PSYC 306: research methods Ch. 1-4

What is epistomology?

Means acquiring knowledge

what nonscientific method of acquiring knowledge is linked to the availability bias?

Method of tenacity

what nonscientific method do conspiracy theorists rely on?

method of intuition

what two nonscientific methods are incorporated into the scientific method?

• empiricism and rationalism

The scientific method uses both reasoning and direct observation in a structured manner

3 factors of the scientific method:

• empirical

• public

• objective

examples of pseudoscience vs. non-pseudoscience:

Pseudoscience: aromatherapy, astrology, Myers-Briggs (MBTI)

• pseudoscience because they fail the falsifiability test

Science: Big 5 trait based model (falsifiable)

method of tenacity

information is accepted as true because it has always been believed or due to superstition supporting it

ex. opposites attract

ex. you can’t teach an old dog new tricks

ex. soap advertisement - saying that you’ll never be fully clean without Dove Soap!

limitations of the method of tenacy

information acquired may not be accurate

beliefs that are widely accepted can be very difficult to change

method of intuition

based on hunches or gut feelings that something is right or not right

• often the quickest method to get answers

Limitation:

no method to separate accurate from inaccurate knowledge

Ex. Psychics rely on this nonscientific method

what is a submethod of the method of authority that relies on accepting any word coming from a trusted adult, often accepting info without doubt or challenge (usually kids to their parents, or believers to their priest)

Method of FAITH —> submethod of the method of authority

what is rationalism

a method of acquiring knowledge through seeking answers from logical reasoning

premise statements vs. the arguement

premise statements = statements presumed to be true

an arguement = a set of premise statements that are logically combined to yield a conclusion

what is logic?

a way of establishing truth when there is no evidence of another truth available

limitations of rationalism

the conclusion is only ever true if both of the premise statements are true

limitation of empiricism

• not everything we observe is truly right (not everything we see is right)

• ex. the vertical vs. horizontal line illusion (the vertical line looks longer than the horizontal, but both are the same size)

• ex. plate of brown sesame noodles or plate of worms? you like the first option better, only until you’re told they’re worms! (you can’t always trust your senses)

what is the scientific method?

formulating specific questions, then systematically finding answers (contains nonscientific aspects like empiricism and rationalism)

at what step of the scientific method does induction play a role?

At step 1) Observe the behaviour / phenomena

what is induction?

using a small set of observations to form a general statement that can apply to an even larger set of observations

when forming a hypothesis, what do you have to think about?

variables that will be part of the hypothesis

variables = characteristics /conditions that can change and can be different for every individual

ex. environment, weather, age, personality

is a hypothesis the final answer to your research question?

NO!! A hypothesis is only a tentative answer and explanation that can be later revised

at what step of the scientific method do we use deduction

Step 3) using the hypothesis to form a testable prediction

Predictions are synonymous to conclusions in a rational arguement

Deduction is part of rationalism (whereas induction is used at Step 1) observing phenomena —> empiricism)

how do you describe the scientific method (is it linear or circular)

circular —> the hypothesis (tentative answer) is always changing, being refined, growing, through new observations being made —> new hypothesis —> testing the predictions —>supporting/refuting the new hypothesis —> refute: then new hypothesis forms again…

3 factors to the scientific method

1) empirical —> uses systematic observations to create a hypothesis and testable predicitions

2) objective —> accounts for researcher bias

3) public —> studies made available to the public once published, allows for replication by other researchers to verify the validity and reliability of the study

why is pseudoscience not a science?

it is not supported by empirical evidence, and is not falsifiable - pseudoscience discounts any negativity toward their “theories”

ex. aromatherapy, astrology, Myers-Briggs Test

Pseudoscience is subjective > objective, pseudoscience sticks with their theory grounded in past science, and never actively tests it (unlike science)

Qualitative vs. quantitative research

Qualitative = making observations and summarizing and interpreting through a narrative report

Quantitative = measuring variables for individual participants, obtaining scores, and then summarizing and interpreting through statistical analysis

different ways to define a research interest area (Ch.2)

By looking at:

1) Personal interests and curiosities

• a particular population

• a particular behaviour

• A general topic

2) Casual observation

3) Reports of others’ observations from informal sources (news, magazines, academic materials)

4) Practical problems or questions (basic vs. applied research)

5) Behavioural theories

difference between basic and applied research

basic research = answers theoretical questions for the sake of new knowledge

applied research = answers practical questions or solves practical problems

what are mathematical models (ex. delay discounting models)

a mathematical representation through an equation or more that defines a theory

• Ex. Newton’s F= ma, models the definition of force through its interaction with mass and acceleration

what are delay discounting models

a mathematical model to represent how much people devalue a reward the longer they have to wait for it

what is a primary source?

a firsthand report written by the individual who actually conducted the research and made the observations

ex. empirical journal articles, dissertations, theses, conference presentations

Note: not everything published in a journal is a primary source (ex. introductory section of a journal article may cite past research)

secondary source

a secondhand report in which an author talks about someone else’s observations

ex. books/textbooks, review articles, meta analyses, introductory section of research reports that summarize past research

literature search ultimate goal

to find a set of published research reports in your area of research interest, and to find an unanswered question (a gap in knowledge) based on your past research findings, that you want to fill/answer

what analogy is used to describe literature

literature is a tree-like structure, where new research grows out of old research (like new branches growing out from the main trunk)

• the top of the tree = current research

• the middle of the tree = newer research

• the base of the tree = old historical research (foundation of the research area)

when you conduct your search and look into secondary sources, what two factors do you want to pay attention to?

1) subject words (the correct terms that describe the study’s variables and participant characteristics)

2) authors (look at the citation section of secondary sources, as you read through many articles, note down recurring authors = they are most likely the leading researchers in that research topic - may be beneficial to look more of their papers up to find an unanswered question of interest)

what are two types of databases used to find published primary/secondary sources

1) full-text databases

• Ex. PsycARTICLES

2) non full-text databases

• Ex. PsycINFO

what does pyscINFO contain?

millions of abstracts - has lots of coverage so more resources under your area of research interest

• a place to find primary and secondary sources through peer-reviewed abstracts

what does psycARTICLES contain?

limited availability of resources, all full-text original publications, not just the abstracts

• a database to find primary and secondary sources through peer-reviewed journals

what are other databases (besides psycINFO and psycARTICLES)

• ERIC = for education literature and resources

• Pubmed/MEDLINE - for medicine, nursing, healthcare, preclinical sciences

• Google scholar = the largest database that covers expansive research topics not just limited to one field

how to screen our a good vs. bad journal article relevant to ur research topic? 4 steps

1) look at the article’s title - is it relevant to your topic?

2) look at the article’s abstract - is it still relevant to your topic?

3) look at the article’s full text (if it’s still relevant)

4) finally, look at the references —> if you’ve made it this far and it’s still relevant, check out other listed authors - they may be of interest to you and your research topic

when do you stop looking for articles related to your research topic?

simply when you’ve found enough that support a potential unanswered question (no limit, just until you’re satisfied)

• that you’ve obtained enough articles to justify new research

You can have as little articles as one to support / justify new research (it must just have enough relevant citations)

what is the most direct path to creating a new research study? (including 2 possible sub-paths)

reviewing already published research articles

1) Looking at the research paper author’s “suggestions for future research” listed at the end of the paper to generate an unanswered question

2) Compare and contrast 2 research papers to generate a question that wasn’t answered by either

what is included in the intro of a research article (3)

1) Lit review

2) Hypothesis

3) Prediction

what is included in the methods section (2) of a research article

1) Participants

2) Procedure

What is included in the results vs. discussion section? (of a research paper)

Results —> 1) statistics

Discussion:

1) Results vs. hypothesis

2) Conclusions

3) Applications

References are cited at the end, separately

what 4 factors makes up a good hypothesis?

1) Logical (the conclusion = hypothesis, with assumed to be true premises)

2) Testable (every variable in the hypothesis can be defined and measured)

3) Refutable (the hypothesis can be falsified - it’s possible for the outcome to be different than the prediction)

4) Positive (the hypothesis is making a claim about something being TRUE (a positive claim) and not about something that is NOT TRUE (unstable, not positive claim)

examples of a true hypothesis (positive claim)

For high school student, there IS a relationship between intelligence and creativity

example of a not positive, unstable hypothesis:

For adults, there is no relationship between age and memory ability

what’s wrong with a hypothesis claiming something as not true?

Goes against science - science is about discovering new research, with the assumption that a relationship doesn’t exist until proven otherwise by research - so if you’re research is already stating a relationship doesn’t exist —> it’s NOT PROVING anything other than the obvious

• this is a non-testable hypothesis —> can’t prove that something doesn’t exist

what is a construct? (hypothetical construct)

hypothetical non-measurable variable (either an attribute or mechanism) that explains and predicts a behaviour (in a theory)

what is a theory?

a set of statements about a mechanism/s that underly a specific behaviour

Ex. Maslow’s Hierarchy of Needs (has a whole bunch of statements that predict human behaviour)

• A good theory generates predictions about behaviour

examples of hypothetical constructs?

• motivation, anxiety, self-esteem

what can be measured around a construct?

since a construct is immeasurable, you can measure the factors that influence constructs and the behaviour that is influenced by the construct

ex. external stimulus —(influences)—> construct —(influences)—> external behaviour

ex. rewards/reinforcers —(influences)—> motivation —(influences)—> school performance

the external stimulus and the external behaviour can be measured but not the construct itself

what is given to a construct so that it can be measured and defined?

an operational definition

= procedure for indirectly measured and defining a variable that can’t be observed or measured directly

limitations of operational definitions on constructs (3)

1. No one-to-one relationship between the variable being measured and the actual measurements produced by the operational definition (the indirect measurement may not be the best measure for the construct)

2. ODs may leave out important parts of a construct

3. ODs may include extra components not part of the construct measured

   1. Ex. a questionnaire that is supposed to measure depression but asks questions that also target other aspects like “willingness to disclose” and one’s “honesty/openness”

. best method to operationalize

consult prior research - see if a good established measure for your construct is already there

if two measures are put on a scatterplot and shows no consistent relationship between the two, is it valid and reliable?

NO, validity and reliability is proven with a +1.00 or -1.00 relationship

what is validity

the measurement is measuring what it claims to be measuring

• ex. height, being measured by measuring someone’s weight = NOT VALID

name the 6 types of validity

1. face validity

2. concurrent validity

3. predictive validity

4. construct validity

5. convergent and divergent validity

what is face validity

the measure looks like it’s measuring the variable its supposed to be measuring

• most superficial, subjective, least scientific judgement of measurement validity

Limitation: the hypothesis may be obvious to participants if there is too much face validity, and that face validity can be very subjective and hard to quantify

what is concurrent validity

that the measurement for a variable gives the same outcome scores as an older established measurement for the same variable

• establishes a sense of consistency between two different measurements

what is predictive validity

that the measurements of construct accurately predict the behaviour it’s meant to predict

Ex. High predictive validity: SRSS (screening scale for at-risk students of behavioural problems) accurately predicts teacher’s ratings of the student’s self control

Ex. Low predictive validity: a buzzfeed employment personality quiz fails to accurately predict if a person is succesful 10 years later (zero correlation between the buzzfeed scores and the person’s employment ratings)

what is construct validity?

the measurement you created for a construct (hypothetical variable) behaves in the same way as the construct itself

• Ex. survey scores on depression strongly correlate with the actual test used to define Depression (Beck’s Depression Inventory)

(This kind of looks like concurrent validity - but its the premise that the survey is accurately predicting depression, the construct, in the way it should be measured and not like saying —- the survey scores strongly correlate with the Big 5 personality trait inventory)

• Ex. Do high scorers on an optimism test actually act upbeat

what is convergent validity?

do two new measures scores correlate strongly with one another? do they converge to produce the same score for the same construct

what is divergent validity? (making sure you construct measure is different from other construct measures)

does your construct measure show little to zero correlation with another construct’s measure? to prove that your measure isn’t assessing other constructs unrelated to your construct of interest

what is reliability? (of a measure)

that your measurement produces consistent results over time (if you measure your construct again with the same procedure, it’ll produce similar results)

Measured score accounts for what 2 things

Measured score = True score + Error

how to resolve measured scores with high error rates?

take the average score

3 common sources of error

1) observer error (by the researcher’s judgements, when there’s subjectivity involved)

2) environmental changes (time of day, temp, weather)

3) participant demeanour changes (their hunger, mood, attention)

what are the 4 ways to test reliability of a measure?

1) Test-retest reliability

2) Parallel forms reliability

3) Inter-rater reliability

4) Split-half reliability

explain the 4 types of reliability tests

1) test-retest

• Using the same measure on the same person, just testing them at different times - should yield the same results = proves the test is reliable

2) parallel forms

• Giving two versions of the same test to someone - they should get similar scores on both tests because it’s essentially the same test

3) interrater

• to account for one researcher’s subjectivity while measuring, we make sure another researcher can measure the participant again with the same procedure —> noticing the correlation between scores obtained by the two different researchers

4) split half

• splitting a test in half, and testing the same person both halves separately —> should yield roughly the same score because it is the same test and each component of the test should be measuring the same thing

• Calculating the correlation between the two test scores (one score for one half, one score for the other half) - should be highly correlated

Is validity a pre-requisite for reliability?

NO - reliability is a prerequisite for validity

but something can be reliable and NOT valid

what is measurement accuracy

the degree to which the measurement conforms to the established average

• accuracy is not used in behavioural research as much bc/ hard to test for ex. how accurate is a person’s depression

what do nominal, ordinal, interval, and ratio determine?

scales of measurement

What two scales of measurement are categorical and qualitative?

Nominal and ordinal

Difference between nominal and ordinal scales of measurement

Nominal = qualitative categories that don’t tell you direction or magnitude, but instead, only tells you if two things are different

• ex. lawyer vs. doctor

• ex. blue brown green hazel

Ordinal = qualitative categories that tell you direction but NOT magnitude

• ex. 1st, 2nd, or 3rd place

• ex. A large coffee is bigger than a small coffee but we don’t know by how much

• ex. very dissatisfied, dissatisfied, neutral, satisfied, very satisfied

• ex. big vs. small (you know that big is greater than small - a sense of direction)

Difference between interval and ratio scales of measurement

Interval = quantitative, with both direction and magnitude, ordered with equal spacing, has an arbitrary zero point (zero is just a reference, and doesn’t mean “absence of”)

• Ex. Celcius/farenheight —> Above or below 20 degrees

• Ex. Above or below sea level

Ratio = quantitative, with both direction and magnitude, ordered with equal spacing, as an absolute zero point (zero means absence of)

• Ex. she is 60 pounds, above zero pounds, a person can’t be negative pounds

Note: Quantities can be described in either interval or ratio, using the same terms:

• Ex. A child is 60 pounds which is -5 pounds relative to the average child weight of 65 pounds = interval scaling

• Ex. A child is 60 pounds = +60 more than being zero pounds = ratio scaling

Interval and ratio scales are the only scales of measurement that can be used to determine means and variances

can ordinal or nominal data be used to compute means and variances?

NO - only interval or ratio data be computes as means and variances

If you don’t know the difference between two sets of measurements being scaled (ex. IQ of 84 vs. 85) then it becomes what kind of data

ordinal

If you DO know the difference between the scaling, then it becomes interval data

what are the 3 modalities of measurement

self report measure

behavioural measures

physiological measures

limitation of physiological measures

the equiptment is expensive, and also may not fully / validly capture the construct

• ex. does measuring heart rate validly measure anxiety/fear

what are behavioural measures?

measuring behaviour through observations, tasks, activities directly

• the most flexible measure modality (out of self reports, physiological, behavioural measures)

when using multiple measures to define a construct, what is one way to limit confusion and complex statistics?

combine the multiple measures into a single score

what are the types of sensitivity problems common when detecting small change between subject measurements

range effects (where a cluster of scores are really high or really low, and its hard to locate individual differences) — including:

• Ceiling effect = cluster of scores really high (ex. if the test is too easy)

• Floor effect = cluster of scores really low (ex. if the test is too hard)

These effects can also occur if the test is good for one group (ex. adolescents) but not suitable for another group (toddlers)

what is an artifact, and what are the two common types of artifacts in a research study?

a nonnatural feature that’s accidentally introduced into a study and can distort the results

1) experimenter bias

2) participant reactivity

example of experimenter bias

an experimenter being too friendly and encouraging to one group of students about to do a performance task

and then being unfriendly to the next group of students

in order to validate their hypothesis

what is the solution to experimenter bias

standardize and automate the study and introduce either single or double blind experiments

single blind: only experimenter has no clue on the study’s expected results so its hard for them to influence the study’s outcome

double-blind: both experimenter and the researcher has no clue on the study’s expected results

what are demand characteristics (part of participant reactivity - type of artifact)

demand characteristics = any type of cues (about the environment, or what they may have heard abt the study before) that may signal to the participant what the hypothesis of the study is, and that could cause them to respond or behave differently

in this case, there are different types of participants that respond differently (4 types)

what 4 types of participant reactivities are there?

1) good subject role —> guessed the hypothesis right, and is now trying to “help” the researcher by making their answers more in line with the hypothesis

2) negativistic subject role —> guessed the hypothesis right, but is trying to do the opposite of the hypothesis to sabotage it

3) apprehensive subject role —> doesn’t know the hypothesis and is worried about how their answers may appear (put them in a negative light) so they try to act desirable > truthful

4) faithful subject role —> they follow instructions properly and they don’t care to know the hypothesis - these ppl are usually either followers of science, or simply apathetic participants = this is the ideal participant

what type of environment eliminates participant reactivity?

a field study - when participants are observed in the natural environment

2 basic parts to ethical responsibility

1) responsibility to the subjects/participants who participate in the study

2) responsibility to science to be accurate and honest in reporting research

what aspects of a study does research ethics shape?

1) which measurement techniques may be used

2) how participants are selected

3) which research strategies are designs are permissible

4) how the study is run

5) How data is analyzed

6) how results are reported

(basically all aspects of a research study is influenced by research ethics)

who oversees human ethics over animal ethics

IRB (institutional review board) » IACUC (Institute of animal care and use committee)

what did the national research act do? (ethics) 1974

Mandated formal regulations protecting research subjects

Note: the declaration of helsinki = international ethical guidelines for medical research (10 years before NRA)

what did the belmont report do

establish 3 ethical guiding principles

1) Respect for persons (informed consent)

2) Beneficience (no harm)

3) Justice

These 3 guidelines are synonymous with “No harm” and “Informed consent”

How many standards goes the APA ethical guideline have?

10 standards - covering the appropriate use and treatment of human participants in research

what is clinical equipose

this is the ethical responsibility for clinicians to provide the best possible treatment for their patients (part of no harm ethical guideline)

3 parts of informed consent

1) information

2) understanding (may require assent from participants, consent from guardians)

3) voluntary participation

passive vs. active deception

passive deception (omission) = withholding certain info from the participant

active deception (commission) = giving misleading information to the participant

what must follow a study after deception tactics are used

a full debriefing

what ethical guideline is violated by deception?

informed consent - the participants did not know they will be getting decieved

this is why a full debriefing is an important follow up after deception is used