Lecture 1: Introduction
Definition: “a series of well-thought-out and carefully executed activities that enable the
manager to know how organizational problems can be solved, or at least considerably
minimized”
Hallmarks of scientific research:
● Purposiveness: knowing “the why” of your research
● Rigor: ensuring a sound theoretical base and methodological design
● Testability: being able to test logically developed ideas base on data
● Replicability: finding the same results if the research is repeat in a similar circumstance
● Precision & confidence: drawing accurate conclusions with a high degree of confidence
● Objectivity: drawing conclusions based on facts (rather than on subjective ideas)
● Generalizability: being able to apply research findings in a wide variety of different
settings
● Parsimony: shaving away unnecessary details, explaining a lot with a little
Research design:
Conclusive research is characterized by clearly
defined phenomena that can be measured by means
of quantitative data
Descriptive: testing the correlational relationship
between two or more variable
Causal: testing the causal relationship between two
or more variables by means of an experiment
Correlation VS Causality
Just because 2 events correlate, does not mean they cause one another to correlate, it can be
due to external events as well
Conditions for causality:
1) X (or IV [independent variable]) and Y (or DV) co-occur (correlate)
2) A logical explanation for the effect of X on Y is needed
3) X (or IV) proceeds Y (or DV) in time [one will exist without the other]
4) No other cause (Z) explains the co-occurence of X (or IV) and Y (or DV)
Lecture 2: Measurement, scales and survey design
1. Sources of error:
,Total error: variation between true mean value in population of the variable of
interest vs observed value
Random sampling error: error due to sample being imperfect representation of
the population
Non-sampling errors: error that can be attributed to sources other than
sampling and can be random or non-random
Non-response error: error when sample participants do not respond
Response error: arises from responses that who give inaccurate answers or
whose answers are mis-recorded / mis-analyzed
● Research error: surrogate information - measurement - population
definition - sampling frame - data analysis
● Interviewer error: respondent selection - questioning - recording -
cheating
● Responded error: inability - unwillingness
2. Questionnaire design
Definition: “a structured technique for data collection consisting of a series of
questions, written (maybe verbal), that a participants answers” ~ it is important
it is structured since you want to compare across all people
Challenges:
1. Translate the information needed into a set of specific questions that
participants can and will answer
2. A questionnaire must uplift, motivate and encourage the participant to
become involeged, to cooperate and complete the task
3. A questionnaire should minimize response error
4. Standardized measurement == generalizability (good)
Design Process
1. Specify information needed: what is it you want to answer and therefore need
to measure? Research problem -> research question -> theoretical framework ->
DV IV (incl moderators, mediators, control) ->
,2. Specify the type of interview method
- Mail & online -> these are self administered questionnaire wso the
questions must be simple and detailed instructions must be provided
(non-response error)
- Online & computer assisted -> complex patterns must be skipped and
randomization of questions can easily be implemented to avoid order bias
(non-response error)
- Face-to-face interview -> because of the interaction, a lengthy and
complex questionnaire with varied questions can be used (bias response
error)
3. Determine the content of individual questions: 1) is the question necessary? 2)
are several questions needed instead of one (avoid double-barreled)
4. Overcoming the participants inability and unwillingness to respond: there will
be no (or inaccurate) answers if the respondent is not informed about the topic ->
may not remember/recall what you’re asking or is simply unable / unwilling to
answer
Guidelines when composing questions:
1) Is the participant informed? -> add filter question
2) Is the participant able to articulate? -> add pictures/map/description
3) Can the participant remember?
a) Omission -> inability to recall
b) Telescoping -> remembering an event occurring more often than it
actually occurred
c) Creation -> the remembered event did not occur at all
4) Additional learnings:
a) The less effort for the participant the better
b) Providing context increases willingness
c) Provide a legitimate purpose for the research
d) Be careful with asking sensitive information
5. Choosing question structure
Unstructured -> open question
- Enable to express general attitudes and opinions
- Enable to indicate other relevant issues
, - High risk of bias
- Coding costly and time consuming
Structured -> specified set of response alternatives and response format
6. Choose question wording
1) Define the issue
2) Use ordinary words
3) Use unambiguous words (regularly, often, occasionally - less than 1 or 2 ..)
4) Avoid leading or biasing questions (presupposition)
5) Avoid implicit alternatives (provide alternatives in the question)
6) Avoid implicit assumption (when making assumption, make them clear)
7) Avoid generalization and estimates (don't ask them to estimate/generalize)
8) Use positive and negative statements (people are more inclined to agree
to statements -> therefore mix)
7. Arrange questions in proper order: opening question (simple/interesting) ->
basic information (research topic) -> classification information
(socio-economic & demographic) -> identification information (postal code,
membership number) -> difficult question (late in sequence)
Funnel approach: general questions should precede specific questions in order
to minimize effect on subsequent question (bias)
Logical order: all questions that deal with a particular topic should be asked
before beginning a new topic
8. Identify the form and layout
9. Reproduce the questionnaire
10. Eliminate problems by pilot-testing
3. Measurement levels
Measurement: assignment of numbers to characteristics of objects according to
pre-specified rules | scaling: generating of a continuum upon which measured
objects are located
Definition: “a series of well-thought-out and carefully executed activities that enable the
manager to know how organizational problems can be solved, or at least considerably
minimized”
Hallmarks of scientific research:
● Purposiveness: knowing “the why” of your research
● Rigor: ensuring a sound theoretical base and methodological design
● Testability: being able to test logically developed ideas base on data
● Replicability: finding the same results if the research is repeat in a similar circumstance
● Precision & confidence: drawing accurate conclusions with a high degree of confidence
● Objectivity: drawing conclusions based on facts (rather than on subjective ideas)
● Generalizability: being able to apply research findings in a wide variety of different
settings
● Parsimony: shaving away unnecessary details, explaining a lot with a little
Research design:
Conclusive research is characterized by clearly
defined phenomena that can be measured by means
of quantitative data
Descriptive: testing the correlational relationship
between two or more variable
Causal: testing the causal relationship between two
or more variables by means of an experiment
Correlation VS Causality
Just because 2 events correlate, does not mean they cause one another to correlate, it can be
due to external events as well
Conditions for causality:
1) X (or IV [independent variable]) and Y (or DV) co-occur (correlate)
2) A logical explanation for the effect of X on Y is needed
3) X (or IV) proceeds Y (or DV) in time [one will exist without the other]
4) No other cause (Z) explains the co-occurence of X (or IV) and Y (or DV)
Lecture 2: Measurement, scales and survey design
1. Sources of error:
,Total error: variation between true mean value in population of the variable of
interest vs observed value
Random sampling error: error due to sample being imperfect representation of
the population
Non-sampling errors: error that can be attributed to sources other than
sampling and can be random or non-random
Non-response error: error when sample participants do not respond
Response error: arises from responses that who give inaccurate answers or
whose answers are mis-recorded / mis-analyzed
● Research error: surrogate information - measurement - population
definition - sampling frame - data analysis
● Interviewer error: respondent selection - questioning - recording -
cheating
● Responded error: inability - unwillingness
2. Questionnaire design
Definition: “a structured technique for data collection consisting of a series of
questions, written (maybe verbal), that a participants answers” ~ it is important
it is structured since you want to compare across all people
Challenges:
1. Translate the information needed into a set of specific questions that
participants can and will answer
2. A questionnaire must uplift, motivate and encourage the participant to
become involeged, to cooperate and complete the task
3. A questionnaire should minimize response error
4. Standardized measurement == generalizability (good)
Design Process
1. Specify information needed: what is it you want to answer and therefore need
to measure? Research problem -> research question -> theoretical framework ->
DV IV (incl moderators, mediators, control) ->
,2. Specify the type of interview method
- Mail & online -> these are self administered questionnaire wso the
questions must be simple and detailed instructions must be provided
(non-response error)
- Online & computer assisted -> complex patterns must be skipped and
randomization of questions can easily be implemented to avoid order bias
(non-response error)
- Face-to-face interview -> because of the interaction, a lengthy and
complex questionnaire with varied questions can be used (bias response
error)
3. Determine the content of individual questions: 1) is the question necessary? 2)
are several questions needed instead of one (avoid double-barreled)
4. Overcoming the participants inability and unwillingness to respond: there will
be no (or inaccurate) answers if the respondent is not informed about the topic ->
may not remember/recall what you’re asking or is simply unable / unwilling to
answer
Guidelines when composing questions:
1) Is the participant informed? -> add filter question
2) Is the participant able to articulate? -> add pictures/map/description
3) Can the participant remember?
a) Omission -> inability to recall
b) Telescoping -> remembering an event occurring more often than it
actually occurred
c) Creation -> the remembered event did not occur at all
4) Additional learnings:
a) The less effort for the participant the better
b) Providing context increases willingness
c) Provide a legitimate purpose for the research
d) Be careful with asking sensitive information
5. Choosing question structure
Unstructured -> open question
- Enable to express general attitudes and opinions
- Enable to indicate other relevant issues
, - High risk of bias
- Coding costly and time consuming
Structured -> specified set of response alternatives and response format
6. Choose question wording
1) Define the issue
2) Use ordinary words
3) Use unambiguous words (regularly, often, occasionally - less than 1 or 2 ..)
4) Avoid leading or biasing questions (presupposition)
5) Avoid implicit alternatives (provide alternatives in the question)
6) Avoid implicit assumption (when making assumption, make them clear)
7) Avoid generalization and estimates (don't ask them to estimate/generalize)
8) Use positive and negative statements (people are more inclined to agree
to statements -> therefore mix)
7. Arrange questions in proper order: opening question (simple/interesting) ->
basic information (research topic) -> classification information
(socio-economic & demographic) -> identification information (postal code,
membership number) -> difficult question (late in sequence)
Funnel approach: general questions should precede specific questions in order
to minimize effect on subsequent question (bias)
Logical order: all questions that deal with a particular topic should be asked
before beginning a new topic
8. Identify the form and layout
9. Reproduce the questionnaire
10. Eliminate problems by pilot-testing
3. Measurement levels
Measurement: assignment of numbers to characteristics of objects according to
pre-specified rules | scaling: generating of a continuum upon which measured
objects are located
