rm1 final

simple random assignment

participants are assigned to groups (e.g., control vs. treatment) using pure chance, such as flipping a coin, drawing names from a hat, or using a computer random number generator; every participant has an equal probability of being placed in any experimental condition

matched random assignment

a procedure for assigning participants to experimental conditions in which participants are first matched into homogenous blocks and then participants from each block are assigned randomly

between-subjects design

different participants are assigned to each condition, meaning each subject experiences only one level of the independent variable

within-subjects (repeated measures) design

the same participants experience all levels of an independent variable, serving as their own control

mixed factorial design

combines features of both between-subjects randomization and repeated measures in an experiment; requires multiple independent variables

internal validity

the degree to which a researcher draws accurate conclusions about the effects of the independent variable

external validity

The º to which results of one study can be replicated or generalized to other samples

confounding

a condition that exists in experimental research when something other than the independent variable differs systematically among the experimental conditions

biased assignment/selection effects

when participants are assigned to conditions nonrandomly

differential attrition

people who drop out of the study differ in systematic ways compared to those who stay

order effects

when a participant’s behavior is impacted due to the order in which they participate in the various conditions of the experiment

practice effects

when participants’ performance improve simply because they complete the dependent variable multiple times

fatigue effects

participants get tired/bored as the experiment progresses

sensitization

participants may figure out the hypothesis → act differently

counterbalancing

used to counteract order effects

placebo effect

a psychological or physiological change that occurs due to a suggestion 

placebo control group

used to counteract placebo effects

experimenter expectancy effect

experimenter expectations alter participant reactions

ways to counteract the experimenter expectancy effect

blinding researchers to condition and standardizing procedures

experiments need 3 things:

1) Manipulation: vary at least one independent variable and assess impacts on responses

2) Random Assignment: be able to assign participants to conditions in ways that guarantees initial equivalence (that groups do not vary systematically)

3) Control: control all extraneous variables that could influence participant responses

_____________ is a great way to achieve these three elements

Randomizing participants to groups

one-way design

an experimental design with one independent variable

factorial design

two or more independent variables

main effect

the effect of a particular independent variable, ignoring the others

A factorial design will have as many main effects as there are ________ variables

independent

Ex: a 2x3 design will have 2 main effects, 3x2x2 will have 3 main effects

An experiment with two independent variables, each of which has three levels, would be described as a ___ design.

3x3

A 3 × 2 × 2 factorial design has __ conditions

12

interaction

any time that the effect of one variable is modified by another

moderator

a variable that moderates or qualifies the effects of another var.

subject variable

not the IV but a potential moderator; e.g., age or sex

posttest-only design

responses are only measured once (post-intro to indie variable)

pretest-posttest designs

responses are measured twice (pre&post intro to indie var.)

null hypothesis testing

determining whether the size of an effect (i.e. differences between correlation size) is larger than what would be expected if the effect was due only to error variance → if so, the difference is accounted for by the IV

null hypothesis

states that the independent variable will NOT have an effect; Correlation will be .00 or that means of the various conditions will not differ

alternative hypothesis

states that the independent variable WILL have an effect; The means from various experimental conditions will differ from each other

type I error

rejecting the null hypothesis when it is true

type II error

failing to reject the null hypothesis when it is false

alpha

the maximum probability that a researcher is willing to make a Type I error; typically set at 0.05

beta

probability of committing a Type II error

• Many things can conspire to obscure the effects of the independent variable and thus lead researchers to make Type II errors → to reduce the likelihood of this, researchers try to design experiments that have high power

rejecting the null hypothesis

concluding that the independent variable had an effect

• Occurs if the difference between the means of the experimental groups is larger than would be expected given the error variance

failing to reject the null hypothesis

concluding that the independent variable had no effect

• The group means differed about as much as one would expect them to 

• We can never “accept” the null hypothesis because we can't conclude 1000% that the independent variable had zero effect whatsoever

statistical significance

a finding that is very unlikely to be due to error variance; <0.05

p-value

the probability that an obtained effect (i.e. correlation) is due to error variance

• If the p is low (<0.05), reject that ho!!!!!!!

confidence interval

a range of values, derived from sample data, that is likely to contain the value of an unknown population parameter (e.g., mean) with a specified level of confidence

95% confidence interval

the range of scores in a sample within which the means of other samples drawn from the same population are likely to fall 95% of the time

• if we repeated the study 100 times and calculated CIs on all of those samples, 95 of the resulting CIs would contain the true population value

statistical power

the probability that a study will correctly reject the null when it is false

• A study’s ability to detect any effect of the independent variable that occurs

• Studies w/ low power may fail to detect the independent variable’s true effect 

• Power is the opposite of beta (power = 1 - beta)

power analysis

a statistical that conveys the power of a study; often used to determine the number of participants needed to achieve a particular level of power

• Researchers usually aim for the level of power = 0.80, making beta = 0.20

• Most published work only has a power level of 0.5…. Awkwarddddd

effect size

the strength of the relationship between two variables, often portrayed as the proportion of variance in one variable that can be accounted for by the other

• factorial designs can have multiple effect sizes

cohen’s d

an indicator of effect size based on the size of the difference between two means relative to the size of the standard deviation of the scores; expresses the size of an effect through standard deviation units

eta squared

an indicator of effect size that expresses the proportion of variance in a continuous variable that can accounted for by a nominal or categorical variable

odd’s ratio

the ratio of the odds of an event occurring in one group to the odds of the events occurring in another group; used when the dependent var. only has 2 levels

ANOVA

used to test mean differences between 3+ groups

f-test

used to test differences between means in ANOVAs

one-way ANOVA

appropriate for a one-way experimental design (one IV that has three or more levels)

Ho: There is no difference in mean literacy scores across the 4 residential communities

two-way ANOVA

appropriate for a two-way (factorial design). 2 x 2 example:

• Treatment 1: Therapy

• Treatment 2: Medication

Ho: there is no effect of factor A, no effect of factor B, and no interaction between them

t-test

an inferential statistical that tests the difference between two means

rejection region

t-statistics that are extreme enough to fall within the rejection region are evidence to reject the null hypothesis of no group differences.

one-tailed t-test

Hypothesis: "Group A will score HIGHER than Group B" (direction

specified)

Rejection region is in ONE tail only (all α on one side).

Use when prior theory or evidence strongly predicts a specific direction

two-tailed t-test

Hypothesis: "There will be a difference" (direction unspecified)

Rejection region is split between BOTH tails (α/2 each side).

Use when you have no strong prediction about which direction the effect will go.

single sample t-test

When: Compare one group's mean to a known population value e.g., Are psychology students' average IQ scores different from the population norm of 100?

independent samples t-test

When: Compare means from two separate, unrelated groups | e.g., Do men and women differ in their levels of risk-taking behaviour?

dependent samples t-test

When: Compare means from the same group measured twice, or matched pairs | e.g.,

Do anxiety scores decrease after an 8-week mindfulness programme?

If t/F exceeds the CV, we ____ the null hypothesis

reject

The _____ t is, the less likely the difference between the means is due to error variance

larger

The larger the t, the _____ the p!

lower

interrupted time series design

O1  O2  O3  O4  X  O5  O6  O7  O8

• The measurements are interrupted by X, the independent variable

local history effect

a threat to internal validity where an extraneous event occurs to one group but not the other

nonequivalent control group design

the group that receives the quasi-independent variable is compared to one or more group(s) that didn’t receive the treatment

posttest only design

aka static group comparison - comparing a group that received the quasi-independent variable w/ 1 that didn’t

pretest-posttest design

both groups are measured before and after the quasi-independent variable X

Quasi-experimental group: O1 X O2

Nonequivalent control group: O1 -- O2

natural experiment

a study in which researchers do not have control over randomization and who is assigned to “treatment” or “control.”

• Nevertheless, there is a source of variability that that makes variability more or less random (i.e., that participants also cannot control)

  • Weather or natural disaster

cross-sectional design

comparing groups of different ages at a single point in time

longitudinal design

the quasi-independent variable is time itself

• O1  O2  O3  O4  O5

• Nothing has occurred between the O’s other than the passage of time

Used most frequently to study age-related changes in thoughts, feelings, and behaviors

ABA design

demonstrates that an independent variable affects behavior, first by showing that the variable causes a target behavior to occur, and then by showing that removal of the variable causes the behavior to cease

• Sort of like an interrupted time design performed on a single participant

case study

a detailed study of a single individual, group, or event

idiographic approach

research that describes, analyzes, and attempts to understand the behavior of individual participants

nomothetic approach

research that seeks to establish generalizable principles

intraparticipant replication

replicating the effects of the independent variable with a single participant

single-case experimental design

 unit of analysis is the individual subject, not a group

• Avg. aren’t used → can’t use statistics like t-tests, F-tests, or confidence intervals

respect for persons

Individuals should be treated as autonomous agents; those with diminished autonomy deserve protection.

In Practice:

• Informed consent process

• Right to withdraw at any time

• Special protections for vulnerable population

benefisence

Do not harm; maximize possible benefits and minimize possible harms to participants.

In Practice:

• Risk-benefit analysis before approval

• Minimizing distress in study design

• Protecting participants from unnecessary harm

justice

The benefits and burdens of research should be distributed fairly across society.

In Practice:

• Equitable selection of participants

• Avoiding exploitation of vulnerable groups

• Communities that bear risks should receive benefits

coercion

when participants agree to participate in a research study because of real or implied pressure from some individual who has authority or influence over them.

minimal risk

a risk of harm or discomfort that is no greater in probability and magnitude than the risks that people ordinarily encounter in daily life or during the performance of routine physical or psychological tests.

confidentiality

maintaining the privacy of participant responses

deception

misleading or lying to participants for research purposes, such as…

• using an experimental confederate 

• providing false feedback to participants

• presenting two related studies as unrelated

• giving incorrect information regarding stimulus materials

debreifing

participants are told about the nature of the study after it is completed

informed consent

Inform participants of the nature of their participation in the study and obtain their explicit agreement to participate

IRB

entity that judges the ethics of research studies