管理学专业essay/report/assignments/paper/research/summary代写
Research Methods
MGT
Quantitative Research Methods
Leanne Ingram
On your devices:
Go to kahoot.it for our
quiz session.
Todays lecture
- Recap of quantitative DESIGNS from last semester.
Including some feedback from Assignment 1.
- Quantitative METHODS most often used in management research and the advantages/disadvantages of these.
- Self-completed Questionnaires
- Structured Interviews
- Structured Observations
- Understanding Reliability and Validity in quantitative research.
1. Recap of quantitative
DESIGNS from last
semester
Including some feedback on
Assignment 1
Quantitative
Research Designs…
- Aim to describe, examine, compare, analyse
######## quantities of things, by collecting numerical
######## data
- Examine the relationship between variables
- USUALLY
- Fit within positivist philosophies
- Are used in deductive research
- Are conducted through experimental research and survey research strategies
Feedback from Assignment 1:
If your research is not aligned
in this way (positivist,
deductive, quantitative) then
your reasons for taking a
non-standard approach need
to be very clear.
Deductive research cycle
Hypothesis
formulation
Data
analysis
Inter-
pretation
Data
collection
Research
question
Research
Objective
Hypotheses
A clear, specific statement of something that is believed to be true but that needs to be tested.
- E.g. Student gradesare positively related to number of job offers received after graduation.
Will identify specific variables, and the nature of the relationship.
Feedback from Assignment 1 :
If you have several
hypotheses, it needs to be
possible for them all to be
true at the same time you
cant predict every possible
outcome.
Feedback from Assignment 1 :
Your hypothesis should state the nature of
the relationship, not just that there will be
one. Are they positively related they
increase/decrease at the same time.
Or negatively related as one increases,
the other decreases.
2. Research
designs/strategies
Experiments
Field experiments
Non experimental designs e.g. Surveys
/ Direct Measurements / Observations
Experimental Design
######## Participants are assigned randomly to
######## intervention or control groups
No Training
Programme
Intervention group
Dependent
Variable:
Interpersonal
Skill
Training
The Effect Of Training On Interpersonal Skill
Independent Variable:
Pre test Control group Post test
Training
Feedback from Assignment 1 :
It is not experimental/quasi-
experimental research if
there is no control group.
Sampling
methods
Stratified
sampling
Simple
random
sampling
Con-
venience
sampling
Cluster
sampling
Quota
sampling
Snowball
sampling
To prevent selection bias, choose the right sampling method!
1. Simple random sampling
- All potential participants have an equal probability of being selected
- Usually gives the most accurate statistics
Example: You let a computer program randomly select 500 students from all students at the UoS.
What is needed?
- A complete, accurate and easily accessible sampling frame that lists the entire population
- Suitable when your research question affects the entire population.
NOTE: feasibilityplays a role is it feasible to interview students that were randomly selected for face-to-face interviews from a sample frame that contained all students of all universities in England?
Stratifiesamplind
samplingrandom Simple g
Convenisamplinence
g samplingCluster
Snowball sampling samplinQuota g
Feedback from Assignment 1 : For simple
random sampling access is needed to
every person in the sampling frame.
2. Stratified sampling
- Sampling frame is divided into mutually exclusive strata i.e. different groups (e.g., public or private sector companies; different industry sectors; different organisational layers…)
- A random sample is taken from each group
Particularly useful when interested in groups that are not equally represented.
Example : There are more UG than PG students at
UoS. If you want a sample of the student
population to reflect this distribution, stratified
sampling might be needed.
Stratifiesamplind
samplingrandom Simple g
Convenisamplinence
g samplingCluster
Snowball sampling samplinQuota g
Feedback from Assignment 1 :
In stratified sampling, some people from
every strata are sampled at random.
3. Cluster sampling
- A population can be divided into a large number of naturally occurring clusters which are, to some extent, random and therefore representative.
- A sample of clusters is selected at random, then all individuals within those clusters are included
Example : Population: All students at the UoS Clusters: Degree Courses. Cluster sampling: a small number of degree courses is randomly selected, all students from the selected degree courses are interviewed.
Stratifiesamplind
samplingrandom Simple g
Convenisamplinence
g samplingCluster
Snowball sampling samplinQuota g
Feedback from Assignment 1 :
In cluster sampling, some clusters are
selected at random and everyone in those
clusters is sampled.
5. Snowball sampling
- Not probability based at all
- Identify an initial person or group of people, and ask them to suggest other participants (etc.)
- Useful when a very specific (and relatively rare) type of participant is needed
Example: You want to survey alumni of the MSc. HRM
from the year 1994 but you dont have their addresses.
You ask the student ambassador of that cohort to
forward the survey to their peers of that cohort, asking
them to forward it to others of that cohort.
Stratifiesamplind
samplingrandom Simple g
Convenisamplinence
g samplingCluster
Snowball sampling samplinQuota g
Feedback from Assignment 1 :
Snowball sampling should only be used to
find very specific and rare participants
(not because it is easy to ask your friends!)
####### 2. Quantitative METHODS most often used
####### in management research and the
####### advantages/disadvantages of these.
- Self-completed Questionnaires
- Structured Interviews
- Structured Observations
Questionnaires, Surveys,
Structured Interviews:
Are they all the same?
Saunders et al. 2016, p. 440. See page 441 for a detailed matrix of suitability of methods.
These are
also known
as
structured
interviews.
They are all kinds of questionnaire or survey.
…we use questionnaire as a general term to include all methods of data collection in which each person is asked to respond to the same set of questions in a predetermined order. (Saunders et al. 2016, p. 437)
####### Structured Interviews
- These involve a researcher recording the answers given by the participant to a predetermined set of questions.
- The possible answers to the questions have already been defined, and the relevant option should be selected.
- Unlike semi-structure or unstructured interviews, the interviews MAY NOT change or adapt the questions all questions are closed.
- The benefit of structured interviews compared to self- completed questionnaires is that the interviewer can confirm that the correct person has taken part.
####### Self-Completed Questionnaires
- The participant records their own answers to a predetermined set of questions.
- The possible answers to the questions have already been defined, and the relevant option(s) should be selected.
- The researcher needs to be confident beforehand that the participant will interpret the questions as intended, as they will not be there to explain.
- Web and mobile based questionnaires can reach a large number of participants, but response rates can be low.
- Compared to structured interviews, the researcher cannot always guarantee that the intended person has taken part.
####### All Survey Methods things to consider.
- The questions that participants will be asked need to be clearly planned and organised in detail before data collection begins.
- Not always possible to follow-up on peoples answers, so the questions need to be clear and relevant first time.
- Think about they type of research question you are trying to answer: – Quantitative surveys are most appropriate for descriptive or explanatory research. – Qualitative methods are usually more appropriate for exploratory or evaluative research.
- Use research literature in the same area to find out what types of data have been used before and if the methods were effective.
####### Types of Questions
- Demographic or factual – Information the participant should know and is likely to be accurate. Usually needed to control for, or explore differences between groups of people. – For example: age, gender, marital status, number of hours worked.
- Attitudes and opinions Participants may need to think about these before answering. Based on the context and the persons own beliefs. – For example: How do you feel about the following statement: My manager supports all team members. (Strongly agree, mildly agree, neutral, mildly disagree, strongly disagree)
- Behaviours and events Participants need to remember actual events and behaviour of themselves or others in a situation from the past, present, or future. – For example: How often do you go to your manager for support?
####### Questionnaire Examples
####### Demographic/Factual Questions
How old are you? ___________
What is your highest level of education?
High School
College or Sixth Form
Undergraduate Degree
Postgraduate Taught Degree
Doctoral Degree
Are you a full-time or part-time employee?
Full-time
Part-time
####### Questionnaire Examples
####### Attitude/Opinion Questions
Do you think the pay and other benefits you receive are fair?
Yes
No
To what extend do you agree with this statement: My employer cares about my well-being.
Strongly Agree
Mildly Agree
Neither Agree nor Disagree
Mildly Disagree
Strongly Disagree
####### Questionnaire Examples
####### Behaviour/Events Questions
In your job, how satisfied do you currently feel at work?
Very Satisfied
Somewhat Satisfied
Neither Satisfied nor Unsatisfied
Somewhat Unsatisfied
Very Unsatisfied
Have you ever taken property
from work without permission?
Never
Occasionally
Daily
Do you plan to leave your job in
the next six months?
Yes
No
Not Sure
How old are you? ___________
What is your highest level of education?
High School
College or Sixth Form
Undergraduate Degree
Postgraduate Taught Degree
Doctoral Degree
Are you a full-time or part-time employee?
Full-time
Part-time
####### The way the question is
####### structured is also important
This looks like an open question BUT is
on a continuous scale. (eg. 1-100 years)
This is a category question only
one answer/category should be
selected.
####### The way the question is
####### structured is also important
Do you think the pay and other benefits you receive are fair?
Yes
No
To what extend do you agree with this statement: My employer cares about my well-being.
Strongly Agree
Mildly Agree
Neither Agree nor Disagree
Mildly Disagree
Strongly Disagree
This is a category question only
one answer/category should be
selected.
This is a rating question using a
Likert Scale only one
answer/category should be
selected, the possible answers are
presented on a continuum.
In your job, how satisfied do you currently feel at work?
Very Satisfied
Somewhat Satisfied
Neither Satisfied nor Unsatisfied
Somewhat Unsatisfied
Very Unsatisfied
Have you ever taken property
from work without permission?
Never
Occasionally
Daily
Do you plan to leave your job in
the next six months?
Yes
No
Not Sure
####### The way the question is
####### structured is also important
This is a rating question using a
Likert Scale only one
answer/category should be
selected, the possible answers are
presented on a continuum.
This is a category
question only one
answer/category
should be selected.
####### What questions should I ask?
The content of your questionnaire should be led by your
research question, and any hypotheses you need to test.
There are three approaches to designing your questionnaire content:
- Adopt a questionnaire which already exists in the literature.
- Adapt a questionnaire which exists, but needs to be changed to fit your research.
- Develop your own questions from scratch.
####### What questions should I ask?
Adopt a questionnaire which already exists in the literature:
- If you want to measure attitudes, opinions or behaviours for a common variable in your subject area, a questionnaire has probably already been developed!
- If an existing measure can provide the information needed, you should check: – Has it been published in a peer reviewed journal? – Do the authors show that it is valid and reliable? If YES, then you can ask permission from the lead author to use the scale.
####### What questions should I ask?
Adapt a questionnaire which exists, but needs to be changed to fit your research:
- If there are questionnaires which ask about similar variables, you may be able to adapt them to suit your purpose the original scale authors should still be credited.
- Again, check the reliability and validity of the scale in the literature.
For Example:
You might adapt a general leadership questionnaire to be specific to a certain organisation.
You might adapt a questionnaire about general well-being, to focus specifically on well-being at work.
####### What questions should I ask?
Develop your own questions from scratch.
- This is the most time-consuming method as questions may need to be tested out beforehand to make sure they are reliable and valid.
- Even when designing your own questionnaire, examples from the literature should guide you to make the questions clear and concise.
- Your dissertation supervisor will be able to give you feedback on your proposed questions to make sure the responses you collect will address your research aims.
- The benefit of using questionnaire(s) from existing literature, is that academics have already spent a great deal of time designing and testing these instruments.
####### An example from my research
One hypothesis in my research is:
Participation in Mindfulness Training at work will be
related to an increase in well-being which is mediated by
an improvement in mindfulness skills when compared to a
control group.
Independent
Variable
Dependent
Variable
Mediator
Mindfulness
Training
(training or no training)
Mindfulness Skills
Measured with the Five
Facet Mindfulness
Questionnaire
(FFMQ; Baer et al.,
2008)
Well-being
Measured with the IWP
Multi-Affect Indicator
(Warr et al., 2010)
####### An example from my research
Other data are required to make the research reliable. These were designed separately with a set of possible responses.
Control variables which may affect the outcomes of the study:
- Age
- Gender
- How many sessions of the course were attended?
- How much mindfulness homework was done?
These are demographic questions
These are questions about
behaviour
####### Structured Observations
In a structured observation, the aim is to quantify behaviour.
Specific behaviours of interest, which have already been decided in line with the research question, are searched for and counted.
Often done using technology – easy to record the activity of many participants, and a great deal of technology use is recorded regularly anyway.
The researcher is a detached observer they have no role in the observed event.
Coding schedules are used which can contain checklists of behaviours to be recorded, either by frequency or sequence.
####### Structured Observations:
####### Flanders Interaction Analysis
####### Categories
Each square represents a 5-second interval.
The observer records which behaviour from the list (if any) is occurring in each interval.
- Teacher asks a question to the group
- Teacher responds to comment
- Teacher explains a concept
- Student responds to a question
- Students asks a question
- Silence
1 1 1 6 4 4
4 2 2 2 6 6
6 5 5 5 2 2
2 2 2 3 3 3
3 6 6 3 3 3
####### Structured Observations
####### Advantages and Disadvantages
Advantages:
- Can collect data in real time, instead of trusting the memory of the participant to report it afterwards.
- Small behaviours that participants may forget about, or not keep track of, can be observed.
- Once a person is trained in the observation measurement, anyone should be able to make an observation objectively. This should also make it reliable as results are replicable.
- Data are collected in their natural setting.
####### Structured Observations
####### Advantages and Disadvantages
Disadvantages:
- If several behaviours occur at once or overlap, coding will be difficult.
- The observer can only measure what they can see/hear
- cannot tell what the underlying reasons or later consequences for the actions are.
- Unless done using technology, the observer needs to be in the location where the observation will occur.
- Data collection can be costly and time-consuming.
- Structured observations can only give limited information on the presence of behaviours other methods will probably be needed to examine why they occur.
3. Understanding Reliability
and Validity in quantitative
research.
Reliability and Validity
######## When using quantitative methods, evaluations are
######## made about the Reliability and Validity of
######## measurement techniques such as questionnaires,
######## and their results.
######## This is different to qualitative research where
######## evaluations relate to Dependability, Credibility ,
######## and Transferability.
######## Make sure you use the correct terms and quality
######## evaluations for your methods.
Reliability and Validity
Reliability: The extent to which data collection techniques will yield similar findings.
- A weighing scale is reliable if it gives the same weight every time the same object is placed on it.
- A weighing scale if not reliable if it gives a different weight each time for the same object.
Validity: The extent to which data collection techniques accurately measure what they intend to measure.
- A weighing scale is a valid way to measure your weight.
- A weighing scale is not a valid way to measure your hair colour.
25kg
Reliability and Validity
Reliability in your research :
The way you measure your variables needs to be as free of measurement error as possible.
- Internal consistency of a quantitative questionnaire can be measured statistically during analysis in SPSS. o By calculating Cronbachs Alpha, we can tell how well the questions in a questionnaire about the same topic overlap each other (are consistent with each other).
- Inter-rater reliability: (structured observations) do different observers find similar results in one event.
- Test-retest reliability: If the same questionnaire is given to someone again, will they answer in the same way?
MASH The Universitys Maths and Statistics Help service, have resources and appointments to help with statistical procedures.
You can book here: https://www.sheffield.ac.uk/mash/bookings/book-stats
25kg
Reliability and Validity
Validity in your research :
There are several types of validity you should consider:
- Construct validity does your questionnaire measure the construct/variable that it is supposed to measure?
- Internal validity can claims be made about the reasons for the findings?
- External validity can the findings be generalised to other settings or groups beyond those in your research?
- Ecological validity do you results make sense in real-life situations?
25kg
Tutorial Exercise
Task (Homework before the tutorial)
With regard to the quantitative question identified in your tutorial group, please write a step-by- step data collection/generation plan, which will include details on:
- A hypothesis to be tested.
- The methods used and the reasons of your choice
- How will you develop them
- How you will employ them
####### Materials are on MOLE
Further reading
- Saunders, M., Lewis, P. & Thornhill, A. (2016).
######## Research Methods for Business Students (7th
######## edition) Chapters 8, 9.3, & 11