Midterm

Question 1

Empiricism

Answer 1

Your knowledge of the world is based on your sensory experiences. That can be studied through systematic observation.

Question 2

Rationalism

Answer 2

the view that regards reason as the chief source and test of knowledge

Question 3

falsifiability

Answer 3

the inherent possibility that a scientific theory or hypothesis can be proven wrong through observation or experimentation.

Question 4

The main assumptions of the physics model and how it shaped modern psychological science.

Answer 4

Sometimes known as positivism, 8 defining principles. – overall – the world exists independent of my senses and it will exist when I am not here - a REALIST ONTOLOGY.

Question 5

Major psychological tensions

Answer 5

Free will (we are governed by forces out of our control), Relativism (people basic psychological processes vary over time and space),

Question 6

A scientific attitude is one where…

Answer 6

…research is carried out systematically, skeptically and ethically.

Question 7

Rationalism

Answer 7

knowledge through logic and reasoning

Question 8

Social Constructionism

Answer 8

Individual focused, that social properties are created through interactions between people rather than having a separate existence. Qualitative research.

Question 9

Indigenous Psychology

Answer 9

a cultural challenge, values the individual, psychology by, from and for the local culture, e.g Kaupapa Maori research, Te Whara Tapa Wha.

Question 10

Research integrity should reflect…

Answer 10

…honesty (in all aspects of research), transparency (open communication and disclosure), accountability (individual and collective responsibility), care and respect (for everyone and everything involved), rigor (appropriate methods and standards)

Question 11

Researcher degrees of freedom?

Answer 11

Collecting and analyzing data until it is statistically significant.

Question 12

False positive psychology

Answer 12

Undisclosed flexibility in data collection and analysis allows presenting anything as statistically significant.

Question 13

Why is research integrity so important in psychological research?

Answer 13

You are more likely to find a false positive (type 1) than a false negative (type 2)
Undermines research trust, also rarely gets published and deters researchers away from conducting or attempting research, waste resources.

Question 14

Mataraunga Maori, Tikanga

Answer 14

The unique way of viewing the world- encompassing both traditional knowledge and culture

Question 15

Mataraunga Maori concepts

Answer 15

Collective possession, collective benefits, treasure, intergenerational connections, guardianship, spiritual, tikanga, verbal assurances,
- Mana of the Matauranga – tradition of oral culture – not picked up and repeated by pen and paper.

Question 16

Te Whare Tapa Wha

Answer 16

The holistic approach to health and health care for Maori, psychological, mental/emotional, physical, spiritual.

Question 17

Where should the end of your ethical research lead?

Answer 17

Thinking about the use to Maori

Question 18

Psychometrics

Answer 18

The science of developing and evaluating reliable and valid measures of unobservable variables

Question 19

Classical Test Theory (CTT) purpose

Answer 19

Proving a theoretical foundation for reliability and validity.

Question 20

Classical Test Theory (CTT) formula

Answer 20

X (observed score)= T(true score, the ideal error free reflection of what we want to find)+E (error, the combination of influences that distort our measures)

Question 21

Random error

Answer 21

Fluctuations due to chance – threatens reliability.

Question 22

Systematic Error

Answer 22

Consistent bias in one direction – threatens validity.

Question 23

Reliability types

Answer 23

Test, re-test, inter-rater, internal reliability

Question 24

Test-retest

Answer 24

The consistency of a measure over time, administering the same test to same individuals at different times would produce a high correlation in results – CTT logic – scores should remain consistent over time.

Question 25

Inter-rater

Answer 25

The consistency of measurements when different observers or raters are involved – CTT logic – different raters should reach similar conclusions if observing the same true score.

Question 26

Internal reliability

Answer 26

The consistency of items in a test. Does your measure actually measure what you want it to (reliability coefficients, Cronbach's a)

Question 27

Convergent Validity

Answer 27

Whether measures of the same construct correlate as expected.

Question 28

Discriminant validity

Answer 28

Whether measures are distinct and not related to other things -we want to make sure we are measuring what we indeed want to measure.

Question 29

Latent Variables

Answer 29

Constructs that are hidden or inferred therefore we can make construct validity claims.

Question 30

Basic Elements of path diagrams

Answer 30

Latent variable (unobservable variable, construct (T) > observed variable (scores on each indicator (aka items, X) > arrows show the direction of influence from cause to effect > e1, e2 (error terms – represent everything else affecting each indicator).

Question 31

Scales

Answer 31

Items share a common cause, designed to capture the intensity or degree of a single underlying concept by assigning different weights to items based on perceived importance, difficulty.

Question 32

Indexes

Answer 32

Items determine an outcome combine multiple indicators, measuring different aspects of a broader concept, assigning equal weights to each indicator.

Question 33

A higher alpha indicates...

Answer 33

...closer to 1 suggests greater reliability – meaning the items are measuring the same concepts consistently >.80 general cut off.

Question 34

Tau Variance

Answer 34

The variance of true effect sizes across studies.

Question 35

mcdonalds w, omega

Answer 35

a sophisticated alternative to alpha based on factor analysis principals, does not assume ‘tau variance’

Question 36

KR-20

Answer 36

a measure of internal consistency reliability on a test, specifically when the items are scored dichotomously (true/false)

Question 37

Probability Sampling

Answer 37

Every population member has known, equal chance of selection. - allows statistical inference to a known population – gold standard for external validity. Increased generalisability.

Question 38

Types of probability sampling

Answer 38

Simple random sampling – significantly large n – sample will be unbiased – will not perfectly replicate the population but no systematic error or bias. Stratified sampling – dividing pop into meaningful subgroups, and a random sample is taken from each group. - Chance of being selected depends partly on which stratum a person is in Quota – identify subgroups but put a quota on how many from each group you are using.

Question 39

Non-probability sampling

Answer 39

Same population members more likely to be selected, cannot statistically generalize to a known population, more common. Weaken external validity.

Question 40

Types of non-probability sampling

Answer 40

Same population members more likely to be selected, cannot statistically generalize to a known population, more common.

Question 41

The external validity of frequency claims...

Answer 41

...depends on how representative the sample is of the target population.

Question 42

When are non-probability samples acceptable.

Answer 42

When the claim does not require precise population estimates but rather tests generalizable relationships or causal mechanisms.

Question 43

Sample Bias

Answer 43

Systematic differences between your sample and the population you want your results to be relevant too.

Question 44

Why does sample size not determine representativeness?

Answer 44

Because it is the way you select participants, the sample size only determines the precision of your estimates. e.g 1000 randomly selected is more generalizable than a 10,000-convenience sample.

Question 45

MoE Margin of Error

Answer 45

A statistical estimate of the expected difference between a sample result and the true pop value. Better with probability samples.

Question 46

MoE and external validity

Answer 46

How generalizable the findings are due to the population studied.

Question 47

Cohen's D

Answer 47

The degree to which the phenomenon is present in the population.

Question 48

Morling

Answer 48

The strength of the relationship between two or more variables – mean differences scores would be the relationship between the two means.

Question 49

Cummings and Calin-Jageman

Answer 49

The amount of anything that’s of research interest.

Question 50

What can effect sizes be understood as?

Answer 50

Point estimates, in that we are talking about a statistic of some type – all of our statistics are making estimates about an effect size – a single point from our data.

Question 51

Raw Effect sizes

Answer 51

Pulling from the specific data units to establish effect size, - Measurement score, Mean difference, Variance, Under standardized regression coefficient e.g how much y increases for x.

Question 52

Pros for raw effect sizes

Answer 52

Directly interpretable, good for well-known scales and measures, common for clinical decisions and policy recommendations.

Question 53

Standardised effect sizes

Answer 53

Standardised across research using specific statistical tests to report them. Scale independent, common in research for testable theories, meta analysis, e.g z scores, correlation coefficients, cohen's d.

Question 54

Confidence intervals

Answer 54

The 'how precise' our estimate is - if we were to repeat the sample many times, about 95% of the resulting confidence intervals are going to contain the true population parameter.

Question 55

How large is the MoE?

Answer 55

About two standard errors wide.

Question 56

What are some assumptions of confidence intervals?

Answer 56

We don't expect data to be much further than our confidence interval. The guess of the CI is that the population parameter is closer to the point, the ends are the most extreme value, values closer to the point estimate are more plausible estimates of the true population mean. The population parameter may not always be in that confidence interval.

Question 57

What does the MoE tell us?

Answer 57

The maximum plausible distance between our estimate and the population parameter.

Question 58

Replication Predictions

Answer 58

There is an 83% chance that someone replicating the study would get a point estimate somewhere in our CI.

Question 59

Typical misinterpretations of CI's.

Answer 59

The confidence interval contains 95% of the sample scores – because sample scores fall within two standard deviations of the mean, 95% probability that population parameter is in the CI, 95% sure the effect is real, We can expect 95% of future sample means to fall in the CI.

Question 60

What would a larger sample size indicate for our standard error and CI?

Answer 60

Smaller standard error, narrower confidence interval.

Question 61

How can we alter variability?

Answer 61

Reducing this will reduce our CI (take out the outliers, target a smaller sampling frame. BUT, means you're sacrificing external validity, reliable measurement - but we can capture the t better.

Question 62

How is the width of the CI governed?

Answer 62

Two MoE

Question 63

MoE equation

Answer 63

MoE= SD (variability of data) x square root of n (sample size) divided by 1 x 2 - 95% critical value (corresponding to the desired confidence level). This can therefore make the confidence interval wider or narrower.