2820- test1

Gordon Allport's aim of psychology

enhancing above the levels achieved by common sense our powers of understanding, predicting behaviour

Introspection

having individuals report on their subjective impressions of various forms of stimulation

Positivist Movement

(Comte) to be scientific psychology was have publically verifiable data, objective observations, replicable results, no metaphysical statements

Lawfulness

behaviour can be understood as a predictable sequence of causes to effects

Determinism

behaviour is solely influenced by natural causes (not free will)

Facts

derived through empirical inquiry. events that are directly, objectively and repeatedly observed, publicly verifiable

Empiricism

all knowledge is acquired through the senses

Nativist viewpoint

knowledge is acquired through contemplation and reason

Naive empiricism/strict positivism

if something is not sensed it doesn't exist

Logical positivism/sophisticated empiricism

based on inference (conclusions are devised from observable evidence

Constructs

conclusions made from unobservable phenomena (uses operationalism)

Operationalism

unobservable is tied to something observable

S.S. Stevens

promoter of operationalism

Reification of a construct

treating the construct as a real entity (converting an adjective to a noun)

Falsifiability

Karl Popper. to be useful an explanation of behaviour must be testable

A good scientific theory makes

specific,testable and possibly false predictions

Occam's Razor

simpler is better

Lloyd Morgan's canon

avoid making more assumptions than necessary

Empirical Laws

assertions accepted as truths on the basis of empirical inquiry, laws do not need to state a cause-effect relation, just have to state that events are regularly associated

Inductive-statistical resasoning

specific --> general
bottom-up

Deductive reasoning

general --> specific
top-down

Hypothetico-deductive method

evaluating theories by generating testable hypothesis, research that gathers relevant data, reevaluating the theory

Refinement phase

devising an idea after reading existing literature

Population

larger group of people in which the researcher is interested

Sample

smaller group that the researcher actually measures

Statistics

info the the sample and used to make predictions about parameter

Parameter

values of the population which are of interest

Parameter

values of the population which are of interest

Single sample design

a single set of measurements is taken and analyzed

Two sample studies

studies involving two populations

between participants design

different people in the two groups

withing participants design

same people used in both groups

Independent variable

are administered by the researcher, participants receive different levels of the variable, can be used to identify a causal relation

Quasi-independent variable

groups formed on basis of characteristics participants already have (preexisting), can only be used to identify a correlation

Factorial Designs

studies that use groups formed on basis of more than a single factor

Main effect

something that effects a subject

Interaction

how main effects interact

Count data

each subject is counted as being in a certain category

Score data

individuals are measured to determine what value of variable they exhibited (like in questionnaires)

One demensional count data

categorizing the subjects on the basis of one dimension

Goodness of fit

how well does what I observe fit what I expect?

Two dimensional count data

categorizing subjects on basis of two dimensions, uses a chi square test of independence, are the two dimensions related?

In order for A to be a cause of B

A and B must be at least correlated, correlation doesn't guaruntee causation

Regression Analysis

using the existence of a correlation to make predictions

Variable

any characteristic that can be identified, measured and can assume different values

Ethnographic research

experience of a group of people

Recurring themes

used in analysis of qualitative research, characterizations that are common to several participants' description

Discrete variable

assumes a countable number of values

Dichotomous variable

a discrete variable that may assume only two values

Continuous variable

has an infinite number of values

Scales of measurement

nominal scale
ordinal scale
interval scale
ratio scale

Nominal scale

-observations are placed into mutually exclusive categories
-can't assign numerical operations
-count data

Ordinal scale

-can be placed in mutually exclusive categories
-intervals are not evenly spaced

Interval scale

-objects are classified into mutually exclusive categories that form a logical order
-no true zero points, but intervals are evenly spaced

Ratio scale

-objects are classified into mutually exclusive categories that form a logical order
-have a true zero point (exhists theoretically but not necessarily in reality)

Descriptive Statistics

-describing sets of data
-summarizes the data in such a way that it aids in the understanding of the data

Inferential statistics

making statements about populations based in samples

Mode

the most frequently occurring measurement value

Median

measurement value below which 50% of the measurements fall

Mean

average
-distribution is negatively skewed when the mean is less than median
-positively skewed when mean is greater than median

Range

difference between highest and lowest values

Measures of variability

variance and standard deviation
-the closer the numbers are to the mean, the smaller the variance (less spread)

z-score

tells you how far away from the mean that value is in terms of the number of standard deviations

a value equal to the mean will have a z-score of

0!

Preexisting bias

having an already formed but nonoperationalized idea about what a psychological concept means

Validity

-a measure actually measures what you want it to measure
-construct validity (criterion-related validity: concurrent/predictive & content validity) and Face validity

Construct validity

-degree to which the measurements reflects the construct of interest
-scores on the measure should allow accurate predictions based on construct that measure is supposedly measuring

Criterion-related validity

the measure correlates highly with some outcome criteria

Criterion-related validity- Concurrent validity

comparing diagnoses of previous test with new test

Criterion-related validity- Predictive validity

the criterion is a future behaviour

Content validity

the degree to which the questions cover the full range of behaviours

Face validity

a measure has the appearance of measuring what it is meant to measure

Reliability

-increased to the degree to which measurements are consistent and do not contain measurement error
-consistency, reliability of measurements
-a measure can be reliable but still be invalid

Interrater reliability

asses the degree to which two raters or judges give the same measure

Test-retest reliability

temporal stability
-same test (or parallel form) is administered on separate occasions

Split-half reliability

-measure of internal consistency
-it`s expected that participants should respond to the items of a single construct in a consistent fashion

p=2rxy/1+rxy

-used to calculate split-half reliability (rxy=correlation b/w the split halves)
-the number of items in each split-half is less than the total number items

N= Desired rxy (1-current rxy)/Current rxy (1-desired rxy)

-N is the factor you should increase the number of items by
-if reliability is too low this shows you how many items to add

α= 1/ [1/k (1/r ̅ - 1) + 1]

Cronbach's alpha
-determine's which is the better split half to use
-k=total #items
-r ̅ = mean inter item correlation

Standardizing

establishing norms for a test

Measurement reactivity

how measures may be affected by factors other than those intended

Nonreactive vs. reactive measures

Nonreactive measures aren't affected by outside factors and are therefore preferred over reactive measures

Ways researcher influences measures

-Biosocial/psychosocial characteristics
-experimenter bias
-Demand characteristics
-artifactual range effects

Biosocial characteristics of experimenter

factors such as sex, age, race
-participants may feel more comfortable if experimenter does not differ in these characteristics

Psychosocial characteristics of experimenter

warmth, status, interpersonal skills
-people tend to score higher when experimenter is warm rather than cold

Experimenter bias

experimenter has an idea of how results should turn out and effects the responces of participants

Blinding

the researcher doesn't know expected results

Single-blind studies

the participant can't guess the expected results

Double-blind procedure

neither participant or researcher know expected results

Expectancy control design

Half the participants get the treatment and half the placebo, then for half of each of these groups the researcher is told the participant got the treatment and other half got the placebo (2 x 2 factorial design)

Social Desirability

participants behave in ways which conform to social norms; may be unwilling to express socially dispproved opinions

Demand Characteristics

cues that communicate to the participant the purpose of the study and expected behaviour of participant, leads to cooperative attitude or negative attitude

Cooperative attitude

demand characteristic, try to perform in the manner they believe they should

Negative attitude

perform in a manner which they believe is opposite what they should

Artifactual range effects

restricting the range of values that measurements may take

Ceiling effect

an artifactual range effect, it obscures real differences that exist between participants

Floor effect

an artifactual range effect, participants' responses will be close to the minimal value of the dependent measure

Self report measures

interview and questionnaire

Categories of interviews

unstructured, informal
structured, formal

Structured, formal interview

uses a set of pre-selected questions whihc are presented in a pre-determined order

Closed-ended questions

a specific set of alternative answers is provided

Open-ended questions

no pre-selected alternatives, a free-flowing verbal response

Unstructured, informal interview

an interview with a purpose but allows for free verbal responses

Discourse analytic view

there is no single objective truth about behaviour

Semi-structured interview

some open-ended questions to get to the closed-ended questions

Constructivist view

knowledge is relative

Grounded theory

-researchers interview individuals about the experience of interest
-theory of the behaviour emerges from the data

Point of Saturation

used in grounded theory, the point at which there are no new emerging themes

Ethnography

-a type of qualitative research
-researcher lives within a group in order to gain an understanding of culture

Action research

-a type of qualitative research
-research on practical issues occuring in every life

Questionnaire

-a self report measure
-used to identify characteristics of people so they can be placed in groups
-initial development involves open-ended questions

Demographic information

info. such as gender, age, religious affiliation etc.

Psychometric properties

validity and reliability

Content analysis

identifying recurring themes

Good questionnaire writing

-wording (avoid biased wording, understandable language, negatives)
-avoid double barreled questions
-positively and negatively keyed items
-use of a scoring key

Double barreled question

at question that asks more than one thing

reverse score

used for negatively keyed items

lie scale

questions used to test of participant is concientious in answering

error choice method

used for sensitive contructs, often there are several opinion based answers but not of them are correct

bogus pipeline

using a fake lie detector

Types of rating scales

-Likert Scale
-Visucal Analogue Scale (VAS)
-Thurstone equal appearing interval scale
-Semantic differential scale

Likert scale

-involves stepwise adjective choices treated as equal intervals
-uses anchors

Visual Analogue Scale (VAS)

participants place a mark on a line
the distance from the anchor to the mark is measured

Thurstone equal appearing interval scale

-each item is rated by a panel as to the extent to which it is consistent with a viewpoint
-each item is then given a score

Semantic differential scale

-participant is given a set of bi-polar adjectives separated by lines

Why is it important to have a description of a novel behaviour?

you can figure out how often it occurs, who, average amount of the behaviour

Why is it important to have a description of a behaviour that's occured for a while?

you can figure out if the behaviour has changed

Why are descriptive studies in general important?

They can identify and describe important behaviour

post hoc ergo proctor hoc fallacy

misidentifying a correlational relationship as a causal one

regression analysis

using the exsistence of a correlational relationship as the basis of a prediction

Criterion variable

the variable that is being predicted

predictor variable

the variable from which the prediction is being made

Third variable problem

two variables covary because they are related to a third variable

spurious correlation

two variables are related only because they are both related to another variable

Directionality problem

existence of a correlation does not provide any information about the potential direction of the relationship

Selection bias

refers to the fact that individuals with certain characteristics may be more likely to choose different things

Covariation

correlation is a necessary condition for causation, but not a sufficient condition

Minimal experimental design

1. use of a true independent variable
2. Random assignment of participants of different levels of independent variable

Demonstration

involves only one group receiving one treatment and thus can't be used to establish causality

Extraneous variable

anything other than the independent variable being manipulated

Internal validity

the variation in the independent variable is the only thing that caused the variation in the dependent variable

Confounding variable

extraneous variable that covaries with the variable under study and thus could account for observed outcome

Threats to internal validity

-history
-biased participant selection
-testing
-instrumentation
-statistical regression to the mean
-experimental mortality/attrition

History

-a threat to internal validity
-events that take place between measurements

Biased participant selection

-a threat to internal validity
-if participants in one group differ initially from those in another group

Testing

-a threat to internal validity
-with multiple testing, subsequent test results may be influenced by previous test results

attitude polarization

if participants are asked about their opinion on a topic on several occasions they may reconsider their position after the first test

Instrumentation

-a threat to internal validity
-in studies with repeated measurements, there is no guarantee that subsequent measurements are comparable to initial measurements

Statistical regression to the mean

-a threat to internal validity
-Zy=rxyZx
-when participants are selected based on extreme scores (changed score on subsequent test is attributable to treatment or just regression, ie. extreme scores tend to move back towards the mean?)

Experimental mortality/attrition

-a threat to internal validity
-when participants to not complete the study
-only a problem if prob isn't random

External validity

the generalizability of results, population external validity and sitution external validity

Population external validity

generalizability of results to people

Situation external validity

generalizability of results to conditions

Threats to external validity

-biased participant selection
-participant mortality/attrition
-measurement reactivity

Biased participant selection

-threat to external validity
-sample is not representative of the population of interest and the results won't generalize

Participant mortality/attrition

-threat to external validity
-if there is nonrandom participant drop out you will be left with a biased sample that won't generalize

Measurement reactivity

-threat to external validity
-any procedure that could make the participant react differently then they normally would

Mundane realism

-aids in external validity
-making studies as similar as possible to the real world

Experimental realism

aids in external validity
-degree to which participants experience the psychological states in which researcher is interested and thus behave in a natural manner
-physical context does not need to be similar

Ecological validity

studies conducted in more naturalistic settings, the term is thought to be unnecessary and misleading

Best possible type of sample

unbiased and representative

Biased samples

-favour the inclusion/exclusion of participants with certain characteristics
-limits internal and external validity

If a sample is biased it cannot be ___________. However even if a sample is unbiased it may not be _________.

representative

Probability sampling

uses random selection, there must be a listing of the population
-simple random sampling, interval sampling, proportionate stratified sampling, cluster sampling

Sampling frame

a listing of the population

Simple random sampling

every member of population has an equal chance of being in the sample

Systematic selection plan/interval sampling

there is a list of population and a random starting point with a constant interval between selections

Strata

segments of the population

Proportionate stratified sampling

proportion of the total participants randomly selected from each trata is made equal to the proportion of the total population

Single cluster sampling

-naturally occuring groups of participants are the clusters
-a random sample of clusters is taken and every member from chosen clusters is sampled

Multistage cluster sampling

-clusters are randomly sampled
-individuals within chosen cluster are randomly sampled

Non-probability sampling

-necessary to assume that the persons selected are similar to the ones that aren't
-convenience sampling, snowball sampling, quota sampling, purposive sampling

Convenience sampling

units of the population that are easiest to measure are used
-limits generalizability

Voluntary response samples

units of population voluntarily respond

Snowball sampling/networking

used with hard-to-acquire participants
-the researcher knows someone with desired characteristic --> this person passes on info about study to similar persons

Quota sampling

participants are recruited until a quota for the sub-groups has been met

Purposive sampling

researcher identifies a specific set of individuals who can provide into needed

Sample size

-quality of a sample is not determined by its size but by how well it suits purposes of researcher
-sample size affects statistical significance, power

Institutional review board (IRB)

reviews ethics of research projects

Tuskagee Study of Untreated Syphilis in the Negro Male

-treatment for syphilis was withheld from african-american males in order to determine damage of untreated syphilis

Dr. Mengele

-carried out unethical medical experiments on subjects at Auschwitz

Nuremberg Code

early code of ethics

Essential features of APA ethical code

-participants must be treated as people
-participants must leave study in no worse condition than when they entered
-informed consent
-voluntary participation
-debriefing
-use of qualified personnel

Informed consent

-researcher must inform participants of all aspects of research that might influence their willingness to participate
-important that consent form is signed, concise

Confidentiality

information collected can be linked to participants

Anonymity

there is no link between info and participants

Reasons to break confidentiality

-if participant is at risk for self-harm or harming others
-suspected risk, incidence of child abuse
-information relating to police investigation

Confidentiality in regards to storage

-storage must be safe a secure
-researcher should say where and how long data will be stored
-must inform participants of plan to archive data
-if computer files aren't anonymous they must be protected

Confidentiality in regards to qualitative studies

-permission must be obtained --> ask if they're allowed to quote with source, quote without souce or just paraphrase

Hyperclaims

claims about the importance of the research that are not realistic

Deception

giving false information about the study

Volunteer bias

some characteristics are more common n people who volunteer for psychological research than those who don't

active treatment control

using an already established treatment as a control group when testing a new treatment