Table of Contents 1.
Epistemological Epistemological and ontological ontological positions ...................... ................................ .......... ..................... ................................ ........... .................. .................. 1
Research Research designs.................... .............................. ............ .. ...................... ................................ .......... .................... ................................ ............ .................... ............................... ............. 3 1.1
Ethical principles principles..................... ................................ ........... .................... .............................. ............ .. ........................ ................................ ........ ................. ................. 5
Quantitative Quantitative r esearch ........................ ................................ ........ ..................... ............................... ............ ....................... ................................ ......... .................... ........................ .... 6 1.2
Reliability Reliability and validity validity ..................... ................................ ........... .................... .............................. ............ .. ........................ ................................ ........ ......... 7
Sampling ................. ........................... ............... ..... ........................ ................................ ........ ..................... ................................ ........... ..................... ............................... ............ ............ ............ 8 1.3
Sample size ..................... ............................... ............ ...................... ................................ .......... .................... ................................ ............ .................... ......................... ..... 9
1.4
Non-probability Non-probability sampling..................... ................................ ........... .................... .............................. ............ .. ........................ ................................ ........ .... 9
Structured Structured interviews interviews ............ ....................... .................... ......... .................... .............................. ............ .. ........................ ................................ ........ ..................... ....................... .. 10 Coding .............. ......................... .................. ....... ....................... ................................ ......... .................... ................................ ............ .................... ............................... ............. ...... 10 1.5
Problems with structured structured interviewing interviewing ..................... ............................... ............ .................... ................................ ............ ............. ............. 12
Self-complet Self-completion ion questionnaires questionnaires.................. ............................... ............... ..................... ............................... ............ ....................... ................................ ......... ........ 12 1.5.1
Steps to improve response rates ..................... ............................... ............ .................... .............................. ............ .. ............... ............... 13
Asking questions ............. ....................... ................... ......... ...................... ................................ .......... .................... ................................ ............ .................... .............................. .......... 13 1.6
Types of questions questions ..................... ................................ ........... .................... ............................... ............. ........................ ................................ ........ ............ ............ 14
1.7
Piloting and pre-testing pre-testing questions...................... ................................ .......... .................... .............................. ............ .. .................... ...................... 15
Structured Structured observation observation ......................... ................................ ....... .................... ............................... ............. ........................ ................................ ........ .................... .................... 16 Secondary Analysis and official official statistics statistics ...................... ................................ .......... ..................... ................................ ........... ..................... .......................... ..... 17 Qualitative Qualitative research ............. ....................... ................... ......... .................... ............................... ............. ........................ ................................ ........ .................... ......................... ..... 18 1.11
Main research methods associated with qualitative research.............. research............................. .................. ... ............ .......... .. 18
1.12
Main steps in qualitative qualitative research......... research..................... ...................... ............ .................... .............................. ............ .. .................... ...................... 19
1.12.1 1.13
Grounded theory .................... ................................ ............ ..................... ................................ ........... ....................... ................................ ......... ...... 19
Reliability Reliability and validity validity in qualitative qualitative research..................... ............................... ............ ....................... ................................ ......... .... 19
1.13.1
Alternative Alternative criteria for evaluating qualitative research .................... ............................... ............. .............. .............. 19
1. Epi Epistemolog stemologic ical al and ontologic ontological al pos osiit ions Empiricism Knowledge gained through experience Deduction Theory -> Hypothesis -> Data collection -> Findings -> Hypothesis confirmed/rejected -> Revision of theory Induction Observations/findings -> Theory
Epistemological positions Positivism Advocates application of natural science methods to study of social reality. y
Only knowledge confirmed by senses can be warranted as knowledge (phenomenalism)
y
Purpose of theory is to generate hypotheses that can be tested (deductivism)
y
Knowledge is arrived at through the gathering of facts that provide the basis for laws (inductivism)
y
Science must be conducted value free (objective)
Realism Natural and social science should apply the same approach to data collection/explanation (same as positivism). There is an external reality. y
Empirical realism Through the use of appropriate methods, reality can be understood.
y
Critical realism We will only be able to understand and change the social world if we identify the structures at work that generate those events and discourses.
Interpretivism Verstehen, hermeneutic-phenomenological tradition and symbolic interactionism. Difference between people and the objects of the natural sciences must be respected. Scientist must grasp subjective meaning of social action. Ontological
positions
Objectivism Social phenomena confront us as external facts that are beyond our reach or influence (social phenomena are independent of social actors). Constructionism (constructivism) Social phenomena and their meanings are continually being accomplished by social actors.
Paradigms Objectivist There is an external viewpoint which the organization can be viewed from. Subjectivist An organization is a socially constructed product. Can only be understood by the point of view of individuals. Purposes of scientific research within paradigms. Regulatory Purpose of business research is to describe what goes on in organizations. Radical Make judgments how organizations should be and how to achieve this. Functionalist Framework based on problem-solving orientation which leads to rational explanation. Interpretative Questions whether organizations exist beyond social actors, thus it must be explained through their experience. Radical humanist Views organization as a social arrangement. Radical structuralist Views organization as a product of structural power relationships.
Epistemological positions Positivism Advocates application of natural science methods to study of social reality. y
Only knowledge confirmed by senses can be warranted as knowledge (phenomenalism)
y
Purpose of theory is to generate hypotheses that can be tested (deductivism)
y
Knowledge is arrived at through the gathering of facts that provide the basis for laws (inductivism)
y
Science must be conducted value free (objective)
Realism Natural and social science should apply the same approach to data collection/explanation (same as positivism). There is an external reality. y
Empirical realism Through the use of appropriate methods, reality can be understood.
y
Critical realism We will only be able to understand and change the social world if we identify the structures at work that generate those events and discourses.
Interpretivism Verstehen, hermeneutic-phenomenological tradition and symbolic interactionism. Difference between people and the objects of the natural sciences must be respected. Scientist must grasp subjective meaning of social action. Ontological
positions
Objectivism Social phenomena confront us as external facts that are beyond our reach or influence (social phenomena are independent of social actors). Constructionism (constructivism) Social phenomena and their meanings are continually being accomplished by social actors.
Paradigms Objectivist There is an external viewpoint which the organization can be viewed from. Subjectivist An organization is a socially constructed product. Can only be understood by the point of view of individuals. Purposes of scientific research within paradigms. Regulatory Purpose of business research is to describe what goes on in organizations. Radical Make judgments how organizations should be and how to achieve this. Functionalist Framework based on problem-solving orientation which leads to rational explanation. Interpretative Questions whether organizations exist beyond social actors, thus it must be explained through their experience. Radical humanist Views organization as a social arrangement. Radical structuralist Views organization as a product of structural power relationships.
Quantitative research Deductive approach to the relationship between theory and research. Testing theories. Incorporated practices and norms of the natural scientific model (positivism in particular). Views social reality as external, objective reality.
Qualitative research Emphasize on inductive approach. Generation of theories. Rejects practices and norms of the natural scientific model. Emphasis on how individuals interpret their social world. Views social reality as constantly shifting emergent property of individuals creation.
Reliability Results of a study are repeatable. Whether the measure is stable or not. Replication Replicate the findings of others. The procedures must have been spelled out in great detail. Validity
Measurement (construct) Does the measure really reflect the concept it is supposed to? Internal Relates to causality. Whether a conclusion that incorporates a causal relationship between two or more variables holds water. Factor that has causal impact = independent variable, effect = dependent variable. External Whether the results of the study can be generalized beyond the specific research context. Ecological Whether or not social findings are applicable to peoples everyday life, natural social setting. The more social scientists intervene in natural settings or create unnatural ones, the more likely the findings will be ecologically invalid. Hawthorne studies.
Trusworthiness Credibility - = internal validity; how believable are the findings? Transferability - = external validity; do the findings apply to other contexts? Dependability - = reliability; are the findings likely to apply at other times? Confirmablility - = objectivity; has investigators values intruded to a high degree?
Naturalism View all objects of study (whether natural or social) to belong to the same realm/Being true to the nature of the phenomenon being investigated/Style of research that seeks to minimize intrusion of artificial methods of data collection.
Research ch des desiigns 2. Resear Classic
experimental design and validity
Control group Eliminate possible rival explanations of a causal finding. Eliminate threats to internal
validity.
y
Testing Subjects may become sensitized to the aims of the experiment (experimenter effect)
y
History Make sure independent variable produces the change
y
Maturation People change which may have implications for independent variable
y
Selection Differences between two groups if they were selected non-randomly
y
Ambiguity about the direction of causal influence Temporal sequence unclear, thus not possible to establish which variable affects the other
Threats to external validity, hence generalizability y
Interaction of selection and treatment To what social/psychological groups can a finding be generalized? Ethnicity, gender, social class etc.
y
Interaction of setting and treatment Can the results of the study be applied to other settings?
y
Interaction of history and treatment Can the findings be generalized to past and future?
y
Interaction effects of pre-testing Subjects may become sensitized to the experimental treatment due to the pre-test.
y
Reactive effects of experimental arrangements Awareness of people participating may influence how they respond
Quasi-experiments Studies with certain characteristics of experimental designs but do not fulfill all the internal validity requirements. Cross
sectional design
Often called social survey design. Surveys, structured observation, content analysis, official statistics and diaries are employed. Nomothetic statements are generated that apply regardless of time and place. y
Collection of data on more than one case Researchers are interested in variation
y
Single point in time in Data on variables of interest are collected simultaneously
y
Collect a body of quantitative or quantifiable data Establish variation between cases
y
Pattern association It is only possible to examine relationships between variables. Problem establishing direction of causal influence
Replicability To the degree that the researcher spells out procedures for selecting respondents, designing measures of concepts, administration of research instruments and the analysis of data. Internal validity Is typically weak since it is difficult to establish causal direction from the resulting data. External validity Is strong when the sample to collect data from has been randomly selected. Ecological validity May be jeopardized by the research instruments (self-completion questionnaires and structured observation schedules) since they disrupt the natural habitat. Longitudinal
design
Used to map change. In terms of reliability, replication and validity it is similar to cross-sectional research. Panel study A sample, often randomly selection national, is focus of data collection on at least two occasions. Cohort study Cohort of people who share a certain characteristic are focus of data collection. Longitudinal designs are better able to deal with problems of ambiguity about the direction of causal influence that plague cross-sectional research. Case
study design
Detailed and intensive analysis of a single case.Single organization/location/person/event. Case study provides a vehicle through which several qualitative methods can be combined, thereby avoiding too great reliance on one single approach. The way they have been designed is heavily influenced by the epistemological tradition of positivism. Idiographic approach Researcher tries to elucidate the unique features of the case. Intrinsic cases Gain insight into particularities of a situation Instrumental cases Focus on using the case as a means of understanding a broader issue or allowing generalizations to be challenged. Particularization rather than generalization is the strength. Case study research is not representative for general cases but good at developing a deep understanding of one case. Types of cases y
Critical case Clearly stated hypothesis and the researcher uses the case to understand under which circumstances the hypothesis will hold and not.
y
Unique case Common focus in clinical studies
y
Revelatory case Opportunity to observe and analyze a phenomenon previously inaccessible to scientific investigation
y
Representative or typical case exemplifies and everyday situation
y
Longitudinal case How a situation changes over time.
Et hical principles y
Harm to participants o
Physical harm, harm to development of self-esteem, stress, career prospects or future employment, inducing to reprehensible acts.
o
Make sure individuals and organizations are not identified or identifiable. > Use pseudonyms
y
Lack of informed consent o
Disguised or covert observation where researchers true identity is unknown. Participants are not given the opportunity to refuse to cooperate.
o
Pe
¡
o
¢
es
ec
Diffic
¤
¢
!
¥
e
£
¨
©
¥
e
¦
¤
¨
y
§
¤
¡
¢
¢
e
e
§
¨
©
¡
¨
e ese c
¤
¦
©
¥
£
e ve
cess
£
¡
¦
se v
ec
¨
©
£
©
¨
¤
es
¨
©
¨
ies
Eve
When
"
y single iece f information is diffic lt to give #
$
%
&
participants are aware of the research er it is unlikely that all are
similarly informed about the nature of the research. y
Invasion o
y
of privacy
Personal information concerning research participants should be kept confidential
Deception is involved
3. Qu
o
When
research is presented as something other than w hat it is
o
Research er
may want to do this to get participants to act more naturally
o
Knowledge
might bias the respondents subse uent behavior '
itative researc 1
0
(
)
Con Elaborate t heory
2
pt building block of theory that represent
3
the points around which business research is conducted.
Devise hypot hesis If
Select research design
a concept is employed in quantitative research
it will have to be measured. Dependent/independent variable
Device measures of concepts
M Select research site(s)
Select research sub jects/respondents
3
4
su
5
3
6
3
nt allows us to delineate fine
diff erences. Gives us consistent device or yardstick for making distinctions Ability to be consistent over time and with other research ers. Provides basis for more precise estimates of
Administ er r esear ch in strument s/coll ect data
the degree of relationship between concepts (correlation).
Pro cess data In
Analyse data
order to give an operational definition to a
concept it is necessary to have an indi
2
4
to . 5
Can be devised through: Develop findings/conclu sion s
question observation schedule (recording of behavior)
Writ e
up findings/conclusion
official statistics e amination of mass media content 7
It
may be pref erable to use multiple indicators and have a number of questions in an interview to tap
a concept. Problem with single indicator y
Incorrectly
classif y many individuals
y
One indicator may capture only a portion of the underlying concept, or be too general
y
Cannot make as fine distinctions
Reliability and validity Reliability
Consistency of measures Stability
Measure of the strength of the relationship between two variables. Perform test-retest method and find high correlation between two observations.
Internal reliability Indicators may not relate to the same thing lack coherence. Can be tested by split-half metodwhere correlation between two halves of indicators are calculated. Inter-observer consistency
Validity Whether a measure of a concept really measures that concept Face
validity Whether the measure reflects the content of the concept in question. Asking other people (experienced and with expertise) Concurrent
validity
Measuring job satisfaction with a criterion (that is relevant and known to differ) employed (absenteeism). To establish concurrent validity we check if people that are more satisfied are less likely to be absent. If lack of correspondence -> the measure might not be addressing job satisfaction.
Predictive validity Future criterion (instead of contemporary as in concurrent validity). Construct
validity Deduction of hypotheses from a theory that is relevant to the concept. Convergent
validity Measure compared to measures of the same concept developed through other methods. If a measure is not reliable, it cannot be valid.
Preoccupat ions of quant itat ive researchers Measurement Carries a number of advantages. Validity and reliability a concern. Causality
Not merely concerned about describing how things are but to say why things are the way
they are. Independent and dependent variables -> causes and effects. Independent variable Manipulated when the experimental design is being employed.
Generalization To say that the findings can be generalized beyond the confinces of the particular context in which the research was conducted.
Probability sampling Researchers seek to generate representative sample to eliminate bias (the sampling is random) Replication
Results should be unaffected by the researcher special characteristics and
expectations. Important to be highly explicit about the procedures.
Crit icism of quant itat ive research y
Fail to distinguish people and social institutions from the world of nature o
Turning blind eye to the differences between the social and natural world. People interpret the world around them which objects of natural science dont.
y
The measurement process possesses an artificial and spurious sense of precision and accuracy o
Measurement by fiat, E.g. respondents are presumed to interpret the key terms of a question similarly
y
Reliance on instruments and procedures hinders the connection between research and everyday life o
Methods rely heavily on administering research instruments to subjects or controlling situations to determine their effects.
y
The analysis of relationships between variables creates a static view of social life that is independent of peoples lives o
Interpretivist epistemological (meaning from individuals point of view) orientation and constructionist ontology (social world is product of individuals, not something beyond them)
4. Sampling Biased
sample One that does not represent the population from which the sample was selected.
Non-probability/non-random sampling Human judgment will affect the selection process Inadequate sampling frame Not comprehensive or inadequate ->Cannot represent the population Non-response Those who choose to participate may in some way differ from those that dont Sampling
error over/underrepresentation of a variable in the sample
Simple random sampling 1. Define population (N) 2. Select/device comprehensive sampling frame 3. Decide sample size (n) 4. Randomize
Systemat ic sample Take every 1 out of 20 employees
Strat if ied random sampling Sample exhibits proportional representation of e.g. the different departments of a firm. Population is stratified by a criterion and a simple random/systematic sample is selected from each strata.
Mult i-stage cluster sampling Primary sampling unit is not the units of the population to be sampled but groupings of those units. These aggregations are known as clusters. y
group 100 largest UK companies by market capitalization
y
sample one company of each of the twelve SIC categories etc.
Sample size Affected by time and cost. It is the absolutesize of the sample that is important, not the relative size. Increased sample size = increased precision (sampling error decreases) The more heterogeneous a population is the larger sample is needed.
Non-probability sampling Samples that dont use probability sampling
Convenience sampling Sample is available to the researcher by virtue of its accessibility. May be impossible to generalize the finding because we dont know what population the sample represents. Probability sampling involves a lot of preparation, so it is often avoided because of the difficulty and cost.
Snowball sampling y
Initial contact with small group of people relevant to the topic and then uses these to establish contact with others.
y
In no sense random. It is very unlikely that the sample will be representative of the population.
y
Better fit with theoretical sampling of qualitative research
Quota sampling y
Used intensively in commercial research
y
Sample is produced that reflects a population in terms of relative proportions of people in different categories (age, gender, ethnicity, religion etc)
Criticism y
Choice is left to interviewer, may be unduly influenced by perceptions -> cannot be representative
y
People may not be typical (underrepresentation of full-time workers etc)
y
Interviewer makes judgments about characteristics which sometimes will be incorrect
y
Difficult to assign people to a certain class
Impossible to calculate standard error etc
y
Arguments in favor y
Cheaper and quicker than probability sampling
y
Dont have to call back to people that did not answer
y
Easier to manage since you dont have to call back
y
Much faster
y
Usefully employed in relation to exploratory work from which new theoretical ideas can be generated Probability samples may be biased too
y
Generalizat ion Even when probability sampling has been employed the findings can only be generalized to the population from which the sample was taken.
Error in survey research y
Sampling error Unlikely one will end up with truly representative sample
y
Non-sampling error Arises from activities related to the sampling process (inaccurate sampling frame and non-response)
y
Data collection error poor question wording, poor interviewing technique, etc.
y
Data processing error faulty management of data, errors in coding of answers
5. Structured interviews Error has an adverse effect on validity of a measure
Intra-interviewer variability Interviewer is not consistent in the way questions are asked/answers recorded
Inter-interviewer variability Interviewers are not consistent with each other about questions/recording of answers Closed question Respondent
is given a limited choice of possible answers. Potential for interviewer
variability is reduced. Facilitate processing of data. Open
question answers need to be sifted and coded so that the data can be analyzed
quantitatively
Coding Introduces another source of error, coding framemay be flawed.
Intra-coder variability coder varies over time Inter-coder variability coders differ from each other Presence of a second interviewer is unlikely to add value.
Telephone y
Cheaper and quicker
y
Geographically dispersed sample is possible
y
Easier to supervise than personal interview
y
Respondents replies will not be affected by interviewer characteristics to the same extent
Limitations y
Lower-income household dont own a telephone, some people may be ex-directory
y
Length of phone interviews cannot be as long as personal interviews
y
Slightly lower response rate than personal
y
Fare less well for the asking of sensitive questions
y
Mobile phones, answerphones etc
y
Telephone interviewers cannot engage in observation
y
Difficult to ascertain if the correct person is replying
y
Cannot employ visual aids
y
Quality of data may be lower
Introduce the research credible rationale Tips: y
Be prepared to call back if respondent is not available
y
Be self-assured, may increase response if you presume they agree
y
Reassure that you are not a sales person
y
Dress in a way that is acceptable
y
Find a time that suits the respondent
Rapport
Establish relationship quickly so that the respondent wants to participate. Cues of
friendliness and smiling, eye contact
Probing Standardized probes Tell a bit more about that, Any reasons you think that?, Mmmmm. Repeat fixed choice alternatives to make it apparent that one must be chosen Dont try to guess a number from something the respondent said (dont put the numbers in his mouth)
Prompt ing Occurs when the interviewer suggest a possible answer to a question to the respondent. All respondents must be given the same prompts! Unacceptable to ask open question and then prompt possible answers to only some.
Crit ical incident met hod Asking respondents to describe critical incidents.Analysis of near-disaster situations.
Project ive met hods, pictoral and photo elicitat ion Inkblots, sentence-completion test, express how they see the identity of the organization by drawing pictures
Verbal protocol Think aloud while they perform a task. Elicit the respondents thought processes while they make a decision or judgment or solving a problem.
Repertory grid technique Identify the interpretative process whereby an individual constructs meaning in relation to his/her social context.
Problems wit h structured interviewing Characteristics
of interviewer Ethnicity (people asked by black interviewer more likely to mention black actors) Response bias
People respond to series of questions in a consistent way but one that is irrelevant to the concept being measured. Acquiescence
Tendency to consistently agree/disagree with a set of questions or items. Social
desirability bias
Answers to questions are related to their perception of the social desirability to those answers. Frame question; imagine what a peer might do instead of what respondent would do.
Problem of meaning When humans communicate they draw on meanings and simultaneously create meanings.
6. Self -complet ion quest ionnaires Questions have to be easy to answer y
Fewer open question
y
Easy-to-follow designs to minimize risk of failure to follow filter questions
y
Shorter to reduce respondent fatigue
Advantages
of self-completion over structured interview
y
Cheaper to administer
y
Quicker to administer
y
Absence of interviewer effect (social desirability, sensitive subjects, interviewer characteristics)
y
No interviewer variability
y
Convenience for respondents
Disadvantagesof
self-completion over structured interview
y
Cannot prompt no one present to help if respondent have difficulties
y
Cannot probe no opportunity to elaborate an answer
y
Cannot ask many questions that are not salient to respondents More likely to become tired
y
Difficult asking other kinds of questions (avoid complex structures such as filter)
y
Questionnaire can be read before answering first question -> answers not independent
y
Do not know who answers
y
Cannot collect additional data
y
Difficult to ask lots of questions
y
Not appropriate for some kinds of respondents
y
Greater risk of missing data
y
Lower response rates
Steps to improve response rates y
Write good covering letter, explaining reasons for research, importance, why recipient was selected, sponsorship, guarantee confidentiality
y
Provide stamped addressed envelope
y
Reminders work, send to non-respondents in two weeks
y
Shorter questionnaires tend to achieve better response rates
y
Clear instructions and attractive layout
y
Do not allow questionnaire to appear unnecessarily bulky
y
Begin with more interesting questions
y
Should comprise as few questions as possible
y
Provide monetary incentives
Do not cramp presentation attractive layout is important! Make sure it is easy on the eye Vertical format more clearly distinguishes questions from answers. Easier to code. Provide clear instructions for how to answer. Tick one box that best represents Never split up a question so that it appears on two pages.
7. Ask ing quest ions Open question respondents can reply however they wish Closed question presented with a set of fixed alternatives
Open quest ions Advantages Answer in their own terms y
y
Allow unusual responses to be derived
y
Do not suggest certain answers
y
Useful for exploring new areas or ones in which researcher has limited knowledge
y
Generating fixed-choice format answers
Disadvantages y
Time consuming for interviewers to administer
y
Answers must be coded, can be unreliable process, variability of coding -> measurement error and hence lack of validity
y
Require greater effort from respondents
y
Variability between interviewers in the recording of answers
y
Tape-recorded transcription immensely time-consuming
Closed quest ions Advantages y
Easy to process answers
y
Enhance comparability of answers
y
Clarify the meaning of a question for respondents
y
Easy for interviewers/respondents to complete
y
Reduce possibility of variability in the recording of answers
Disadvantages Loss of spontaneity in respondents answers y
y
Difficult to make forced-choice answers mutually exclusive
y
Difficult to make forced-choice answers exhaustive
y
Respondents might interpret answers differently
y
Irritating when respondent is not find any category that applies to them
y
Impersonal feel to the interview
Types of quest ions y
y
Personal factual o
Age, gender, employment, income etc.
o
Questions about behavior
Factual about others o
y
y
Informant factual o
Size of firm which they work at etc.
o
Characteristics of an entity of which they have knowledge
About attitudes o
y
Whether they believe something is true or false
About normative standards and values o
y
Likert scale
About beliefs o
y
Personal information about others
Indicate what principles of behavior influence them
Questions about knowledge
Important to bear in mind the different types of questions. They force you to clarify in your own mind what you are asking about. It guards against asking questions in an appropriate format. When building Likert scale it is best not to mix different types of questions.
Rules of t humb y
Questions should always be geared towards asking research questions
y
Decide exactly what it is you want to know
y
Put yourself in the position of the respondent
y
Avoid terms such as regularly, often etc.
y
Avoid long questions o
y
Avoid double-barreled questions o
y
Do you agree with ?
Avoid questions asking two questions o
y
Lacking a frame of reference How satisfied are you with your job?
Avoid leading questions o
y
Question that asks two things
Avoid very general questions o
y
Respondent might feel tempted to omit
When did you last discuss with supervisor/line manager?
Avoid questions that include negatives o
Easy to miss the word not
y
Avoid technical terms
y
Does the respondent have the requisite knowledge?
y
Make sure that there is symmetry between closed questions and its answers o
y
Yes/no
questions should have matching answer (Definitely/Not at all)
Make sure that the answers provided for a closed question are balanced o
Choices balanced, middle answer should be neutral
y
Memory problems
y
Dont know/no opinion o
If not included may force people to express a view they dont hold
o
If included they might tick not to be bothered to think about the issue
Vignette quest ion Presenting respondent with one or more scenarios and then asking them how they would respond when confronted with the situation y
Reduces possibility of unreflective replies
y
Scenarios must be believable
Pilot ing and pre-test i ng quest ions y
To ensure that survey questions operate well.
y
Also ensures that the research instrument as a whole functions well.
y
Open questions can be asked in the pilot to generate fixed-choice answers
y
Piloting provides interviewers with experience of using the schedule and can infuse them with a greater sense of confidence
y
If virtually everyone answers a question in the same way it is not a variable.
y
Find out if respondents feel uncomfortable with certain questions
y
Questions that seems not to be understood
y
Determine the adequacy of instructions to interviewers
y
Should not be carried out on people that might be part of the sample
Using existing questions
y
Have been piloted for you
y
Measurement qualities known
y
Allows for comparison with other research
8. Structured observat ion Systematically observing the behavior of individuals Observation
schedule
y
Clear focus necessary exactly who/what to observ
y
Forms taken by any category of behavior must be mutually exclusive (not overlap) and inclusive
y
Recording system must be easy to operate
y
Problem: Observation schedules sometimes require interpretation by the observer
Observing
behavior
y
Record in terms of incidents
y
Observe and record in terms of short periods of time
y
Observe and record for quite long periods of time. Watch and record continuously.
y
Time sampling
People sample on a random basis Time If certain people are sampled more than one time, dont always observe them the same time of the day If a structured observation study is conducted over a relatively short span of time, issues of the representativeness of findings are likely to arise. y
ad libitum sampling observer records whatever is happening at the time
y
focal sampling specific individual observed for set period of time
y
scan sampling group is scanned at regular intervals and everyones behavior is recorded
y
behavior sampling entire group watched and observer looks for particular behavior
Compared to interviews and quest ionnaires y
Provides more reliable informations about events
y
Greater precision regarding their timing, duration, and frequency
y
Greater accuracy in the time ordering of variables
y
More accurate and economical reconstructions of large-scale social episodes
Reliability y
Inter-observer consistency is a concern
y
Intra-observer consistency Application of the observation schedule may vary over time
y
No easy matter to achieve reliability in structured observation.
y
Observer fatigue and lapses in attention
y
Observers can be trained
Validity y
Is the observation instrument administered as it is supposed to be? o
y
Variability between observers , measure unreliable thus not valid
Do people change behavior because they are observed? o
reactive effect
Crit icism y
Risk of imposing inappropriate or irrelevant framework on the settings being observed. o
y
Solution to precede with unstructured to set variables and categories
Structured observation is rarely able to get at intentions behind behavior. Observer cannot get a grasp of the meaning.
y
Generate lots of bits of data, difficult to see the big picture.
y
Neglects context within which behavior takes place
y
Not all behaviors can be observed
May work best accompanied with other methods Interviews and questionnaires more flexible, offers the opportunity to reveal both attitudes and behaviors.
9. Secondary Analysis and off icial stat ist ics Data that have been collected by other researchers or by other organizations in the course of their business.
Advantages secondary analysis y
Cost and time Tiny fraction of the resources
y
High-quality data o
Rigorous sampling procedures
o
well established procedures for following up non-respondents
o
national samples covering wide variety of regions
o
generated by experienced researchers
o
control procedures to check quality of the data
y
Opportunity for longitudinal analysis
y
Subgroup or subset analysis specialized categories of individuals
y
Opportunity for cross-cultural analysis
y
More time for data analysis
y
Reanalysis may offer new interpretations
y
o
Certain variables impact on other variables
o
New theoretical ideas may suggest new analyses
o
Alternative method of analysis
Wider obligations of business researchers
Limitat ions secondary analysis y
Lack of familiarity with data
y
Complexity of data
y
No control over data
y
Absence of key variables
Off icial stat ist ics y
Data have already been collected, saves time and expense. Can give more complete picture if not based merely on sample.
y
Problem with reactivity less pronounced
y
Cross-sectional and longitudinal and sometimes cross-cultural
Process by which official statistics are generated involves element of interpretation. Problem with convergent validity. the dark figure.
Reliability and validity Reliability seems jeopardized because definitions and policies regarding the phenomena to be counted vary over time.
Ecological fallacy Problems of validity. Variations may be due to other factors than those measured.
10. Qualitat ive research Tends to be concerned with words rather than numbers. y
Inductive view of the relationship between theory and research; former generated out of the latter.
y
Epistemological position described as interpretivist. Stress is on understanding the social world by examining the interpretations by the participants of that world
y
Ontological position described as constructionist. Social properties are outcomes of the interactions between individuals.
y
Distinctiveness not solely absence of numbers.
Main research met hods associated wit h qualitat ive research y
Ethnography/participant observartion
y
Qualitative interviewing
y
Focus groups
y
Language-based approaches to the collection of qualitative data
y
o
Discourse analysis
o
Conversation analysis
Collection and qualitative analysis of texts and documents
Main steps in qualitative researc
8
1. General research question
2. Select relevant site(s) and sub jects
5b. Collection of further data
3. Collection of relevant data
4. Interpretation of data
5. Conceptual and theoretical work
5a. Tighter specification of the research question(s)
6. Writing up findings/conclusions Distinction between definitive and s nsiti in concepts. By quantitative research a concepts @
A
9
becomes fixed through elaboration of indicators. = strait jacket on social world. Gr
B
unded t eor C
D
Allowing theoretical ideas to emerge out of ones data (Check Pettigrew)
Reliability and validity in qualitative researc E
y
Ext
y
Int
F
G
P
n l H
G
n l
H
F
li bilit whether or not, when there is more than one observer, research team S
R
Q
li bilit degree to which a study can be replicated I
Q
R
P
agree about what they see and hear (inter-observer consistency) y
Int
T
n l
lidit wheth er or not there is a good match between research ers observations
W
X
V
U
V
and the theoretical ideas they develop. Congruence between concepts and observations. y
Ext
Y
`
n l a
b
lidit degree to which findings can be generalized across social settings. c
a
Alternative criteria for evaluating qualitative researc
d
Should be valued and evaluated based on diff erent criteria Trustwort iness d
Credibilit
e
Ensuring that
research is carried out according to the canons of good practice and submitting
research findings to the members of the social world who were studied for confirmations that the investigator has correctly understood the social world.Respondent/
f
ember validation.
Transf era bilit e
The research tends to be oriented to th e contextual uniqueness and significance of the particular case that was studied. Researcher are encouraged to producethi k des riptions - rich accounts of g
the details of a culture.
g
Dependability
Should adopt an auditing approach. Ensuring that complete records are kept of all phases of the research process. Confirmability
Complete objectivity is impossible but the researcher should have acted in good faith. Should be apparent that he/she overly didnt allow personal values. Authenticity y
Fairness
y
Ontological
y
Educative authenticity Better appreciate the perspectives of other members of their social
represent different viewpoints among members of social setting? authenticity Help members to arrive at better understanding of their social world?
setting? y
Catalytic
y
Tactical authenticity Has research empowered members to take steps for action?
authenticity Has research acted as impetus to members to change their situation?
Points of affinity with action research.
Preoccupat ions of qualitat ive researchers At the level of epistemology qualitative researcher are more influenced by interpretivism. Seeing
through the eyes of the people being studied People can attribute meaning to events and to the environment. Goal of seeking to probe beneath surface appearances. Description and emphasis on context
Provide a great deal of descriptive detail when reporting the fruits of their research. Often full of seemingly trivial details, but they are important. It is recommended to provide thick descriptions of social settings, events, and often individuals.
Emphasis on process Tends to view social life in terms of processes. Show how events and patterns unfold over time. Flexibility
and limited structure
Many qualitative researchers are disdainful of approaches that entail imposition of predetermined formats on the social world. Qualitative research tries not to delimit areas of enquiry too much and asks fairly general rather than specific questions. Concepts
and theory grounded in data
Usually inductively arrived at from the data that are collected.
Crit ique of qualitat ive research Too subjective Impressionistic and subjective. Findings rely too much on researchers unsystematic views about what is significant and important.
Difficult
to replicate
Unstr ct red and reliant pon the researchers ingen ity.Very diffic lt to replicate.Responses of i
i
i
i
i
participants likely to be affected by characteristics of the researcher. Interpretation will be infl enced i
by s bjective learnings of the researcher. i
Problems of generalization How can j st one or two cases be representative of all cases. Findings of q alitative research are to i
i
generalize to theory rather that to populations p
Lac q
of transparency
It is often not obvio s how the analysis was cond cted. i
i
Contrasts Quant itat ive vs Qualitat ive Numbers
Words
Point of view of researcher
Point of view of participants
Researcher distant
Researcher close
Theory/concepts tested in research
Theory/concepts emergent from data
Static
Process
Structured
Unstructured
Generalization
Contextual understanding
Hard, reliable data
Rich, deep data
Macro
Behavior
Artificial setting
Micro Meaning Natural setting.
Similarit ies - Quant itat ive and Qualitat ive y
Both are concerned with data reduction
y
Both are concerned with answering research questions
y
Both are concerned with relating data analysis to the research literature
y
Both are concerned with variation
y
Both treat frequency as a springboard for analysis
y
Both seek to ensure that deliberate distortion does not occur
y
Both argue for the importance of transparency
y
Both must address question of error
y
Research methods should be appropriate to the research question
Act ion research y
Experiments are on real problems within an organization and are designed to assist in their solution.
y
Involves iterative process of problem identification, planning, action and evaluation.
y
Leads to re-education, changing patterns of thinking, and action
y
1. Problem identification
Intended to contribute both to academic theory
4. Eval ation
2. Planning
h
and practical action
3. Action
y
It should have implications that relate to situations other than the one that is studied
y
Should be usable in everyday life, should also be concerned with theory
y
Should lead to the generation of emergent or grounded theory
y
Action researchers should recognize that the research will have practical implications and they must be clear about the take aways from the project.
Can involve the collection of both quantitative and qualitative data. The investigator becomes part of the field of study. The report is seen primarily as a discussion document, which presents a number of action strategies from which collaborators will jointly select a course to take.
Cognit ive mapping Complementary to action research. Maps can be used as a problem-solving device by researchers. Commonly used management consulting technique.
Primary function: Reflective thinking about problems to enable steps to be taken towards its solution Enable managers to understand the way others in the group perceive a problem.
Et hnography and part icipant observat ion Organizational ethnography (writing about people/culture) concerned with social relation that are related to certain goal-oriented activities. Ethnographic approach implies intense researcher involvement in the day-to-day running of an organization, so that the researcher can understand it from an insiders point of view.
Access Most difficult step is to gain access to the social setting. Strategic planning, hard work and dumb luck. y
Friends contacts colleagues, academics
y
Champion within organization
y
Through top management
y
Offer something in return
y
Provide clear explanation of your aims and methods and be prepared to deal with concerns.
y
Be prepared to negotiate
y
Be honest about the amount of peoples time you are likely to take up
y
Hanging around, loitering
Covert
not disclose the fact that you are a researcher. Ethical considerations.
Ongoing access People will have suspicions of you y
y
People worry what thay say may get back to bosses or colleagues
y
May sabotage, engage in deceptions, misinformation and not allowing access to back regions
Smooth
the path
Play up credentials Be non-judgmental when things are said about informal activities
Roles for et hnographers omplete participant r
Fully functioning member of the social setting. His true identity is not known.
Participantas-observer
Same as complete participant but members aware of the researcher
s
bserver-asparticipant
Researcher is mainly an interviewer. Some observation but little involves participation
r
omplete observer
The researcher does not interact with people. Unobtrusive forms of observation.
Most sampling in qualitative research entails purposive sampling of some kind.
Theoretical sampling Better for qualitative research than probability. Aims to discover categories and their properties and to suggest the interrelationships into a theory. Grounded theory advocates that you sample in terms of what is relevant and meaningful for your theory.
Interviewing in qualitat ive research Different from interviewing in quantitative research! y
Less structured than quantitative (structured interview) which is supposed to maximize the reliability and validity of measurement of key concepts. In qualitative there is an emphasis on greater generality in the formulation of initial research ideas and on interviewees own perspectives.
y
Much greater interest in interviewees point of view.
y
Rambling or going off tangents is often encouraged.
y
Interviewers can depart significantly from any schedule or guide that is being used.
y
The researcher wants rich, detailed answers. Interviewee may be interviewed more than one, sometimes several occasions.
Unstructured interview Interviewee is allowed to respond freely, interviewer follows up on points that seem worthy. Preferred if the researcher is to gain a genuine understanding of the world views of the members. Semi-structured
interview List of questions on fairly specific topics to be covered, interview guide
Preferred if the interviewer is in the beginning of the investigation with a fairly clear focus. Formulation of research question should not be so specific that alternative avenues of enquiry that might arise during the collection of fieldwork data are closed off.
Prompting Get the interviewee to think more about the topic and provide a more detailed response Qualitative researchers are often interested also in
how the
interviewee says things. Therefore
recording might be good but it may also disconcert respondent who become alarmed at the prospect of their words being preserved. When tape recorder turned off Unsolicited accounts source of revealing information or views.
Advantages of part icipant observat i on vsqualitat ive interviewing Seeing
through other s eyes t
Prolonged immersion in a social setting makes the researcher better equipped to see as others see. Learning
the native language
In order to understand a culture, the language must be lerned.
The ta en for granted Matters interviewees take for granted are less likely to surface u
Deviant
and hidden activities
Much of sabotage, resistance at work and criminal activity is known from participant observation Sensitivity
to context Allows the context of peoples behavior to be mapped fully Encountering the unexpected and flexibility Unstructured nature makes it more likely to uncover unexpected topics or issues. Naturalistic emphasis Has potential to come closer. Interview is a disruption of normal flow of events.
Advantages of qualitat ive interviewingvspart icipant observat ion Issues resistant to observation Some issues are not amenable to observation Reconstruction of events
Ethical considerations Hidden hardware etc Reactive
effects
Presence of participant observer will result in reactive effects. Observers disturb the situation. Less
intrusive in people s lives v
Participant observation takes up lots of peoples time Longitudinal
research easier Limits to the amount of time a researcher can devote to observation.
Greater breadth of coverage Researcher is not limited to the restricted range of people, incidents and localities Specific
focus
Focus groups y
Emphasize specific theme or topic to be explored indepth
y
Can save time and money instead of carrying out individual interviews
y
Practitioner interested in the way which individuals discuss a certain issue as members of a group
y
Researcher aims to provide a fairly unstructured setting
The dynamics of focus groups could make individuals define business problems in new and innovative ways. y
Understand why people feel the way they do
y
Participants are able to bring to the fore issues they deem to be important
y
Individuals argue with each other and challenge each others views
y
The way individuals collectively make sense of a phenomenon can be studied. Can be regarded more naturalistic
Criticism of reliability and lack of realism. Moderator can vary Variation in the interpretation of transcripts Better representativeness can be achieved by stratifying the population and greater reliability can be achieved with different moderators. y
Focus group will work best if recorded. o
The researcher interested in who expresses views.
o
Makes it possible to study the processes by which meaning is collectively constructed.
o
Interested in both what people say and how they say it.
Theoretical saturation, perhaps twelve to fifteen. More groups will increase complexity.
Size Six to ten Groups more than eight difficult to manage Consciously over-recruit since people might not show up. Smaller groups demand more since every participant must contribute more
Moderator Not intrusive or structured May be necessary to refocus if it goes off at a total tangent. Allow discussion to flow freely/intervene to bring out salient issues
Part icipants Snowball sampling Only natural groups not always possible Natural groups may also not bring up taken-for-granted assumptions
Limitat ions of focus groups y
Researcher has less control over proceedings than with individual interviews
y
Data hard to analyze. Huge amounts quickly produced
y
Difficult to organize. Persuade people to turn up at particular time
y
Recordings time-consuming to transcribe
y
Possible problems of group effects
y
Not appropriate when they might cause discomfort among participants
Language analysis Conversat ion analysis Fine-grained analysis of talk as it occurs in interaction in naturally occurring situations. Indexicality The meaning of an act, spoken words or utterances, pauses, sounds, depends upon context they are used Reflexivity Spoken words are constitutive of the social world in which they are located. Recordings of naturally occurring conversations and their transcription for the purpose of intensive analysis of the sequences of interaction revealed in the subsequent transcripts.
Three assumpt ions of C A y
Talk is structured
y
Talk is forged contextually
y
Analysis is grounded in data
CA fails to capture body movements Adjacency pairs
Talk involves linked phases. Question followed by an answer; invitation response;
greeting returned greeting.
Discourse analysis Forms of communication other than talk. Company mission statements; websites; email messages Anti-realist
denies that there is an external reality
Constructionist
Recognition that in the process particular depiction of reality is built up
Discourse shapes the rules that determine how we speak and act in relation to a given topic. y
Reading the detail
y
Looking for rhetorical detail: ways in which arguments are constructed
y
Looking for accountability
y
Cross-referencing discourse studies
Narrat ive analysis Elicitation and analysis of language that is sensitive to the sense of temporal sequence that people, as tellers of stories about their lives or events around them, detect in their lives and surrounding episodes and inject into their accounts.
Rhetorical analysis Analysis of argumentation, literary devices (metaphor, synecdoche, metonymy and irony) Contrast Subject is described in terms of opposite to reinforce a point List together we are strong, together we can make progress, together we Headline-punchline/puzzle-solution Position-taking Pursuit, repetition, alliteration
Documents as sources of data Criteria
for assessing the quality
Authenticity
Genuine and of unquestionable origin?
Credibility
Free from error and distortion?
Representativeness Meaning Clear
Typical of its kind?
and comprehensible?
Organizational
documents Likely to authentic and meaningful, but may not be accurate representation of how different actors perceive the situations. Documents linked to other documents Disagreements may be suppressed Documents are texts written with distinctive purposes in mind
Semiot ics Sign, signifier etc.Focuses on the way that messages are communicated as systems of cultural meaning. Uncover the hidden meanings that reside in texts as broadly defined.
Hermeneut ics Understanding/interpretation of texts. Bring out the meaning from the perspective of its author.
Historical analysis Documents and other artifacts to trace history of an organization. y
Understanding of contemporary organizations relies on having an awareness of how they developed historically.
y
Historical analysis can reduce the ideological biases that are embedded in current fasionable trends in organization theory and practice.
y
Historical analysis enables interpretation of existing organization arrangements as the result of intentional or implicit decisions made in the past
y
Theories can be subjected to more radical test.
Break ing down t he quant itat ive/qualitat ive divide Researchers perceive that there is a difference between qualitative and quantitative research in terms of strategy.
Natural science model and qualitat ive research Difficulties y
No agreement on the epistemological basis of natural sciences
y
Researchers dont do and write the same things
Similarities y
Empiricist overtones theory grounded in datat constitute a manifesto for empiricism
y
Specific problem focus Qualitative research can be employed to investigate specific, tightly defined research questions.
y
Hypethesis- and theory-testing Analytic induction or grounded theory
y
Realism Epistemological basis of the natural sciences.
Quant itat ive research and interpret ivism Quantitative researchers frequently address meanings. Quant itat ive research and construct ionism Construction of social world. Conventional quantitative content analysis can be useful for this.
Epistemological and ontological considerat ions Carry with them a cluster of commitments. However there is no perfect correspondence between research strategy and matters of epistemology and ontology.
Problems wit h quant itat ive/qualitat ive contrast Behavior
vs meaning
Quantitative research frequently involves the study of meanings in the form of attitude scales. Qualitative research frequently entails examination of behavior in context
Theory tested in research vs emergent from data Survey-based studies are often more exploratory than simply theory-testing Numbers vs words Qualitative researchers sometimes undertake limited amount of quantification of their data. many, often, some
Artificial
vs natural
Naturalism of focus groups is assumed rather than demonstrated. Participant observation, the researcher can be source of interference that makes situation less natural.
Quant if icat ion in qualitat ive research Thematic analysis Implicit quantification when identifying themes and the elevation of some themes over others. Q uasi-quantification frequently, rarely. Combating
anecdotalism Give numbers to do this
Mixed met hods research Arguments
against mixing
Epistemological commitments integration of research strategies ignores assumptions underlying research methods and transforms qualitative inquiry to procedural variation of quantitative inquiry. Separate paradigms Values, methods, assumptions are incompatible between paradigms. Integration will be at superficial level within single paradigm.
Triangulat ion Cross-check results of one method by employing another method associated with the other research strategy. When triangulation is undertaken, there is a possibility of a failure to corroborate the findings.
Qualitat ive research facilitates quant itat ive research Providing hypotheses helpful as a source of hypotheses or hunches that can subsequently be tested Aiding measurement
In-depth knowledge acquired through qualitative research can inform the
design of survey questions.
Quant itat iveresearch facilitates qualitat iveresearch Prepare selection of people to interviewed, or companies to be selected as case studies. Filling
in the gaps
Static
and processual features Quantitative research tends to bring out a more static picture of
social life, qualitative is more processual Q ualitative
facilitates interpretation of relationship between variables Influenced by the
independent variable but which has effect on the dependent variable. Qualitative research can provide explanation. Study different aspects of a phenomenon
1
a puzzle Outcomes of research are not always easy to anticipate 2. Mixed methods research, like mono-method, must be completely designed and conducted. w
Solving
Macro vs macro
