Psy 3001W Exam 3 UMN

Complex Design

Two or more IVs manipulated simultaneously in a single experiment

Inferential Statistics

Used to determine whether the results match what would happen if we were to conduct the experiment again and again with multiple samples

Null Hypothesis

The population means are equal - the observed difference is due to random error

(IV had no effect)

Research Hypothesis

The population means are not equal

(IV had an effect)

Statistical Significance

A significant result is one that has a very low probability of occurring by chance if the assumptions hold

Indicates that there is a low probability that the difference between the obtained sample means was due to random error

Probability

The likelihood of the occurrence of some event or outcome

Alpha Level

The probability required for significance

Sampling Distribution

Probability of obtaining possible outcomes (usually a table)

t test

Used to examine whether two groups are significantly different from each other

t = group difference/within-group variability

s

standard deviation

s^2

variance

Degrees of Freedom (df)

Used to determine critical value

n1 + n2 - 2

F test / Analysis of variance

Used to ask whether there is a difference among three or more groups, or to evaluate the results of factorial designs (2+ IVs)

Systematic Variance (Between Group Variance)

The deviation of the group means from the grand mean (mean score of all individuals in all groups)

Error Variance (Within Group Variance)

The deviation of the individual scores in each group from their respective group means

Confidence Interval

Most likely range of actual population values

Type I Error

Made when we reject the null hypothesis but the null hypothesis is actually true (alpha)

Type II Error

Made when we accept the null hypothesis although in the population the research hypothesis is true (beta)

Power

1 - beta (Type II Error)

Pearson r correlation

Used to describe the strength of the relationship between two variables when both variables have interval or ratio scale properties

Use Chi-square when:

IV is nominal and DV is nominal

Use t test when:

IV is nominal (2 groups) and DV is interval/ratio

Use one-way analysis of variance when:

IV is nominal (3 groups) and DV is interval/ratio

Use Pearson correlation when:

IV is interval/ratio and DV is interval/ratio

Use Analysis of variance (factorial design/F test) when:

IV is nominal (2+ variables) and DV is interval/ratio

Use Multiple regression when:

IV is interval/ratio (2+ variables) and DV is interval/ratio

p-value

Probability of obtaining that value of the statistic or a more extreme value of the statistic if the null hypothesis is true

p < .05 = significant

p > .05 = not significant

Effect Size

r = .15, small effect size

r = .30, medium effect size

r = .40, large effect size

Power Analysis

Given the effect size and level of significance, determine N needed to detect effect

Single-case experimental design (single-subject design)

Uses one subject. Seen in applied behavior analysis

Baseline

Control period where subject's behavior is measured

Reversal Design (ABA Design)

A (baseline period) -> B (treatment period) -> A (baseline period)

Multiple Baseline Design

Effectiveness of treatment is demonstrated when a behavior changes only after the manipulation is introduced. Multiple circumstances.

Across Subjects Multiple Baseline Design

Behavior of several subjects is measured over time. The manipulation was introduced at a different point in time.

Across Behaviors Multiple Baseline Design

Several different behaviors in a single subject are measured over time.

Across Situations Multiple Baseline Design

Same behavior is measured in different settings

Quasi-experimental design

Allows us to examine the impact of an independent variable on ta dependent variable, but causal inference is much more difficult because of the lack of random assignment and others.

One-Group Posttest-Only Design

Lacks control or comparison group

Participants -> IV (treatment group only) -> DV

One-Group Pretest-Posttest Design

Measures participants before and after the manipulation

Participants -> DV Pretest -> IV (treatment group only) -> DV Posttest

Threats to internal validity not controlled by quasi-experiments:

History

Maturation

Testing

Regression toward the mean

Subject attrition (mortality)

Selection

History

Events that occur during participation that affect behavior

Maturation

Changes due to the passage of time that affect behavior

Testing

Taking a test can affect subsequent testing

Instrumentation

Changes in measurement instruments (including observers) over time

Regression toward the mean

Extreme scores are likely to be followed by more moderate scores

Subject Attrition (mortality)

Participants selectively drop out of the experiment

Selection

When control and experimental groups are chosen in such a way that they are not equivalent

Nonequivalent Control Group Design

Participants in the control and experimental groups are not equivalent

(selection)

Participants -> IV (treatment and no treatment groups) -> DV

Nonequivalent Control Group Pretest-Posttest Design

Groups are not equivalent, but we can look at changes in scores

Participants -> DV Pretest -> IV (treatment and no treatment groups) ->DV Posttest

Propensity score matching

Matching individuals in the control and treatment groups based on propensity scores

Interrupted time series design

Examine a series of observations before and after a treatment to look for canges in behavior

Control series design

Interrupted time series design with a control group

Single Case Experimental Design

NOT a case study

Behavior recorded during baseline period, researcher manipulates IV, behavior is recorded again

Problems in all experiments

Contamination

Experimenter expectancy effect/observer bias

Reactivity

Contamination

Communication between participants

Reactivity

When participants behave differently because they know they are being studied

Developmental Research Design

used to study changes in behavior associated with age

not random assignment

usually a correlational study

Cross-Sectional Method

Randomly select participants from different age groups

Different participants make up each age group

Relatively fast

Cohort effects (cohort = group of people born at the same time

- Effects due to unique experiences of a particular generation rather than age itself

Longitudinal Method

Single group followed over time and measured repeatedly

takes a long time

cohort effects not an issue

conclusions may not apply to other generations

Mortality issues

multiple-observation effects

-Improved performance on test

-Trying to be consistent over time

-Reactivity

Interaction

When effects of one variable depend on the level of another variable

Factorial Designs

2 x 3

Number of numbers = number of IVs (2)

Value of numbers - levels of each IV (one has 2, the other has 3)

Product of numbers indicates total number of conditions (6)

Sequential Method

Mix between Cross-Sectional Method (first part of sequential method) and longitudinal method

Main Effect

The effect each variable has by itself

IV x PV Design

Factorial design with an experimental and non experimental variable (eg sex)

Simple Main Effect

Examines the mean differences at each level of the independent variable

Independent Groups (between subjects) Design

A different group of participants is assigned to each of the conditions

Repeated Measures (within subjects) Design

The same individuals participate in all conditions

Solomon four group design

Half of the participants are given the pretest, and the other half receive the posttest only

Literature Review

Researcher reads a number of studies about a topic and writes a paper summarizing and evaluating the literature

Meta-Analysis

Researcher combines actual results of studies and uses statistics to compare a given finding across multiple studies (focus on effect size)