Scientific Foundations Of Psychology Week Six Important Notes

gyat daymn man ;(

Correlation Research Designs

Non-experimental approach used to investigate relationships between two or more variables

• No manipulation

Relationship

Factor between two variables and is the manner extent of how their scores covary

Correlation Coefficient And Their Meanings

• .0 - .2 = very weak

• .2 - .4 = weak

• .4 - .6 = moderate

• .6 - .8 = strong

• .8 - .1.0 = very strong

Negative Correlation

When one factor increases and causes another factor to decrease

Positive Correlation

When one factor increases, causing another factor to increase

Interpretting Correlation Scientifically

• Any theory that cannot achieve a correlation of at least .9 isn’t deemed to be successful

• However, when potentially looking at elementary correlates of intelligence, any correlation over .3 is great

Limitations And Applications Of Correlational Designs

• Cannot establish cause and effect relationships

• Uncontrolled third variables can confound results, leading to mistaken conclusions

• Direction is causality remains uncertain and reverse causation is possible

Outlier

Observation that lies an abnormal distance from other values

Alternative Manipulative Experiments

• Non-Experimental

• Correlational

• Passive Observational

• Survey

• Observational

Cross-Sectional Desgisn

Involves two variales whch both may be scores or nominal catergories or a mix of both

• Would try and utilise three variables as the third may introduce a control to confoundign variables

• more demanidng n context to number of participants

Naturalistic Research Design

Include less artificial aspects to allow for higher generalizability

Natural Variation

By removing the nuisance of sources of variation, the key relationship is the strongest

• doing experiment with no manipulation highlights natural variation

Cases That Require Non-Manipulation Studies

• Naturalistic Research

• When Manipulation Isnt Possible

• Establishing An Association

• Natural Variation

• Comparing Size of Associations

• Prediction and Selection

• Explanatory Models

• Structure

• Developing/Refining Measures

• Temporal Change

• Temporal Direct Of Associations

Internal Consistency Of A Psychological Scale

May be measures as the split-hald reliability but often measured as the Bronbach’s Alpha

Third Variable Issue

• Hard to find casual relationships within non-experimental design

• Both variables are, to a degree, correlated with a third variable; may potentially bring out the original correlation

• third variable may increase/decrease correlations

Size Correlation Between Two Variables Is Reduced When

• range or variation of scores on one variable is restricted

• When the satter of the scores of two variables about the correlation line is fairly constant over the entire length of that line

Correlation Line

Straight line which we could draw through the points of a scatter gram

• If scatter scores around line aren’t consistent, it isnt possibel to determine the size of correlation

Standard Deviation

Measure of variance

• Certain samples may lead to no results due to limitations/restrictions but doesnt ean there is no relationship

Challenges Within Cross-Cultural Conduction

• Trabslation Difficulties

• Moderacy and Extremelity Bias

• Aceueesecens bias

• Reference Group Effect

• Deprivaion Effect

Challenges Within Cross-Cultural Conduction - Translation Difficulties

Accurate translation in surveys and questionnares critical

• goal is to have at least two bilingual translators to ‘back-translate’

Challenges Within Cross-Cultural Conduction - moderavy And Extremity Bias

Similar people may rate themselves quite differently on surveys. and questionnares

• Standardizing responses allows for comparison of respond patterns within cultures

Challenges Within Cross-Cultural Conduction - Acquiescence Bias

Differences in the extent to which people generally tend to agree/disagree with a statement on a survey/questionnare

Challenges Within Cross-Cultural Conduction - Reference Group Effect

When people compare themselves in similar research

Challenges Within Cross-Cultural Conduction Deprivation Effect

Surveys measure what people want to have rather than what they actually have

Ethnocentrism

the assumption that the ways of one’s own culture are the best ways of doing things.

Two-Eyed Seeing (Eruaptmunk)

Approach for reconciling Western research methods and theories with indigenous knowledges

• proposes to promote better understandings and approaching with strengths from both indigenous and western knowledge systems

Four R’s for Both Ways Knowledge

• Respect

• Relevance

• Reciprocity

• Responsibility

Benefits Of Experiments When Doing Cross Cultural Work

• Common alternative to cross-cultural compaison of subjective scales

• IV manipulated

• DV Measured

• culture can’t be assigned

• Compares means to two groups within each culture

• responses biases and reference groups held constant

Cultural Priming

Makes certain ideas more accessible to participants

• Different cultural scripters and available to people

Culture-Level Measures

Rather than directly investigating people, sometimes researchers look at cultural products around them

Correlation Research

Description of relationships between variables

• Not cause-and-effect relationships

• used in exploratory research to provide ideas of levels of association to develop hypothesis that can be tested by experimental research

Experimental Research

• Uses a homgenous sample of participants who are randomly allocated to groups of treatments including a control treatment, and tested on outcome measure

• manipualtion used to study effects on outcomes, while keeping other factors constant

Relationship

Manner and extend of how scores covary

• corrleation coefficient provides a single numerical estimate of the strength of direction of the relationhsip betwen teo variables

Factors Effecting Correlation

• non-linearity

• sampling from a restricted range

• heterogenous subsamples

• outliers

Simple Sampling

Measure everyone who comes in, can be anyone

• completely unbiased

• everyone has equal chance

• time consuming

Systematic Sampling

Every amount of person after a certain amount of people is chosen for the sampe

• its completely random and its not just anyone, but is every 20th or 30th person

• nothing is directly chosen

Stratified Sampling

Choose specific samples in order to get a better perspective on the specific subgroups