance of being selected assubjects.

  • Factors affecting sampling design

    The relevant target population of focus to thestudy

    The parameters we are interested ininvestigating

    The kind of sample frame is available

    Costs and Time are attached to the sampledesign and collection of Data

  • Determining the sample size

    The decision about the how large the sample sizeshould be can be very difficult one. These factorsaffecting the sampling decision are The research objective The extent of precision desired(the confidence

    interval) The acceptance risk in predicting that level of

    precision(confidence level) The amount of variability in the population itself The cost and time constraints In some cases, the size of population itself

  • Executing the sample process

    In this final stage of sampling process, decisionwith respect to thethe target population,the sampling frame,the sample technique, andthe sample size have to be implemented.

  • Example: A young researcher was investigating the

    antecedents of salesperson performance.

    To examine his hypotheses, data were collectedfrom the chief sales executive in the Pakistan(the target population) via mail questionnaire.

  • The sample was initially drawn from thepublished business register (the samplingframe), but supplemented with respondentrecommendations and other additions, in ajudgment sampling methodology.

    The questionnaires were subsequentlydistributed to sales executives of 450 companies(the sample size).

  • Non response and non response errors

    A failure to obtain information from a number ofsubjects included in the sample

    Those who do respond to your survey aredifferent from those who did not on (one of the)characteristics of interest in your study

    Two important sources of non response errorsare not at homes and refusals

  • Reducing the rate of refusals The rate of refusals depends, among other

    things, on the length of the survey, the datacollection method and the backing of research.

    Decrease in survey length, personalinterviews/questionnaire instead of mailquestionnaire and the sponsorship of theresearch often improve the overall return rate.

  • Recap

    Sampling is the process of selecting the rightindividuals

    Sample is used to represent the whole data orpopulation

    Sampling process include defining population,sample frame, sampling design, sample sizeand sampling process

  • Research Methodology

    Lecture No :15(Sampling Design / Probability vs Non probility)

  • Recap

    Sampling is the process of selecting the rightindividuals

    Sample is used to represent the whole data orpopulation

    Sampling process include defining population,sample frame, sampling design, sample sizeand sampling process

  • Lecture Objectives

    Differentiate between probability and nonprobability sampling

    Learn about the types of probability sampling, itsadvantages and disadvantages

    Learn about the types of non probabilitysampling, its advantages and disadvantages

    Issues relevant to sample design and collection

  • Probability SamplingUnrestricted or simple random sampling

    Technique which ensures that each element inthe population has an equal chance of beingselected for the sample.

    The simple random sampling is the least biasand offer the most generalizability.

  • Probability Sampling

    The major advantage of simple randomsampling is its simplicity.

    The sampling process could becomecumbersome and expensive.

    Example: Choosing raffle tickets from a drum,computer-generated selections, random-digittelephone dialing.

  • Simple random sampling

  • Probability Sampling

    Restricted or complex probability sampling:

    It is an alternate to simple random samplingdesign, several complex probability samplingdesigns can be used.

    Efficiency is improved in that more informationcan be obtained for a given sample size usingthe complex probability sampling procedures.

  • Probability Sampling

    The most common complex probability samplingdesign1. Systematic sampling2. Stratified sampling3. Cluster sampling

    1. Area sampling

    4. Double sampling

  • Probability SamplingSystematic Sampling: Technique in which an initial starting point is

    selected by a random process, after which everynth number on the list is selected to constitutepart of the sample.

  • Sampling interval (SI) = population list size (N)divided by a pre-determined sample size (n)

    How to draw: 1) calculate SI, say (200/20)=10 2) select a number between 1 and SI randomly, i.e. 1-10 3) go to this number as the starting point and the item on the list

    here is the first in the sample, e.g 3 4) add SI to the position number of this item and the new position

    will be the second sampled item, e.g 3+10=13 5) continue this process until desired sample size is reached.

  • For systematic sampling to work best, the listshould be random in nature and not have someunderlying systematic pattern.

    E.g: Office directory with the Senior Manager,Middle manager .names are listed in eachdepartment. This can create as systematicproblem

  • Probability SamplingStratified Sampling: Technique in which simple random subsamples

    are drawn from within different strata that sharesome common characteristic. Within the groupthey are homogenous and among the groupthey are heterogeneous.

  • Probability Sampling

    Stratified SamplingExample: The student body of CIIT is divided intotwo groups (management science, engineering)and from each group, students are selected for asample using simple random sampling in each ofthe two groups, whereby the size of the sample foreach group is determined by that groups overallstrength.

  • Probability Sampling

    Cluster Sampling Technique in which the target population is first

    divided into clusters. Then, a random sample ofclusters is drawn and for each selected clustereither all the elements or a sample of elementsare included in the sample.

    Cluster samples offer more heterogeneity withingroups and more homogeneity among groups

  • Probability Sampling

    Area samplingSpecific type of cluster sampling in which clustersconsist of geographic areas such as counties, cityblocks, or particular boundaries within a locality. Area sampling is less expensive than most other

    sampling designs and it is not dependent onsampling frame.

    Key motivation in cluster sampling is costreduction.

  • Probability Sampling

    Area samplingExample: A city map showing the blocks of the cityis adequate information to allow the researcher totake a sample of the blocks and obtain data fromthe resident therein.Example: If you wanted to survey the residents ofthe city, you would get a city map, take a sample ofcity blocks and select respondents within each cityblock.

  • Probability Sampling

    Single stage and multistage cluster sampling Single stage cluster sampling involves the

    division of population into convenient clusters,randomly choosing the required number ofclusters as sample subjects, and investigating allthe elements in each of the randomly chosenclusters

    Cluster sampling can also be done in severalstages and is then known as multistage clustersampling.

  • Probability Sampling

    Example: If we were to do a national survey of theaverage monthly bank deposits, cluster samplingwould be used to select the urban, semi urban andrural geographical location for study. At the nextstage particular areas in each of these locationswould be chosen. At the third stage, banks withineach area would be chosen.Example:

  • Probability Sampling

    Double sampling: A sampling design where initially a sample is

    used in a study to collect some preliminaryinformation of interest, and later a subsample ofthis primary sample is use to examine the matterin more detail.

  • Probability Sampling

    Double samplingExample: A structured interview might indicate thata subgroup of respondents has more insight intothe problems of the organization. Theserespondents might be interviewed again and againand asked additional questions.

  • Non-Probability Sampling

    Convenience Sampling: Sampling technique which selects those

    sampling units most conveniently available at acertain point in, or over a period, of time.

  • Non-Probability Sampling

    Convenience Sampling: Major advantages of convenience sampling is

    that is quick, convenient and economical; amajor disadvantage is that the sample may notbe representative.

    Convenience sampling is best used for thepurpose of exploratory research andsupplemented subsequently with probabilitysampling.

  • Non-Probability Sampling

    Judgment (purposive) Sampling: Sampling technique in which the business

    researcher selects the sample based onjudgment about some appropriate characteristicof the sample members.

    Example: Selection of certain students who areactive in the university activities to inquire aboutthe sports and recreation facilities at the university.

  • Recap Simple random sampling and restricted

    sampling are two basic types of probabilitysampling.

    Probability ( Simple Random, Systematic,Cluster, Single stage/multistage, Doublesampling)

    Non Probability (Convenience, judgment)

  • Research Methodology

    Lecture No :16( Sampling / Non Probability, Confidence and Precision, Sample size)

  • Recap Lecture

    Systematic ,stratified sampling, cluster, area anddouble sampling are the common types ofcomplex sampling.

    Convenience, judgment, quota and snowballsampling are the common types of nonprobability sampling.

  • Lecture Objectives

    Non Probability Based sampling (Quota/snowball)

    Discuss about the precision and the confidence.

    Precision and Confidence

    Factors to be taken into consideration fordetermining sample size.

    Managerial implications of sampling.

  • Non-Probability SamplingQuota Sampling:This is a sampling technique in which the businessresearcher ensures that certain characteristics of apopulation are represented in the sample to anextent which is he or she desires.

  • Non-Probability Sampling

    Quota SamplingExample: A business researcher wants to determinethrough interview, the demand for Product X in adistrict which is very diverse in terms of its ethniccomposition.

    If the sample size is to consist of 100 units, thenumber of individuals from each ethnic groupinterviewed should correspond to the groupspercentage composition of the total population of thatdistrict.

  • Quota Sampling

    Example: Quotas havebeen set for gender only.Under thecircumstances, its nosurprise that the sampleis representative of thepopulation only in termsof gender, not in terms ofrace. Interviewers areonly human;.

  • Non-Probability SamplingSnowball Sampling : This is a sampling technique in which individuals

    or organizations are selected first by probabilitymethods, and then additional respondents areidentified based on information provided by thefirst group of respondents

  • Non-Probability Sampling

    Snowball Sampling The advantage of snowball sampling is that smaller

    sample sizes and costs are necessary; a majordisadvantage is that the second group ofrespondents suggested by the first group may bevery similar and not representative of the populationwith that characteristic.

    Example: Through a sample of 500 individuals, 20antique car enthusiasts are identified which, in turn,identify a number of other antique car enthuiasts

  • More Snowball SamplingMore systematic versions of snowball sampling canreduce the potential for bias. For example,respondent-driven sampling gives financialincentives to respondents to recruit peers.

  • Issues in Sample Design and Selection

    Availability of Information Often information onpotential sample participants in the form of lists,directories etc. is unavailable (especially indeveloping countries) which makes somesampling techniques (e.g. systematic sampling)impossible to undertake.

  • Resources Time, money and individual orinstitutional capacity are very importantconsiderations due to the limitation on them.Often, these resources must be traded againstaccuracy.

  • Issues in Sample Design and Selection

    Geographical Considerations The number anddispersion of population elements maydetermine the sampling technique used (e.g.cluster sampling).

    Statistical Analysis This should be performedonly on samples which have been createdthrough probability sampling (i.e. not probabilitysampling).

    Accuracy Samples should be representative ofthe target population (less accuracy is requiredfor exploratory research than for conclusiveresearch projects).

  • Issues of precision and confidence indetermining sample size

    Precision Precision is how close our estimate is to the true

    population characteristic. Precision is the function of the range of

    variability in the sampling distribution of thesample mean.

  • Population and Sample distinctiveness

    Sample Statistics( Mean, Std Deviation, Variance) andPopulation parameters ( Mean, Std Deviation,Variance)

    Compare the Sample estimates and populationcharacteristic. Where the estimates should be therepresentative of the population charactertics

    Sample statistics (mean, sd, ..) should berepresentative of the population parameters(mean,sd )

  • Issues of precision and confidence indetermining sample size

    Precision:How close are the estimates to the population.

    While expecting that the population mean would itfall between (+,- )10 points or (+,-) 5 points basedon the sample estimates is precision.

    The narrower the more precise our statement is

  • E.g: The average age of the a particular classbased on the sample is between 20 and 25Or it between 18 and 28.

    How close are the estimates to the population.

  • Confidence Confidence denotes how certain we are that our

    estimate will hold true for the population. The level of confidence can range from 0 to

    100%. However 95% confidence is theconventionally accepted for most businessresearch.

  • The more we want to be precise the less confidentwe become that our statement is going to be true.

    So at one level we want to be accurate in ourstatement but on the other we taking a higher risk ofproved incorrect.

    In order to maintain the precision and increase theconfidence or increase the precision and theconfidence we need to have a larger sample.

  • Determining sample size

    Roscoe (1975) proposes the following rules ofthumb for determining sample size.

    Sample sizes larger than 30 and less than 500are appropriate for most research

    Where sample sizes are broken into subsamples(males/females, juniors/seniors etc.), a minimumsample size of 30 for each category isnecessary.

  • Determining sample size

    In multivariate research (including multipleregression analysis), the sample size should beseveral times (preferably ten times or more)as large as the number of variables in thestudy.

    For simple experimental research with tightexperimental controls (matched pairs, etc.),successful research is possible with samples assmall as 10 to 20 in size.

  • Tools and mathematical equations are availableto establish the right size of the sample.

    Refer to the book for the sample size calculationequation.

    Standard Tables are available

    Use a software like RAO calculator available onthe internet.

  • Types of Sampling Designs

    Sampling Designs

    Non-probability Probability

    Convenience Judgmental Quota Snowball

    Systematic Stratified Cluster Other SamplingTechniques

    SimpleRandom

  • Managerial Implications

    Awareness of sampling designs and sample sizehelps managers to understand why a particularof sampling is used by researchers.

    It also facilitates understanding of the costimplications of different designs, and the tradeoff between precision and confidence vis--visthe costs.

  • Managerial Implications

    This enables managers to understand the riskthey take in implementing changes based on theresults of the research study.

    By reading journal articles, this knowledge alsohelps managers to assess the generazibility ofthe findings and analyze the implications oftrying out the recommendations made therein intheir own system.

  • Recap Non Probability based sampling ( Precision we estimate the population parameter

    to fall within a range, based on sample estimate. Confidence is the certainty that our estimate will

    hold true for the population. Roscoe (1975) rules of thumb for determining

    sample size. Some sampling designs are more efficient than

    the others. The knowledge about sampling is used for

    different managerial implications.

  • Research Methodology

    Lecture No : 3

  • Recap lecture 2 Broad problem: the entire situation where

    one sees a need for problem solving andresearch.

    Literature Review: To understand theproblem more detail information is needed

    Different sources of information is gatheredfrom books, reports, published researchpapers etc.

  • Recap lecture 2

    More aspects of research are exposed. More variables which play an important role

    are uncovered. This allows us to develop a more robust

    theory. We start documenting a comprehensive

    review To conclude we are able to identify the gaps

    and develop our precise problemsstatements.

  • Lecture Objective

    Revisit at literature review Why literature review is important Methods of writing a review Contents of a review

    Developing a theoretical framework Identify the relationships and the theory

    supporting the relationship Describing Variables

    Describing what are variables and thedifferent types

  • Purpose of Literature Review

    Every research study requires the researcher toreview pertinent literature on the topic.

    1. To avoid unnecessary duplication of research. 2. To identify variables that may influence the

    problem 3. To identify promising procedures and

    instruments 4. To limit the problem.

  • Two steps in conduction literature review

    Survey of literature (search) Documenting of the literature (write)

  • Survey of literature

    Survey different sources Books Research Articles Theses Conference preceding

    You can obtain them from Libraries Internet Online databases (Full text, abstract)

  • Documenting the literature

    Three activities are involved whiledocumenting the literature which you havesurveyed Method of documenting the list of reviewedarticles. (Modes)

    Referencing and quoting the studies (Cite) Organizing and documenting the contents of thereviewed articles (writing the review)

  • Method of documenting the list ofreviewed articles.

    References / Bibliography is a list of work that isrelevant to the main topic arranged in analphabetical order.

    The difference between reference list andbibliography is that reference list is a subset ofthe list of articles which have been referenced inthe research.

    Bibliography is a list which includes all thereferenced and non referenced articles in yourresearch but are relevant to your research

  • Examples of Modes of Reference listingThere are different modes of referencing in businessresearch. For example the APA (Publication of Manualof the American Psychological Association), ChicagoManual Style, Harvard style, Turabian Style.Each manual specifies with examples how books,newspaper, research journal are to be referenced inyour research. Following are the example in APA style

  • Referencing and quoting the studies

    Cite the references in the body of the paperusing author-year method of citation; i.e.surname of author(s) and the year ofpublications

    E.g. Kaleem(2004) has shown. In the recent studies of employee motivation

    (Freeman,2007 ;Mitnzberg, 2007) it has In 1997, Kyle compared the different models of

    motivation.. As pointed out by (Tucker & Snell, 1989),..

  • Referencing and quoting the studies(cont )

  • Organizing and documenting thecontents of the reviewed articles

    While writing the review the text needs to arrangedin the following manner

    1. Introduction - Importance of the subject , states the purpose or scope of the review

    2. Define the key concepts What are the different definitions found in the

    literature. Which definition is better or much closeryour research objective.

  • Organization of a LiteratureReview:

    3. Critical review - Describe the relationships between the different

    variables identified in the previous studies Do not list one study after another, but rather

    classify, compare & contrast as they relate to yourproblem statement.

    Organize the review around different themes. 4. Summarize

    states the status of what exists on the topic andidentifies the gaps which provide the rationale foryour study.

  • Example of a short Review

    Pg 44

  • Example of a short review

    Introduction to Organization effectiveness Identified the problem and the purpose

    No consensus on the how to conceptualizeand measure OE

    Summarize the previous work and identify thegaps in the literature Variables from different streams related to the

    OE uncovered Leading to the forming of the research

    questions

  • Questions What could be the dimensions used for

    measuring OE ? What factors effect the OE ?

    Once the research questions have been statedthen one is ready to develop a theoretical framework of their research

  • While developing your theoretical frame workyou basically

    Theorize on the bases of your belief that howare certain phenomena's are related.

    So theoretical framework is a representation ofyour beliefs on how certain phenomena ( orvariables or concepts ) are related to eachother(model) and an explanation of why youbelieve that these are associated with eachother (theory)

  • Theoretical Framework

    So there are two components to theoreticalframe work Identification of variables and their

    relationship Describing the relationship with arguments

    While identifying the different variables we needto differentiate between the different kinds ofvariables

  • Variables

    Any thing that can take on different or varyingvalues is a variable

    Values can be different at various times for thesame object or person or at the same time fordifferent objects or persons E.g. Production units (Employee 1 (10 units on Monday) Production

    units (Employee 1 (11 units on Tuesday) Production units (Employee 2 (12 units on Monday) Production units (Employee 2 (10 units on Tuesday) Attendance at department x on Monday(10), Tuesday(2)

  • Types of Variables

    Independent Dependent Moderating Mediating

  • Types of Variables

    Dependent (Criterion Variable) primary interest Describe or explain the variability or predict it. We study what variables influence dependent

    variable So by studying these we might able to find a

    solution of the problem E.g. Sales are low , employee loyalty is

    dropping

  • Independent (Predictor variable) Which influences the dependent variable The influence might be positive or negative When independent variable is present the

    dependent variable is also present. With each unit of increase in independent

    variable there is an increase or decrease inthe dependent variable

    E.g. Advertising on sales, recognition onloyalty

  • Moderating (surfaces in between theindependent and dependent at a given time)

    Mediating (Effects the relationship betweenindependent and dependent)

  • Exercise : List the independent variable

    A manager believes that good supervision andtraining would increase the production level ofthe workers.

  • Recap

    Literature Review involves searching anddocumenting

    There are different formats of Documenting(APA)

    There is a structure of review (importance,objectives, definitions, relationships identified,gaps)

    Theoretical framework is representation of yourbelief on how variables related and why

    Variables are of 4 different kinds

  • Research Methodology

    Lecture No : 5

    (Theoretical Framework - Hypothesis Development)

  • Recap

    Types of Variables Independent, Dependent, Moderating, Mediating(Intervening)

    Examples of relationships with each other Developing of Theoretical Framework

    Variables, logical Relationships, Directions,Explanations

    We wanted to break down a problem intoeasily measurable into testable cases.

  • Exercise

    A production manager is concerned about the lowoutput levels of his employee. The articles that hereads on job performance frequently mentionedthree variables as important to job performance: skillrequired by job, rewards and satisfaction. In severalof the articles it was also indicated that only if therewards were attractive to the recipients, didsatisfaction, and job performance increase nototherwise.

  • Theoretical Framework( Description and Discussion of the Variables)

    In this section of theoretical framework we need toprovide the description of the variables and theirrelationships with different variables. For example..

    Rewards are two types, intrinsic and extrinsic ..,where as job enrichment is making the job morechallenging and utilizes all the skills of theemployeewhen the.. . Rewards are known toenhance the satisfaction of employees which leads tohigher organization performance But for someemployees the rewards are not attractive hence doesnot contribute to the satisfaction of employee .etc

  • Theoretical Framework(Schematic Diagram)

    JobEnrichment

    Rewards

    EmployeeSatisfaction

    OrganizationPerformance

    Attractionfor

    rewards

  • Research Questions

    Does job enrichment and rewards influence theperformance ?

    Does the satisfaction intervenes the relationshipbetween rewards and performance?

    Does the satisfaction intervenes the relationshipbetween job enrichment and performance?

    Does attractiveness of the rewards moderate therelationship between rewards and satisfaction.

  • Hypotheses Development

    The research problem could be better solved whenwe formulate the appropriate research questions.

    The logically placed relationships need to be tested. So we develop statements which would be easily

    testable Formulating such testable statements is calledhypothesis development.

  • Hypothesis Statements

    A hypothesis can be defined as a logically speculatedrelationship between two or more variablesexpressed in the form of a test able statement.

    Different Hypotheses statements can be drawn fromthe theoretical framework developed earlier.

    E.g. Ha1: Job Enrichment leads to higher job satisfaction Ha2: If rewards are offered the job satisfaction level be high Ha3: Organization performance is effected by job enrichment

    through satisfaction

  • The logical relationships have been now stated in atestable format.

    We need to statistically examine the relationshipbetween the variables Rewards and satisfactionor Job Enrichment and Satisfaction

    We need to also statistically establish that thesatisfaction mediates the relationship betweenrewards , job enrichment and organizationperformance

  • We need to statistically see if there is positivecorrelation between these variables is significant(large enough) then we would state that thehypotheses have been substantiated(proved)

    In social sciences we call a relationship statisticallysignificant when we are confident that 95 times outof 100, the observed relationship will hold true.

  • It is through data analysis our logical relationshipsare tested.

    In case our hypothesis are not proved then we wouldsearch for possible reasons. May be some othervariables which influence the relationship e.g. somemoderating variables.

    It is again the literature which can provide us withthe directions. Hence a good literature review isimportant.

  • Hypotheses Statement Formats

    Hypotheses statements could be to test Difference between groups Relationship between variables

    The statements could be in the shape of Proposition (suggestion) If-then Else statement

    Theses statements could be direction or nondirectional

  • Examples of different formats of Hypothesesstatements

    Difference between groups There is difference between the motivation level of menand women

    Relationship between variables There is a relationship between age and job satisfaction

  • Proposition style Employees who are more healthy wil