research final

nominal

different groups NO # example- gender

ordinal

categories ranking

example- level of education

interval

more sophisticated statistical treatments; the intervals between levels are equal in size.

ratio

 have both equal intervals and an absolute zero point

bar graph

use separate and distinct bars for each piece of information. 

nominal and ordinal, looking at a single variable

histogram

uses bars that are side-by-side next to each other.

interval/ratio,  interpretations of the shape

Central tendency

a single number or value that describes the typical or central score among a set.

Central tendency nominal

Mode: the most frequent score

Central tendency ratio/interval

Mean: obtained by adding all the scores and dividing by the number of scores

variability

the amount of spread in a distribution of scores.

Standard deviation

the average deviation of scores from the mean

On average how much do scores vary?

• Symbolized as s and abbreviated as SD in scientific reports.

Variance

the square of the standard deviation

Inferential statistics

whether the results match if repeatedly conducted with multiple samples.

Descriptive statistics

used to describe the strength of the effect or relationship:

• Measures of effect size (Cohen’s d) are used to describe differences in a quantitative variable between groups or conditions

null hypothesis

the population means are equal and that the observed difference is due to random error.

Null Hypothesis Testing

• Based on the probability of the difference being related to sampling error

Retain the null hypothesis

(and accept that the results are due to sampling error and do not represent real differences in the population)

Reject the null hypothesis

(and accept that the results from the sample are representative of the population)

One-tailed tests

specifies a direction of difference between the groups.

Two-tailed tests

does not specify a predicted direction of difference.

Sample size

the total number of observations—has an impact on determinations of statistical significance.

• Greater size produces more confidence.

Type 1

when we reject the null hypothesis but the null hypothesis is actually true.

based on the choice of significance or alpha level.

Type 2

when the null hypothesis is accepted but in the population the research hypothesis is true.

Factors that affect Type 1 error

The lower the significance or alpha level, the lower the probability of an error.

• .05, .01., or .001

p-hacking (type 1 error)

an exploitation of data analysis in order to discover patterns which would be presented as statistically significant, when there is no underlying effect.

Pre-registration

the practice of publicly sharing a research and data analysis plan before starting a study

Factors that affect Type 2 error

affected by the studies power which is related to three factors:

• Significance (alpha) level;

• Sample size; and

• Effect size.

Quasi-experimental design name

an experiment where the researcher does not have full experimental control

Quasi means “resembling”

→ resembles a true experiment

True experimental design

• Uses random assignment to treatment group(s)

• The researcher designs the treatment

• Requires the use of control groups

Quasi-experimental design

NATURALLY OCCURS, NO CONTROL

• Uses non-random assignment (naturally occurring groups)

• Researcher does not have control over the design of the treatment

• Control groups are not required (but recommended)

One-group

a dependent variable is measured for one group of participants following a treatment (no comparison group)

Example- car accidents, and after watching 13 Reasons why

Non-equivalent groups

a researcher measures a dependent variable between existing groups where only one of the groups experiences the treatment

• Researchers choose groups as similar as possible to reduce confounds

• Most common type of quasi-experimental design

two groups one different than the other, no random assignment and manipulating exposure. (existing groups or one group that experiences a change)

Interrupted times series

multiple measurements taken at intervals over a period of time before and after treatment with one group

• Multiple assessments of the DV over time

Cross-sectional method

persons of different ages are measured at the same point in time.

Longitudinal method

the same group of people is observed at different points in time as they grow older.

• Most famous Harvard (1938)- young men from Harvard, and one from a different background

Factorial designs

compares more than one independent variable (or factor) on a single measured variable

• The simplest factorial design—a 2 × 2 factorial design— has two independent variables, each having two levels.

simples experimental design

are only two levels of the independent variable (IV).

Factorial Design format

# x # The number itself represents how many levels the IV has number of conditions

Each number represents an IV

Main effects

the effect that each IV has on the dependent variable independently

One to consider for each IV in the study

The example has TWO

• (1) Time of instruction

• (2) Setting

Interaction

situation in which the effect of one independent variable on the dependent variable changes depending on the level of another independent variable.

Interactions cannot be obtained in a simple experimental design

One for each set of IVs

• Ex: 2 x 2 has one interaction (A x B)

• Ex: 2 x 2 x 2 has three interactions (A x B, B x C, and A x C)

Types of interactions

• crossover X

• spreading

Single-case experiment designs

a type of quantitative, experimental research that involves studying in detail the behavior of a small number of participants

• Also called single-case design and small-N design.

Involves a detailed analysis of a single person, or small set of individuals, change in behavior or attitude based on a manipulation.

Single-case experiment designs used…

during a baseline period, followed by experimental manipulation and repeated measurement of the dependent variable.

Reversal design

a single-case design in which the treatment is introduced after a baseline period and then withdrawn during a second baseline period.

ABA reversal design:

Phase A: baseline is established for the DV (no treatment)

Phase B: researcher introduces treatment

Phase A: researcher removes treatment

Multiple baseline design

observing behavior before and after a manipulation under multiple circumstances.

Across (a) different individuals, (b) different behaviors, or (c) different settings.

Across participants (a)

Baseline is established for each participant

• Then, treatment is introduced at a different time for each participant

Across behaviors (b)

Baselines are established for the same participant(s)

Treatment is introduced at a different time for each behavior (aka each dependent variable)

Across settings (c )

• Imagine you are interested in how positive attention effects reading in different settings

IV: positive attention

DV: reading in different settings

When are multiple baseline designs preferred?

when you want to study the effects of an intervention on multiple behaviors, individuals, or settings without the need to withdraw treatment

 data analyzed for single-case experiments

Inferential statistics are typically not used

• Relies heavily on visual inspection of data

Involves:

• Plotting individual participants’ data

• Examine the plots carefully

• Make judgements about the IV’s effect

Factors considered in single-case experiments analysis

• Level: changes in level of the DV from condition to condition

• Trend: gradual increases or decreases in the DV across observations

• Latency: time it takes for the DV to begin changing after a change in condition

Effects of restriction of range with correlation

????

Null vs research hypothesis

????

Type of inferential test statistic for testing group differences on a quantitative variable

(hint: the one use for study #1)

T- test

designed to compare the means of two groups, allowing you to determine if there is a statistically significant difference between them

Type of inferential test statistic for two quantitative variables (hint: the one used for study

#2)

Pearson correlation coefficient (r) - which measures the strength and direction of the linear relationship

between the two variables

Can quasi-experiments support causal claims?

????

Advantage of factorial designs

??????

How to interpret main effects

????