PSY 212 FINAL EXAM

Steps of the Scientific Process

1. observation
2. question
3. hypothesis
4. test
5. analysis
6. conclusion
7. REPEAT

theory-data cycle

scientists collect data to test, change, or update their theories

Applied Research

to tackle real world problems
- research you can apply to the world

Basic Research

knowledge just for knowledge
- research just to know

Elements of a Journal Article

1. abstract
2. introduction
3. method
4. results
5. discussion
6. references

abstract

- general summary of the article
- always start with the abstract

introduction

- will go through other existing research
- might mention theory that based its entire study on

method

- will always mention participants
- how they were recruited
- will always talk about material used in the study
- anything they're using in their study
- will include measures & procedures

results

- more technical
- will try to summarize what certain terms mean

discussion

- go over major findings and why they found it
- talk about limitations & future directions

references

- will cite every article for this paper
- any methods that they borrowed, etc..

searching for a journal article: SEARCH TERMS

- "or" \= gives you more
- "and" \= gives you more
- "not" \= gives you less
- * asterisk \= helps you get all related terms

searching for a journal article: DATABASES

- PsycINFO (library in UH)
- Google Scholar

Types of Hypotheses

- describing a phenomenon or population
- difference between groups
- relation between variables
- mediator of relation between variables
- moderator of relation between variables

describing a phenomenon

- focuses on a SINGLE variable within a SINGLE population (ONE variable, ONE population)
- works in:
--\> percent
--\> proportion
--\> frequency
--\> amount

difference between groups

- focuses on the difference between two or more groups on a variable
--\> observational
--\> causal
- you can look at only 2 or many more

relation between variables

- focuses on a positive or negative relation between two variables
- POSITIVE relation: as one variable increases, the other increases
- NEGATIVE relation: as one variable decreases, the other decreases

mediator of relation between variables

- focuses on a variable that EXPLAINS the relation between two other variables
- EXPLAINING a relationship !!
- EX: pizza --\> increase work motivation (the mediator) --\> productivity

moderator of relation between variables

- focuses on a variable that CHANGES the relation between two other variables
- CHANGES. a relationship
- EX: experience increases, salary also increases
--\> moderator of GENDER may change strength/direction

Types of Research

- descriptive
- correlational
- experimental

descriptive research

- aims to observe and describe naturally occurring phenomena or behaviors

correlational research

- aim to describe and PREDICT behavior
- explain the relation between two or more variables

experimental research

- aim to demonstrate a causal relation
--\> A causes B
- its own category with its own goal
- the only type of research that can provide a cause & effect relationship

Research Claims

- frequency claims
- association claims
- causal claims

frequency claims

- describes a particular rate or degree of a single variable
- EX: 15% of undergrads live on campus
- telling you the frequency of behavior

association claims

- when one variable changes, the other variable changes too
- two measured variables
- TERMS: predict, covary, related, associated, correlated
- EX: in a scatterplot, it shows us the pattern of responses between two variables

causal claims

- a claim arguing that a specific change in one variable is responsible for influencing the value of another variable
- TERMS: cause, increase, decrease, influence, affect, change

Measured Variable

- variable whose values are observed and recorded
- EX: note-taking (electronic or written)

Manipulated Variable

- variable whose values a researcher controls
- EX: assign how you note-take

Nominal Variable

- variable whose values are NAMES
- sometimes referred to as a categorical variable
- EX: religion

Numeric Variable

- variable whose values are NUMBERS

interval variable

- distance between each consecutive value is equal
- EX: age (53 & 54 yrs are considered the same)

ratio variable

- has an absolute point
- zero means a lack of that variable

ordinal variable

- the distance between each consecutive value is not equal
- EX: placement in marathon (1st, 2nd, 3rd place)
--\> distance between each runner may not be the same

Independent Variable (IV)

- X
- manipulated
- applied in experimental research

Dependent Variable (DV)

- Y
- measured
- applied in experimental research

Conceptual Def.

- focuses on the variable as an abstract, theoretical construct
- what IS happiness?

Operational Def.

- focuses on the observable or measurable construct
- how will I MEASURE happiness?

Measurement of Variables

- self-report
- other-report
- observational
- physiological

self-report measures

- ask people to report their own thoughts, behaviors, and experiences either in real time or retrospectively
- BENEFITS:
--\> often efficient
(you can do it very easily and inexpensively while getting lots of info.)
--\> privileged access
(you would be able to tap into what people's thoughts and perspectives are just by asking them)
- DRAWBACKS:
--\> potentially biased
(in the way that questions are asked which may not be accurate)
(the answers may be biased AKA providing a socially acceptable response)

other-report measures

- ask teachers, caregivers, or other observers to report about someone else
- BENEFITS:
--\> often efficient
--\> different perspective
(getting more of an unbiased perception by asking somebody else)
- DRAWBACKS:
--\> potentially biased
--\> potentially distorted

observational behavioral measures

- capture a variable through behavioral expressions of it
- BENEFITS:
--\> more objective
--\> EX: counting how many drinks someone has in a night
- DRAWBACKS:
--\> less efficient
(time-consuming)
--\> situational
(distorted based on the context in which you're measuring)
(behavior may be different in different contexts)

physiological measures

- capture a variable through physiological indications of it
- BENEFITS:
--\> more objective
- DRAWBACKS:
--\> less efficient
(very time-consuming and expensive)

Three Types of Reliability

- test-retest
- interrater
- internal

test-retest reliability

- consistency in the pattern of scores
- most relevant for constructs that are theoretically STABLE
- EX: personality, intelligence
- EX: measuring head circumference
--\> you should still get scores in the same range

interrater reliability

- independent observers come up with consistent findings
- most relevant for observational measures
- EX: operationalizing aggression on a 1-10 scale

internal reliability

- only relevant for measures that combine multiple items that measure the same construct
- people should respond consistently across items
- EX: do you get energy from people? \= yes
--\> which means: do you get energy from being alone? \= no
- ANSWERS MUST BE CONSISTENT

Construct Validity

- captures how well a conceptual def. has been translated into an operational def.

Five Types of Construct Validity

1. face validity
2. content validity
3. criterion validity
4. convergent validity
5. discriminant validity

face validity

- does it appear to measure what it says it is measuring?
- align well with conceptual def.
- subjective (making your own judgement calls)

content validity

- does the measure capture ALL PARTS of the defined construct?
- EX: student achievement defining performance in school (operationalization \= measure test scores in math class)
--\> low content validity because it's only looking at ONE content of that goal
- subjective (saying "that doesn't capture what it needed to capture")
- EX: depression

criterion validity

- the measure is associated with a concrete behavioral outcome it theoretically SHOULD be associated with
- it SHOULD make sense it is associated with, theoretically

convergent validity

- does the measure show a similar pattern to other related measures?

criterion validity VS convergent validity

- criterion: focuses on KEY BEHAVIORAL outcomes
- convergent: focuses on PATTERN among SIMILAR measures

discriminant validity (AKA divergent validity)

- does the measure discriminate between constructs that different?
- what the construct involves & doesn't involve

* reliability and validity are two independent measures of assessments *

a measurement can be reliable without being valid
- however if a measurement is valid, it is most likely reliable

Internal Validity

- the extent to which we can identify a single, trustworthy explanation of results

- threats to internal validity \= aspects of the study that limit interpretation of its findings

External Validity

- the extent to which we can GENERALIZE results of research beyond specific instances in the study

- threats to external validity \= aspects of the study that limit application of its findings

ways assess validity empirically

- scatterplots
- correlation coefficients
--\> can assess criterion, convergent, and divergent validity
- known-paradigm
--\> can assess criterion validity

ways to assess reliability empirically

- scatterplots
--\> test-retest reliability
--\> interrater reliability

- correlation coefficient
--\> test-retest reliability
--\> interrater reliability
--\> internal reliability

Population

- the people, places, or things you want to learn more about through research (want to make a conclusion about)
--\> population of interest
--\> target population

Sample

- the smaller subset of the target population included in your study
- EX: UH students \= pop.
--\> UH students in a certain class \= sample

Sampling Methods

- probability sampling
- non-probability sampling

probability sampling

- every member of the target population is known and their selection into your sample is through a RANDOM PROCESS
- each member of the target population has AN EQUAL CHANCE of being selected for the sample

- simple random sampling -

- every member of the pop. has an equal chance of being selection
- time-consuming, sometimes impossible, very expensive
- EX: using a random number generator
--\> birthday party \= putting names into a jar and choosing names 4 teams

- stratified random sampling -

- individuals are randomly selected from predetermined categories
- within each category, you are still going through a randomized process
- EX: race, gender
--\> studying taxes \= stratify based on income and randomly sample within income brackets

- other types -

- systematic \= sample based on two randomly selected numbers
--\> EX: start with 7th person, select every 3rd person
- cluster: randomly selecting from pre-existing clusters
--\> randomly select 3 high schools and include all students from those schools

non-probability sampling

- NOT every member of the population is known and their selection into your sample is NOT through a random process

- convenience sampling -

- using a sample of people are easy to contact and are readily available to participate
- easy access, not resource intensive
- most common technique in psych. research
- EX: sona \= sampling college students in psychology courses

- quota sampling -

- similar to stratified random sampling
- fulfilling a quota for each specified subset using non-probability methods
- EX: college sample \= first years, second years, etc..
--\> smokers VS non smokers

self-selection (sampling biases)

- do those who volunteer to participate DIFFER SYSTEMATICALLY from the target population?

Experimental Design Characteristics

- manipulation
- measurement
- comparison

manipulation

- independent variable (IV)
- variable w/ categories within which participants grouped
- manipulate IV to create two or more conditions

measurement

- dependent variable (DV)
- outcome of interest that is measured
- measure all conditions (all levels of IV)
- EX: measure anxiety symptoms

comparison

- compare average (the mean) scores on DV for each condition
- happens in the analysis step
- "which group has a higher level of anxiety?"

- comparison conditions -

- control group
- comparison group

\-- comparison group --

- alternate intervention (comparing something to something else)
- can have multiple comparison conditions

\-- control group --

- not giving the treatment

business as usual (control group)

- one gets the condition and one doesn't

placebo (control group)

- take a fake version of a medication

waitlist (control group)

- one group receives treatment, one group gets the treatment when the study concludes

control

- all extraneous variables are control methodologically
- (extraneous variables \= all variables other than IV & DV)
- anything that can influence your outcome
- EX: desire to go to class, mental/physical health

- holding constant - (control)

- removing an extraneous variable by ensuring it applies to every participant
- we make sure everyone gets the SAME experience
- EX: measure during semester

- randomization - (control)

- random assignment: assigning participants to condition randomly.
- EX: heads -\> condition A ;
tails -\> condition B
- each participant has an equal chance of being in each condition

Experimental Designs

- independent groups design
- within groups designs
- factorial designs

independent groups design

- separate groups of participants are placed into different levels of the independent variable
- "between subjects" or "between groups" design

- BENEFITS:
--\> confirm that random assignment worked!
--\> track CHANGE in performance over time
- COSTS:
--\> different participants need to be recruited for each condition which can be difficult & expensive

- posttest-only design -

- participants are randomly assigned to the IV and are tested on the DV once
- EX: randomly assigning students to take notes on their laptops or on paper, and BOTH are taking a comprehension test
- exposing students to the IV and collect the DV once

- pretest/posttest design -

- participants are randomly assigned to the IV and are tested on the DV twice - BEFORE & AFTER exposure to IV
- EX: testing students' GRE scores, then putting them thru a certain class, and then testing their GRE scores again

within groups design

- ALL participants experience ALL levels of the IV
- BENEFITS:
--\> participants in your groups are equivalent because they are the same participants and serve as their own controls
--\> within-groups designs require fewer participants than other designs
- COSTS:
--\> internal validity threats
--\> practice & carryover effects may affect the outcome

- repeated-measures design -

- participants are measured more than once, after being exposed to each level of the IV
- EX: first, testing choco. with confederate, then rate choco. then, taste choco. alone, then rate choco.
--\> you are getting all the conditions of the IV

- concurrent-measures design -

- participants are exposed to all levels of IV at roughly THE SAME TIME, and a single DV is measured (& shared)
- DV is often looking at attitudes or preference

factorial designs

- two or more independent variables (IVs)
- researchers will study each possible combination of the IVs
- EX: young drivers, old drivers, on phones, not on phones
- 2x2 design

- interaction effects -

- whether the effect of one IV depends on the level of another IV
- effect of Independent Variable A depending on level of Independent Variable B

- main effects -

- Overall effect of Independent Variable A (collapsing across B)
- Overall effect of Independent Variable B (collapsing across A)

* Experiments have HIGH internal validity and LOWER external validity *

HIGH INTERNAL VALIDITY
LOW EXTERNAL VALIDITY

Quasi-Experimental Research

- doesn't use random assignment
- BENEFITS:
--\> high external validity
--\> real world examples
--\> solution for ethical constraints
- COSTS:
--\> no randomization (lower internal validity)

nonequivalent group design

- when researching the difference between groups and manipulation/randomization are not possible or ethical

nonequivalent control group posttest-only design

- has at least ONE experimental group and ONE comparison group
- no random assignment to conditions
- EX: organ donations \= doctors want to measure those who use the opt in VS opt out method for being a donor

nonequivalent control group pretest/posttest design

- has at least one experimental group and one comparison group
- no random assignment to condition
- EX: looking at plastic surgery and self-esteem

nonequivalent control group interrupted time-series design

- you know there is some intervention that's gonna happen but not WHEN it's gonna happen
- measurement overtime with some intervention in between
- when researcher doesn't know when that intervention happens and has no control over it