Please enable JavaScript.
Coggle requires JavaScript to display documents.
W1 - Advanced Research Methods Introduction, W3 - Scale Construction, W2 …
W1 - Advanced Research Methods Introduction
Experimental Design
Manipulates the IV
Measures the DV
Can assume causation
Measures differences between groups
Experimental group
control or comparison group
random assignment
Statistical tests
Two groups = T-test
More than 2 groups = ANOVA
Correlational designs
Examine relationship between Vs that are NOT manipulated
Cannot assume causation
Why?
Reverse causation
maybe get the predictor mixed up
the effect of unmeasured V
Can assume they relate but not causation
Positive correlation = as scores on one increase scores on the other also increas
Negative correlation = as scores on one increase scores on the other decrease
statistical tests
correlational (relationship between 2 vairables)
Regression (one or more Vs predict scores on another V)
Survey Methods
good at measuring DV
inexpensive
easy to administer
disadvantages
response bias
unclear wording
could trigger want for social desireability
poor time resolution
participant guesses right/wrong answer ("demand")
poor for measuring things people are unaware of
W3 - Scale Construction
Scales
Quantify constructs
Represents a set of items to measure a construct
Data collected - provides insights that can be analysed to make informed decisions
Principles
Clarity of items
Ambiguity can lead to undermining the validity of results
Avoidance of double-barreled items
Make sure an item only measure 1 concept in 1 item
Balanced scales
Helps mitigate repsonse bias
Strongly Disagree neutral or strongly agree
Steps in scale construction
Define the construct
precise definition of what youre trying to measure
Eg; testing smart phone addiction - you must first define what constitutes as addiction in this context
Generate Initial Pool of Items
serves as fundamental starting point
Refine items
Ambiguous or redunadnt items should be removed
Not long and boring - participants will reflect this in results
Some questions are allowed to be similar (to test for consistency) however try not to be repetitive
Types of Scales
Likert scale
Prevalent
Series of items with agreement choice
Strongly disagree to Strongly agree
Show in between options - agree, neutral, disagree
5-point scale
Semantic Differential Scales
Measuring meaning of concepts by ranking them on opposite dimensions
Eg; rate on modern vs traditional, cheap vs expensive, trustworhty vs untrustworthy
Only show the two extremes as options - can choose somewhere in the middle but no label for "neutral"
Thurstone Scales
Order of items has been specifically chosen with predictions
also known as "equal appearing intervals"
The designer assigns scores to each response based on the items position
Eg; how do you like our service? 1,2,3,4,5,6,7,8,9,10
Cultural sensitivity
ensure scale items are culturally appropriate
Relevant to target population
Consider language, cultural norms etc
typically youre dealing with a diverse population
Respnose format
Decide whether numerical, visual or verbal scale will work best- be more accurate
Eg; consumers rate attraction to visual brand packaging
Pilot Testing
Facilitates refinement of the scale
Ensures it is comprehnsible and meaningful
Will show shortcomings
Enhances reliability and validity
Scale Evaluation
The measurement tools used to quanitfy/measure abstract concepts or variables
Ensures reliability and validity
Must be accruate, consistent and appropriate for study's objective
Types of Scales (types of data)
Nominal
Used for labelling variables without any quantitative value or ordering
Categorise data into groups/classes
Eg; male, female, non-binary
Eg; toyota, ferrari, mercedes
Ordinal
ranking among categories - not consistent or measureable
Eg; likert scale
There is an order but difficult to measure the exact same
Eg; high school diploma, bachelors degree, master degree, PHD
Interval
Maintain order
Have equal intervals between points
Lack a true zero point
Eg; temperature 0 degrees doesnt mean a lack of temperature
Eg; Calendar dates
Ratio
Have all properties of interval scale PLUS a true zero point
Eg; Age 0 years = absence of age, weight of 0kg = absence of weight
Importance of choosing the correct Scale Type
Nature of variables influences choice of scale
Eg; if you're measuring something categorical without any order like gender- nominal scale is appropriate
Research objectives
If you need to perform statistical tests eg; the mean - you need at LEAST an interval scale
if you require a true zero a ratio scale is needed
Reliability in scale eval
Refers to consistency and stability of measurements over time
internal consistency = consistency of reposnses within the scale items
test retest reliability = consistency of responses over time
inter-rater reliability = consistency of responses between different raters/different people
Methods for assessing reliability
Methods for testing reliability
eg; Intra class correlation for test retest
Validity in scale evaluation
Validity = extent to which a scale measures what it intends to measure
Types
Content validity = measures all aspects of the construct
Criterial = scale predicts or correlates with external criterie
Construct validity = the scale measure the theoretical contstruct it claims to measure
Methods of assessing validity
Content analysis to measure content validity
factor analysis to measure consructive validity
Correlation analysis to measure criterial validitiy
Pitfalls in Scale eval
Overlooking cultural or contextual differences
Using scales without propervalidation
Ignoring participation
Best practicies
Ensure appropriate sample size
Use multiple methods for validation
Seek expert feedback
Conduct thorough literature review
W2 -Ethical and Cultural Considerations in Psychological Research
Research Ethics
Acting in the right spirit out of respect and concern for others
Should be considered in
Design of research
Respectful to individual and society
Ethical reporitng
Ethical use of resources and outputs
Importance
Maintain rights, welfare and dignity of particpants
Ethical principles
Honest, respectful
Confidential
Use research to benefit others
Should follow NHMRC, National Statement on Ethical Conduct in Human Research (2023)
You must submit an application of ethical consideration when conducting research
Human Research
With or about people, their data or samples
Surveys/interviews/focus groups
Undergoing treatment - psychological, medical etc
Accessing or using personal data/records
Being observed
Collection of biological samples
Values
Research merit and integrity
Compels researchers to be honest in all aspects of research - review of literature design, methods, implementation, use of results
Justice
Selection and inclusion of participants
Fairness in recruiting
Benefits of research is fairly distributed
Beneficience
Benefits of research must outweigh risk, harm
Convey risks to participants
Respect
Recognition of intrinsic value
Respect for customs and cultural heritage
Culture
Large part of who we are/how we perceive our environment
Cultural norms and values help us understand peoples ways of doing things
Individualist cultures - more independent, autonomous, self focused
Collectivist - group oriented, interdependent, maintaining harmony
Cross cultural differences
Survey questions could be phrased in ways that may be perceived differently
Eg; "i feel responsible for my brother's failures" - collectivist yes more likely to
Sense of obligation for people in collectivist culture to participate in research where people in individualist cultures tend to be less influenced by social pressures
Language of the research
Eg; may not be same words for same experiences in all cultures "Depression" in Hindi there is no singular word for it
Condescending (Eng) vs condescendiente (Esp) = down to earth and nice
Be careful with translations
Cultural lens
We bring our own cultural understandings to our research
Our cultural way of being can impact the people we study
Eg; white people studying black people, history of being subjugated by us
Step out of our own value system and look at wide lens
Cultural safety
Feeling respected as a participant
Only determined by the individual
Avoid stereotypes
Develop trust
Understand influence of culture shock
Having humility
Cultural Humility
Process of self exploration, self criticism to elarn from others
Engagement with another person with intent of respecting their values, customs, beliefs
Culturally humble
Ask how your bias may impact
Cultural competency
Have genuine understanding of culture - without bias
Engagement in ethically and socially just research is not just about following a code of ethics, its having ethical sophistication
Establish relationships
Looking after people and their knowledge
National Statement on Ethical Conduct in Human Research (NHMRC 2023) :
Reciprocity
Equitable and respectful engagement with and inclusion
Respect
Respect rights of people to bold and express different values, be aware how your own beliefs may bias views of others
Equity
Having equitable partnerships with communities, actively engage with people and communities, inclusion of research
Responsibility
Minimise harm- accountable to individuals, family and community
Survival
Protection
Spirit and Integrity
Respect past present future generations knowledge
Cultural continuity (ATSI only)
Respecting personal and collecitve bonds to community and find ways to work toegether
Men, Heart and Mind Study (Brown, Mentha, Howard, Rowley 2016)
Differences in expression and consequences of depression across cultures
Very few culturally specific measure of depression or stress
Researchers conducted interviews with ATSI people to develop a model of what psych stress looks likes
Questionnaire contained items that would be relevant
Piloted on 189 ATSI men
Chronic stress, social deprvation formed experience of stress
Sense of injury
Weak spirit, homesickness, worry, anger
ERLI (Jones, Collyer et al 2020)
Purposeful dev of multilingual checklist to respectfully cater for diversity of language backgrounds in indigenous children
Adaption of the Macarthurt BAtes CDI
Dev collaboratively in partnership with local indigenous people and local education and health services
Culturally appropriate language assessment
Considers individuals, family, community factors
Phase 1: Interviews to dev checklist
Phase 2: Administering checklist to caregivers
Key findings
ERLI has culturally relevant items
Improvements in vocabulary with age, a different pattern to the prievious literature with standardised mainstream assessments
Has clinical utility
Bivariate correlation and Bivarite Regression
B Correlation
example : link between studying hours and academic scores
3 types of correlation
Positive
Negative
NO correlation
Assumptions of Pearsons Correlation
Linearity
Relationship between x and Y
Ways to check = scatter plot, linear regression, residual plot
Independence
Different observations of x,y are independent from each other
Experimental design and data collection method review
Normality
both x and y must follow a normal/gaussian distribution AKA bell curve
QQ plot, shapiro wilk/kolomogrov
Equal variance or homoscedasticity
the variance of y doesnt depend on x. spread of y values remains constant
way to check = Breusch-Pagan test
Interpreting correlation coefficients
if r is smaller than 0.3 = weak relationship
if r is between 0.3-0.7 = mid relationship
over 0.7 = strong relationship