research methods
Chapter 1 - Introduction to research
Methods of acquiring knowledge
How are we getting the answers to our questions ex: looking at reviews, getting other's opinions
Non Scientific approaches
Just because they aren’t scientific, doesn't mean not useful, combined in a scientific way to be useful
Can all be used within the scientific approach
Method of tenacity
Beliefs or knowledge that has been over time
Accepting info as true because it has always been believed or present
We as humans cling to superstitions because we believe it will help us or we will do better
Ex: Advertising slogans (repetition); habits; cliches
When we keep hearing the same things, we start to believe it's true (like commercial slogans)
Method of Intuition
A hunch or feeling
Ex: instinct; “gut feeling”; what “I feel like” doing
Method of Authority
Accepting information or answers from an “expert”
Good starting point- quick and easy way to obtain answers
The limitation: of going to authority is that it can be inaccurate
Does not always provide accurate information
Often accepts expert's statements as fact
People can be wrong, Google can be wrong, not all “experts” are always experts
Ex: google, professors/ researchers
Rational Method
Using logical reasoning to draw a conclusion
Premise statement = assuming is facts
Ex: All 3-year-old children are afraid of the dark (premise statement)
Michael is a 3-year-old boy (premise statement)
Therefore, Michael is afraid of the dark (logical conclusion)
Method of Empiricism
Empiricism: Answering questions by direct observation or personal experience
Based on the philosophy that all knowledge is acquired through our senses
The closest we could get to a scientific approach but still has some limitations
Will use it, but needs a little more to it to completely make it a scientific process
Sometimes can’t fully answer questions through empiricism, which could be impossible or dangerous
Scientific approach= Scientific method
The Scientific Method
Formulate specific questions and systematically find answers
Making an initial observation, making a hypothesis, research questions
Combines several previous methods of acquiring knowledge
It is:
Empirical
Structured or systematic observations
Public
Observations can be evaluated
Replication
Shows that they are consistent findings by replication
Can generate more information and insight on how they can be related
Objective
The data outcome isn't skewed by bias
Science vs. Pseudo-Science
Science: Evidence gathered from careful, systematic, and objective observations
Careful about how we collect data, showing the conclusion we have
Pseudoscience: Presented as science, but missing key components
Missing research backing them up
September 5: The Scientific Method
The Scientific Method
Make Observation
Something you see that spurs your interest
Inductive reasoning: Use a small set of specific observations to make a generalized conclusion
Leading to conclusion that may not be accurate
Ex: All brown dogs in the park today are small dogs. Therefore, all small dogs must be brown
Induction = increase
Deductive reasoning: Use general statement to reach a conclusion about specific examples
Ex: All cats have a keen sense of smell. Mittens is a cat, so mittens has a keen sense of smell
Can be inaccurate if saying all and not most
Ex: All swans are white. Jane is white. Therefore, Jane is a swan
Can be inaccurate because Jane could be a white human not a swan
Deductions = decrease
Sources of topics
Informal observations
Practical problems or questions
Previous research
Personal interests & curiosities
Self-reflection
Something in healthcare, maybe the business side of it or the health insurance logistics/ do less people get help because of the fear of health insurance coverage or the health expenses
Form Hypothesis
Variables: factors that change or have different values or different cases
Ex: weather, health status, age
Hypothesis: statement that describes or explains a relationship between variables; “best guess”
Can lead to different observable and measurable predictions
Logical: the logical conclusion of a logical argument
Testable: variables, events, and individuals can be defined and observed
Refutable: can be demonstrated to be false
Positive: must make a positive statement about the existence of something
Ex: age & emotion regulation; The younger you are the more emotional you are
Create Testable Prediction
Take the developed hypothesis and apply it to a specific, observable situation
Ex: When there is an audience present, extroverted participants are going to do better on a task. When there is no audience, extroverted participants are going to do worse on a task.
Original Hypothesis: sleep deprivation affects concentration
Prediction 1: students who are sleep deprived will have more difficulty concentrating
Prediction 2: students who are less sleep deprived will have a greater ability to concentrate
Systematically test prediction
Done through data collection
Descriptive research
Correlational research
Relationships between variables
Experimental research
Cause and effect between variables
Important to record exactly what occurs
Analyze data & draw conclusions
Examine extent to which findings relate to the hypothesis
Do they support, refute. Or refine the hypothesis?
Communicate findings
Share results within and outside the scientific community
Allows for replication of research and future studies
Can also inspire future research
Start the process over again, science is a continuous cycle
Literature Review
Two goals
Gain familiarity with research
Find research to serve as the basis for research idea
Defines the current state of knowledge
Helps identify gaps that your study can potentially fill
Sources
Primary sources: First hand reports of research
Ex: Empirical journal articles (an experience or observation the researcher is reporting)
Secondary sources: Secondhand reports that discusses someone else's observations
Ex: (News article, bibliography, a review, textbooks, introduction section of research reports)
Good starting point: recently published secondary sources
Book or lit reviews
Make note of:
Subject words
Author names
The Article Review
Find primary journal article in library database using advanced search
Collect just enough information
Keep an open mind
Be flexible
Start with looking at the title, review the abstract, skim full text. Read articles & take notes, check references
September 12:
The Research Process
After forming a hypothesis, select a research strategy
Research strategy
What are my goals
Will dictate what kind of strategies you use
General approach and goals of the study
Three different goals that will guide our research
Describe:
Descriptive research
Describes characteristics of a specific variable or phenomena
Data: Presented as averages, percentages or x variables
Ex: On average, local college students study 12.5 hours outside of class each week and get 7.2 hours of sleep each night
Prediction
Correlational research
Measuring the correlation or relationship between two or more variables for each individual
Does not attempt to explain the relationship
Have different shapes or directions it can follow
Relationships can be:
Linear
straight
Curvilinear
Does not follow direct, even relationship
Positive
Direct relationship
As one variable increases the other variable also increases
Negative
Inverted relationship
As one variable increases, the other variable decreases
Correlation does not cause causation
It describes a relationship but does not explain it
Explanation
Answers cause and effect questions about relationship between variables
Creates treatment conditions by changing the level of one variable
Ex: increasing the amount of exercise causes a decrease in cholesterol levels
Independent variable: Intentionally manipulated by the experimenter
Dependent variable: Responds as a result of the manipulation of the independent variable
Measured by the experimenter
APA Style Writing- Research Manuscripts
Manuscript Elements
Title page
Abstract
Introduction
What do we know? What dont we know
Funnel into the study
Method
What did we do?
How did we measure
Who was involved
For other people to replicate the study
Results
What did we find?
Were there relationships between variables?
Discussion
More broad
What did we get from the research?
What are some takeaways
What are limitations?
References
Citing Work
In-text citation
“previous research found”
Is throughout the text
Include author's last name and year of publication
Either
Cite source in parenthesis in sentence
Ex: “Associations between variable x and variable y have been found in past research (Example, 2019).”
Use source as the subject of a sentence
Ex: “Research by example (2019) suggests…”
Paraphrase a point rather than using a direct quote
Use direct quotations only when it is necessary to preserve the essence of the original statement
Reference citation
Put citations at bottom
Reference pages
Author element
Date element
Title element
Source element
Reynolds, S. M., & Wickline, V. B., Bruner, A. R., Steele, E. (2020). Miniature horses have big advantages: Improved stress, mood for both airport travelers and college students. PSI CHI, The International Honor Society in Psychology, 25(1), 77-84.
Writing style
Keep language impersonal
Avoid colloquial sayings or jargon
Sparingly use personal pronouns to describe what you did
Consider verb tense
Results: use past tense (“scores decreased”)
Drawing conclusions on discussion: Present tense (“the data suggests”)
Unbiased language
Call people what they prefer to be called
Use appropriate terms to refer to groups of individuals
Did the author say they were causing something when they actually weren't, using inappropriate words/ it wasn't causal or experimental
September 19: Defining and Measuring Variables
Defining Variables
Measurement: Assigning scores on a variable
Psychometrics: Psychological measurement
Variables: Have different values for different cases
Temperature
Overall fitness
Age
Tangible, easily observed and measured
Ex: height and weight
Intangible, abstract, can’t be easily seen or measured
Ex: depression, self-esteem, knowledge
Constructs: Hypothetical or intangible variables
Help explain and predict behavior in a theory
Help know how someone will react in a given situation
Ex: pain, depression, mood, focus, stress levels
Operational Definitions: How to precisely measure constructs
How are we defining it so we can capture in a researches study
Memory -> recall a list of words
Hunger -> number of hours of food deprivation
How to operationally define variables:
Review previous research
Use conventional/ previously used operationalization
Use multiple types of measures
Self-report
Straight forward
Validity/ accuracy might be questionable
Physiological
Seeing what is physiologically going on in someones body, if we can measure the construct
Set up different measures on someone's skin, measure skin levels
Measure cortisol in body
Brain scans
Expensive
Behavioral
Observing behaviors
What could represent a construct
Looking at someone's sleep
Making sure there are multiple people observing incase you miss an observation, clearly define behaviors
Limitations of operational definitions:
The way we define the variable is not the same as the construct itself
Might leave out important components of a construct
Might include extra components not part of the construct
Ex: measuring style on the amount of designer things they wear. But also measuring their money and inaccurately measuring style because they might not have enough money for designer
Scales of measurement
Nominal scale
Representing qualitative differences in the variable measured
Ex: numbers on a jersey- doesn't actually mean anything “#1 isn't the best on the team”
Ex: sports: baseball, basketball, soccer - doesn't say of one is better than the other or the best just shows the difference
Would be more beneficial to measure how many people said they played that sport
Ordinal scale
Represents differences in ranking
Indicates direction of differences
Who came first second third in a race…
Does not show anything about magnitude
But doesnt show by how much time they won by
Ex: first, second, third; preferences
Interval scale
Numbers mean something
Measuring a series of equal intervals between differences in the variable
Zero point in arbitrary
Not a meaningful number
Ex: Temperature in Fahrenheit; likert scale
Can go below zero degrees, doesn't mean there is no longer any temperature when you go below zero
Ratio scale
Series of equal intervals
True zero point
Zero = absence of something
Ex: weight in points, temperature (kelvin scale)
Can't be zero points
Internal & ratio difference
Interval: zero point in arbitrary
Ratio: zero point is meaningful
Introduction page lab
What is a construct? Something that can not be directly observed or measured. ->
turning that into a measurement: self report, physiological, behavioral
Validity and Reliability of Measurement
Reliability: Degree of consistency of a measure
Is same individuals are measured under the same conditions, results should be the same
Over time, across different items, and across researchers
Inconsistency comes from errors
Sources of error
Observer error: Error made by the individual making the measurements
Environmental changes: Difficult to attain the ideal identical circumstances
Participant changes: Participant changes between measurements
Reliability across time
Test-retest reliability: Compares scores of two successive measurements of the same individuals and correlates the scores
Ex: measuring cortisol levels at multiple different times to find average
Internal consistency
Are these different items in a measure consistently related to each other
Split-half reliability: Splits the test in half, computes a separate score for each half, and calculates correlation between the two scores for a group of participants
During a study, if they correlate then they are close enough/ relate to each other
Reliability across researchers
Inter-rater reliability: Agreement between two observers who simultaneously record measurements of the behaviors
Used for observation of behaviors
Reliability & validity
Reliability is a prerequisite for validity
A measurement procedure cannot be valid unless it is reliable
Validity
Whether a tool measures the variable it claims to measure
Are we actually measuring what we think we’re measuring?
E.g.: Does an IQ test measure intelligence accurately?
Face Validity
Simplest, least scientific
Whether a tool appears to measure what it’s supposed to measure
“Logical” validity
Content Validity
Whether scores on a measure “covers” construct of interest
Concurrent Validity
Whether the scores obtained from a new tool correlate with scores obtained from a more established tool of the same variable
Criterion Validity
Whether scores on a measure correlate with other variables expected to be related
Measure criterion in the future = predictive
Predictive Validity
Whether scores obtained from a tool accurately predict certain relevant behaviors
Convergent Validity
Strong relationship between scores from 2+ different methods of measuring the same construct
Divergent Validity
Showing little or no relationship between the measurements of 2+ different constructs
Other Aspects of Measurement
Multiple measures
Sensitivity of measurements (range effect)
Ceiling effect: Scores cluster at high end of scale
Floor effect: Scores cluster at low end of scale
Research Ethics
Researchers responsibility to respect all individuals affected by studies
Ethics does not equal morality
They are related but are not the same
Morals has a more of individual connotation
Can be impacted by our upbringing, religion
Ethics is more of a right and wrong over a general sense
Researchers are in a position of power/ control
An element of influence
A sense of authority
Feel pressure of what they are asking of you if say the researcher is in the front of the room
Must ensure:
Welfare and dignity of participants/ subjects
Accuracy in reporting results
History of research ethics
WW2: Nazi “experiments” on unwilling participants
Unwilling participants
Resulted in torture
Held accountable in trial for things they did
Nuremberg Code (1947)
Created after WW2
To guide how humans will be treated in research
Ten guidelines for ethical treatment of human subjects
Served as future guideline for how ethical principles will be developed
Foundation for current ethical guidelines
Tuskegee Syphilis Study
Led to the Belmont Report (1979)
Belmont Report
Principle of Respect for Persons: Individuals must be free to consent to research; those unable to consent need special connection
Children, prisoners, people with disabilities, more vulnerable populations like pregnant people should have more protection
Principle of Beneficence: Researchers must minimize risks and maximize benefits
Principle of Justice: Requires fair and non exploitative procedures for the selection and treatment of participants
How important it is to have established ethics
APA Guidelines
Ethical guidelines for the treatment of human participants in research (1973)
Ethical codes are constantly evolving as needed
October 5:
Privacy and Confidentiality
Researchers must protect participants confidential information including:
Attitudes and opinions
Because it can add bias
Can also be personal information
Measures of performance
To protect participants identity and information
Demographic characteristics
Informed Consent
Must inform participants about the study so they can voluntarily decide to participate
Tell participants what will happen in the study, but not why
Provide information in accessible language
Obtain informed consent from legal guardians, when necessary
Participant can decline to participate without consequences
Deception
Purposefully withholding information or misleading participants regarding the study information
Passive deception (or omission): Withholding information
Not telling them the whole purpose of what you're going to be doing
Active deception (or commission): Intentionally mislead participants
Confederates: “Pretend” research participants who actually are hired/ work for researcher
Benefit needs to outweigh the risk
Cannot conceal factors that may cause physical harm or severe emotional distress
Must provide a debriefing after participation
Debriefing
Researchers must provide an explanation of studies purpose afterwards
Includes:
The true purpose of the study
The contribution of the study
How and why deception was used
How any negative effects may be minimized
Any other answers to participants questions
Institutional Approval
Researchers must submit proposal to their governing board to receive approval before conducting research
Institutional Review Board (IRB)
Composed of scientists and nonscientists
Examines proposed research involving humans
Minimization of risk to participants
(notes)
Plagiarism & Fraud
To falsify or misinterpret data intentionally
Why do researchers commit fraud?
To remain competitive in their academic environment
“Public or Perish”
Avoid Fraud
Allow for replication of studies
Peer review process
Data must be shared with others if requested
Establishing consequences of being guilty of fraud
Protecting Animal Participants
Why study animals?
Understand animal psychology for their own sake
Understand human behavior by generalizing it from animals
Conduct research that is impossible to conduct with human subjects
General care and maintenance of animal subjects
Maintain adequate housing conditions, food, sanitation, and medical care for research animals
October 10:
Methods Section
Provides a detailed description of:
Who were the participants in the study
How the variables were defined and measured
How was the study carried out
Allows for future research to replicate what has been done
Participants Subsection: Tells us who was involved in the study (the number of participants, eligibility/ exclusion criteria, demographic variables, and any other relevant characteristics)
Measures/ Materials Subsection: Identification of the variables in the study and how they were operationalized (defined and measured)
How are you defining and measuring the variables
Procedure: Step-by-step process used to complete the study
What did they do?
Selection procedures
Settings and locations
Any payment to participants
Survey Research
Survey research: Quantitative & qualitative method
Uses self-report (questionnaires or interviews)
Pay attention to sampling
Goal: Obtain a “snapshot” of a group at a particular time
Can measure variables and demographics
Conducted via phone, mail, online
Typically nonexperimental (descriptive, correlational) but can be used in experiments
Constructing Questionnaires
Open-ended: Allows participants to respond in their own words
“What do you think of the food on SU’s campus?”
Pros: Flexibility
Cons: Different participant interpretation; more challenging to analyze
Help by being more specific in the question
Types of questions
Close-ended (restricted): present a limited number of response alternatives
“What is your favorite grocery store?”
Aldi
Weis
Wegmans
Other (please specify): ___________
Pros: Easy to analyze
Cons: More limited in answer options
Rating scale: Select a numerical value on set scale
Ex: To what do you agree with this statement?
“Stats is easier for me than it is for most students”
Strongly disagree
Disagree
Neutral
Agree
Strongly agree
Pros: Easy to understand
Cons: potential for response set
Writing survey questions
Brief
Relevant
Unambiguous (clear)
Can only be interpreted in one way
Specific
Objective
Avoiding two questions in one
Suggestions
Using existing measures
Reliable
Valid (measuring what is intended to be measured)
Review relevant research
Compare across studies
Consider the order of questions
Sampling
Place demographic questions at the end
Place sensitive questions in the middle of the survey
Group together questions with the same topic
Keep format/ language relatively simple
(Control variables beyond emotions) (note whatever else they collected data on)
Probability sampling
Odds are choosing person in population are known
Know exact size of population and all individuals
Individuals have equal chance of being chosen
Simple random sampling
Randomly selecting participants from list of population
Equal chance of selection
Selections are independent
Systematic sampling
Choosing every nth participant from the list
N = (population / sample)
If population = 5,000 and sample = 100, choose n = 50
Stratified random sampling
Divide population into strata (into different groups)
Randomly choose from strata
Guarantees representation, but may overrepresent some groups
Proportionate stratified random sampling
Same as stratified random sampling
Proportions in sample = proportions in population
Cluster sampling
Selecting clusters (pre existing groups) in the population
Ex: Sampling entire classrooms
Problem: Not really random
Nonprobability sampling
Odds of selecting individual = not known
Convenience sampling
Selecting individuals available and willing to participate
To offset issues:
Try to make sample representative
Quota sampling
Create quotas of individuals to sample
Researcher who is looking to obtain a sample of 30 students -> set quota of 15 girls and 15 boys
When quota of 15 girls are met, stop recruiting girls
Snowball sampling
Snowball sampling: Existing research participants help recruit additional participants for the study
Experimental research
Observe cause & effect between two variables
Manipulate the IV
Changing in an intentional and systematic way to create different conditions
Measuring the DV
How does the variable change after manipulating IV
Compare scores across conditions
Based on different manipulations and comparing them
Control all other variables
Controlling everything else that is not related to IV or DV
Directionality problem: Relationship between two variables does not explain which is the cause and which is the effect
When the relationship between two variables don't say what is the cause and what is the effect, don't know what causes the other
Third-variable problem: When a third (unidentified) variable is responsible for relationship between two other variables
Treatment conditions: Experimental groups of participants that differ along the value of the IV
Levels: Different values of IV
Define different treatment conditions
Manipulation
Control group & Experimental group
Manipulation & the third variable problem
Control of other variables (other than the IV & DV)
Researcher must eliminate all confounding variables
October 31:
Inferential Statistics
Hypothesis test: Determines whether sample data provides evidence to support hypothesis
Ho (no difference, no relationship, nothing really exists) (null hypothesis)
HA (research hypothesis)
Errors in Hypothesis Testing
Type 1 error: Rejecting Ho when it is actually true
Concluding there was a significant effect, when there wasn't one in reality
Type 2 error: Retaining Ho when it was actually false (failed to notice it when doing hypothesis testing)
Concluding there wasn't a significant effect, when there in fact was
Factors that Influence Hypothesis Testing
Increase number of scores and participants in sample
Larger sample size, more evidence there is a significant relationship between variables
Make sure to measure variables to get enough variance/ difference in scores
A sample mean with high variance
How do I know which data analysis strategy to use?
Use the right statistics
Depends on your data
Research design
Type of variable
Continuous, categorical (nominal)
Scales of measurement:
Nominal, ordinal, interval, ratio
Descriptive research: One group of participants; not interested in relationships between variables
Nominal: mode
Ordinal: median
Interval/ ratio: mean and standard deviation
Correlational research: One group of participants; examining relationship between 2+ variables
Pearson correlation
Regression
Chi-square test for independence
Experimental research: 2+ groups of participants’ comparing means of scores on certain variables
T-tests
(1) one-sample t-test; (2) two-sample t-test; and (3) two-sample paired t-test
used to determine if there is a significant difference between the means of two groups and how they are related. T-tests are used when the data sets follow a normal distribution and have unknown variances
One-way ANOVA (2+ levels of IV)
Two-way ANOVA (2+ factors) (more than one independent variables)
Nominal or ordinal scales: Chi-square test for independence
Regression analysis: A commonly-used analysis technique, which examines the relationship between variables
Factorial ANOVA: Compares means across two or more independent variables
Factorial Research Designs
Factorial Designs
Experiments with 2+ factors (IVs)
Highlights how factors independently and jointly influence the DV
Can use for between-subjects design, within-subjects design and mixed design
Can include both true IV and quasi-IVs
Reduce variance in between-subjects designs
Replicate and expand a previous study
Evaluate order effects in within-subjects design
Main Effects
How each independent variable affects the dependent variable
Psychotherapy intervention (IV) -> Wellbeing (DV)
Motivation to change (IV) -> ^
Interaction effect
How two factors can interact to have different effects on dependent variable
When they both interact you see the most benefits
Interaction
When drugs and alcohol are combined they are working together to do the same thing and can be potentially deadly
Two factor study will have 2 main effects and an interaction effect
Notation
Shows number of factors and their levels
(2x3) the number of factors
^ the actual number tells us the levels of factor
A table can be used to show separate treatment conditions and their main effects
High-self esteem Low-self esteem
Audience
No audience
If lines are not parallel then there is an interaction happening
If the lines are parallel there is going to be no interaction
That we didn't find an effect on perceived group defensiveness
Based on findings there was not a significant difference of perceived groups defensiveness
Limitation: smaller sample size, elements of the design/ what we could do differently/ for future research
Want to link it back to what previous research has found
Why did we find the findings that we did
Conclusion section is optional
Discussion Sections
Final written section of the research report
Restates hypothesis (why?) and summarizes results
Interprets findings, implications, and possible applications of the results
Also addresses limitations and areas for future research
Non Experimental & Quasi-Experimental Research
Internal Validity: Degree to which causal relationship between variables has a single, clear explanation
Not explained by some other variable
Why do experiments tend to be high in internal validity?
(Hint: think about the design of experiments)
Overview
Non-experimental & quasi-experimental can “look like” experiments
Between-subjects (AKA nonequivalent group designs)
Within-subjects (AKA one group designs)
Between Groups Design
Differential Research Design
Compares pre existing groups to establish differences between them
Example: Self-esteem scores between children in 2 guardian vs. single guardian homes
Posttest-Only Nonequivalent Control Group Design
Compares pre existing groups’ scores after treatment
Group 1: [Treatment] – Measurement
Group 2: [No Treatment] – Measurement
Pretest-Posttest Nonequivalent Control Group Design
Compares two nonequivalent groups before and after treatment
Group 1: Measurement – [Treatment] – Measurement
Group 2: Measurement – [No Treatment] – Measurement
Within-Subjects Designs
Pretest-Posttest Designs
Compare scores for one group before and after treatment
Group 1: Measurement – [Treatment] – Measurement
Threats to internal validity:
History, testing effects, maturation, regression to the mean
Time Series Designs
Measures participant scores in a series before and after treatment or event
Group 1: Measure – Measure – [Treatment] – Measure – Measure
Review / Examples:
Dr. Jackson is interested in the effectiveness of an anti drug education program on high school students’ attitudes toward illegal drugs. He measured the attitudes of students during week 1, implemented the anti drug program week 2, and measured attitudes again week 3.
o x o - pretest-posttest design
Dr. Michaels is interested in the relationship between emotional regulation and traumatic brain injuries. She measures levels of emotional regulation between athletes who have experienced TBIs and athletes who have not experienced TBIs. She then compares emotional regulation scores between the two groups.
o - differential
Developmental Research Designs
Nonexperimental research to study behaviors related to age
Cross-sectional
Longitudinal
Cross-sectional
Compares groups of different ages
Example: compare 20 year olds, 30 year olds, 40 year olds
Longitudinal
Compares same group as they age
Example: measure same group of individuals across 10 years (20 years→ 30 years)
November 21: Correlational Research
Purpose of correlational research:
Measure relationship between variables
No attempt to manipulate variables of interest or control extraneous variables
Strengths
Studies things that can’t be examined experimentally (not feasible or unethical)
High external validity
Limitations
Can’t assess causality
Third-variable problem
Directionality problem
Low internal validity
Measuring the Relationship Between Variables
Characteristics of Correlations
Direction
Whether relationships are direct or indirect
+ or -
Strength (or magnitude)
How strong is the relationship between variables
0.0-1.0
Direction
Positive relationship: Variables change in same direction
Variable X ⬆ Variable Y ⬆
Variable X ⬇ Variable Y ⬇
(AKA direct relationship)
Negative relationship: Variables change in opposite direction
Variable X ⬆ Variable Y ⬇
Variable X ⬇ Variable Y ⬆
(AKA indirect relationship)
Strength
Measured using Pearson’s correlation coefficient (r):
Sign (+/–) indicates the direction of the relationship
Numerical value (0.0–1.0) indicates strength
Strength of Relationships
Strength - continued
Statistical significance means that the correlation is unlikely to have been produced randomly
With a small sample, it’s possible to obtain a strong correlation when there’s no relationship between the variables
Increasing sample size increases likelihood that a correlation represents a real relationship
Correlation does not imply Causation
Relationships
Correlational designs are more likely to be impacted by third-variable or directionality problems
Third-variable problem: Relationship between 2 variables is due to some other 3rd (unidentified) variable
Example: More churches = more crime because both are found in higher population areas (3rd variable)
Directionality problem: Unclear if a relationship between 2 variables is due to X causing Y or Y causing X
Example: Unclear if violent video games cause aggressive behaviors or if more aggressive kids seek out violent video games to play
Third variables can create spurious correlations
When variables are statistically related but are not directly related
Complex Correlation
Researchers can examine relationships among multiple variables
Example: How does income and health relate to happiness?
Application of Complex Correlation
Factor analysis
Identify underlying factors across a larger set of variables
Statistical control
Can include potential third variables in the analysis to help control for them
Regression
Can predict one variable given another variable
November 28: Descriptive Research
Descriptive Research Strategy
Goal
Observe and describe variables as they naturally occur
Often used to begin to understand a topic that has not been previously researched
Observational Research
Researcher observes and records behavior of individual(s)
Ex: Observing people in a park, looking at interactions, in classrooms
Behavioral Observation
Naturalistic Observation: No researcher intervention
Ex: Shoppers behavior at the store, Children's playground interactions
Participant Observation: Researcher interacts with participants and becomes one of them
Researcher becomes one with the population they are studying
More involved
Ex: Ethnographic research
Structured Observation: Researcher sets up a situation likely to produce the desired behavior in participants
Less natural
Ex: observing parent-child interactions in lab
Observation Considerations
Do not disrupt or influence behaviors
Demand characteristics and reactivity
Hints or clues that a researcher might give to tell participants what is expected to be found
Reactivity: Engage in behavior because they know they're being observed
Can be addressed by:
Concealing observer - blend in, dress a certain way, hidden
Habituating participants to the observer- get people used to them being there if you can't hide them; be in space for a while before starting observation
Observations are subjective
Can be addressed with:
Inter-rater reliability- degree of agreement between two different observers
Cohen's Kappa- how similar are the rates and responses
Percent agreement- how closely the responses align
Quantifying Observations
Frequency method: Count instances of each specific behavior
Duration method: Fixed-time observation
How long they engaged in the behavior
Interval method: Divide observation period into a series of intervals and record behavior during each interval
Making Observations
When observing complex situations:
Record and replay situation to gather observations
Three techniques:
Time sampling: Record observations in intervals
Event sampling: Focus on one behavior at a time
Individual sampling: One participant at a time
Survey Research
Case Study Research
November 30: Abstract
Purpose
Identifies key elements of the study and paper
Summary of the study- gives readers an overview of the study/ main idea
Includes (1-2 sentences each)
General statement for each pieces of intro, method, results, discussion
Statement of the problem/ research topic
Hypothesis / prediction(s)
Methods used (number of participants, materials, procedure(s)
“A total of were surveyed using qualtrics/ were randomly assigned”
Findings / results
Conclusions or implications (broader takeaways or applications of the findings)
“This study adds to the existing knowledge of in-group and outgroup findings”
“These findings can help increase insight on — topic”
Formatting
Written as single paragraph at the very beginning of your paper
On its own page after title page before introduction
150-250 words
Importance
Helps readers determine the relevance of your paper to their research
Should have enough information to make sense to someone without reading entire article
December 5:
Chart
Descriptive Research
Content Analysis
Measures occurrence of events in media
Literature, movies, TV, etc.
Establish behavioral categories
Use frequency, duration, or interval methods
Use multiple observers
Qualitative Research
Explores meaning/ understanding of phenomena, experiences, perspectives, and constructed reality
Non-numerical data
Collecting Qualitative Data
Field notes
Description of what you're observing
Internal feelings in the setting
Interviews
Semi-structured interview conversation
Before: Construct/ pilot interview questions
During: Record (with consent); use a quiet space; talk less/ listen more; build rapport with participant
After: Keep notes of any pattern or thoughts; transcribe & analyze data
Focus groups
Group interview
Typically 1 interviewer, 5-8 participants
Pro: More participants at once
Con: Participant (dis)comfort with discussing topics
Which one to use: Qual or Quant?
Mixed Methods
Combines qualitative & quantitative approaches
Examples:
Qual -> hypothesis generation,
Quant -> test hypothesis
Survey/ interview with both quant and qual questions