Psych 202 Gallimore: Ch 2

Reliability of sources

Ask:

What am i being asked to believe?

Is the data being represented correctly?

Are they mistaking correlation for causation?

What determines:

-evidence

-replication

-

What DOES NOT DETERMINE:

-author/news outlet

-length

descriptive statistics

researchers can describe and summarize the mathematical results of their study

central tendancy

center batch of scores

mean, median, and mode

Variablility

can explain how similar or different a group of numbers are from eachother

Inferential statistics

use sample results to infer what is true about the broader population

statistical significance

a statistical statement of how likely it is that an obtained result occurred by chance

Null hypothessi

We start by assuming that there is NO relationship

Reject the null hypothesis

researcher concludes that the null hypothesis is wrong and that the independent variable had an effect on the dependent variable

large effect size/ strong correlation = more likely to be statistically significant

Statistically significant if p value is LESS THAN .05

significant result

low p-value

does NOT MEAN the hypothesis is true, rather the data is unlikely to occur under the null hypothesis

non-significant result

high p-value

does not mean a false hypothesis, rather it is likely that the data occurred under the null hypothesis

meta-analysis

a procedure for statistically combining the results of many different research studies

External validity

when a sample can generalize to the population

usually good when we use random sampling techniques

To cross populations we must replicate the study

internal validity

can we rule out alternative explanations

confound

when the experimental groups differ on more than just the independent variable

variable

something that has multiple states or values, while other aspects of a study remain constant

measured variable

observed and recorded in numeric form

some variables can only ever be measured

physiological measure

takes biological data from participants

manipulated variable

one who's levels the researcher controls by assigning different participants to different levels of that variable

self-report neasure

describing yourself in response to a survey or interview

observational measure

info gathered by directly watching a participant

operational definitions

specify exact process for determining the levels or values

random assignment

used in experiments to randomly assign different levels of the independent variable

increases internal validity

random sampling

used in correlational studies, experiments, observational study or a survey to select random participants from a population

increases external validity

informed consent

researcher must explain the procedures

ensure people are not coerced into participating

debrefing

the post-experimental explanation of a study, including its purpose and any deceptions, to its participants

Institutional Review Board (IRB)

determines whether the study upholds the communities ethical standards

decision making around 3 ethical principles:

- autonomy

- beneficence

- justice

Autonomy

-must give informed consent

-cannot be coerced though intimidation or high payouts

-protects vulnerable populations

beneficence

-evaluated on its risks and benefits to the participants and on the research's potential benefits to society

Justice

-people that bear the burden of the research should be representative of the people who will benefit from it

Animal welfare act

regulates how researchers must treat animals

Guiding principles of animal research

replacement: find alternatives when possible

refinement: minimize or eliminate animal distress

reduction: use the fewest amount of animals

IACUC (Institutional Animal Care and Use Committee)

committee that approves or denies permission for studies involving animals

Theory data cycle

theory, hypothesis, design, observe, compare

theory

set of propositions about what people do and why

hypothesis

prediction about what will happen based on the theory

data

observations from a study

replication

study has been conducted more than once

VERY IMPORTANT

journals

where researches share their scientific research in specialized scientific publications

Effect size

CORRELATION

magnitude of the relationship between manipulated and measured variables

- absolute value of r

—spread out dots: weak effect size

—medium

—close dots: strong effect size

r of .1 = weak effect size

r of .3 = medium effect size

r of .5 or greater = large effect size

Correlation coeffcient

quantification of direction and strength of correlation

d

statistical calculation of the effect size representing the difference between two means

the average response of the two groups differed by 'd' of a standard deviation

d = .20 ; small

d = .5 ; medium

d > .80 ; large

naturalistic observation

observing and recording behavior in naturally occurring situations without trying to manipulate and control the situation

observational research

measure a variable of interest by observing and recording what people are doing

case study

observational research method where one or two individuals are studied in depth

often have a unique condition

correlational research

'what kinds of people do this' 'what's associated with what'

measure 2+ variables in order to understand the relationship between them

data is presented in a scatter plot

negative correlation- inverse relationship

positive correlation- direct relationship

zero correlation- no relationship

corrlation

even a strong correlation does not mean we can say one thing CAUSES another

correlational studied can NEVER support casual claims because they can NEVER rule out a third variable

strong correlations lead to more accurate predictions

causation

One event leads to another

1. two variables must be correlated

2. must know which variable came first in time

3. no reasonable alternative explanations

experimental research

'why people do this' 'what causes these behaviors'

can support casual claims

independent variable

hypothesized cause

dependent varible

hypothesized effect

placebo condition

help researchers separate physiological effects of a treatment from a persons expectation that the treatment might effect them

descriptive research

'what people do'

measuring how people think feel or behave

cannot test relationships

researches one variable at a time

validity

appropriateness or accuracy of some claim or conclusion

construct validity

how well variables were operationalized to capture the variables of interest

Reliablility

degree to which a measure yields consistent results each time it is administered

descriptive statistics

graph or computation that describes characteristics of a batch of scores such as distribution, central tendency, or variability

frequency distribution

bar graph

scores on x

# of people on y

Variablility

the extent to which score differ from one another

standard deviation

how much a batch of scores varies around a mean

Why might we night be able to replicate a study?

differences in sample, materials procedures, or geography

false positives

statistically significant findings that do not reflect the real effect

small samples

larger samples provide more precise estimates because variability is lower

HARKing

hypothesizing after the results are known

-when ideas are generated after results of a study area known, it is important to replicate that finding in a second study

P-hacking

removing extreme scores, computing scores in different ways, or run different statistics

underreporting nonsignificant event

happens when researchers only report variables that showed strong effects and not mentioning others

open science

sharing ones data

preregistration

The practice of posting a study's method, hypotheses, or statistical analyses publicly, in advance of data collection

avoids HARKing

operational definition

a statement of the procedures used to define research variables