Research Methods pt 2

FINAL

between subjects

compare across groups

within subjects

differences within a subject, within a group

mixed design

between and within

block design

using an extraneous variable as an independent to control or measure its effects

\
Ex: Everyone receives same treatment, researcher looks at severity as different groups

"True" Experimental Design

Control group

Randomized to groups

(Needs BOTH)

Quasi-Experimental Design

An experiment that does not require random assignment to conditions.

(lacks random assignment and/or comparison/control group)

One group, pre-post test

Within subjects

No randomization

\
Ex: Kids with Dyslexia receive reading pre-test, complete 6 weeks of READRIGHT treatment, then receive same reading post-test immediately following treatment

One-way, repeated measures over time

Within subjects

No randomization

Must have three or more group means (i.e., must test participants 3+ occasions)

\
Ex: Test people with aphasia on the Boston Naming Test prior to treatment \[Time 1\], immediately post-treatment \[Time 2\], and three months post-treatment \[Time 3\].

Non-equivalent, control group (pre-post test)

Between subjects

No randomization

1+ treatment/experimental group AND control groups (s)

Used with populations that are intrinsically not comparable

\
Ex: Test auditory attention before and after 3 months on gluten-free diet in kids with ADHD, kids with auditory processing disorder, and typical kids (control group, matched for age)

Non-equivalent, post test

Between subjects

No randomization

With or without control group

Often when pretest is not possible

\
Ex: Anxiety scores among people with hypertension in affluent communication vs working class communities post COVID.

Factorial Designs

Two or more independent variables are manipulated simultaneously

\
Ex: Two treatments + how many times a week

Multivariate Design

Two or more dependent variables

\
Ex: Effect of Early Start Denver Model (ESDM) for toddlers with ASD on IQ, adaptive behavior and ASD Dx. Toddlers randomly assigned to 1 of 2 groups (EDSM or control \[standard care\]) for 2 years.

Single Subject Design

Individual performance even if there are multiple individuals

Considered "best" when measuring behavioral change or modification at individual level

Withdrawal Design (Reversal Design)

Refers to the removal of intervention/treatment during one or more phases to demonstrate the effects that withdrawal has on target behavior

\
Allows researchers to demonstrate cause-effect relationships between behavior and intervention/treatment

Multiple Baselines Design

a single-subject design in which the effectiveness of a treatment is assessed on two or more behaviors or across two or more situations

Alternating Treatments Design

Compare the events of two or more treatments on the same target behavior

Rapid alternation of two or more treatments

Distribution

Summary of the frequency of individual values or ranges of values for a variable

Central Tendency

An estimate of the center of a distribution of values

mean, median, mode

Dispersion or variability

Spread of values around the central tendency

Correlation

A single value (r) that describes the degree of relationship between two variables

Significance (p-value)

If it is less than 0.05 (typically ≤ 0.05) is statistically significant

Conclusion Validity

Is there a relationship between the cause and effect?

Internal Validity

Is the relationship causal?

Construct Validity

Can we generalize to the constructs?

External Validity

Can we generalize to other persons, places, times?

Type I Error

Concluding that is a relationship when in fact there is not (false positive)

Finding a Relationship Where There is NONE

Type II Error

Concluding that there is no relationship when in fact there is (false negative)

Finding NO Relationship Where There IS One

Trustworthiness: Credibility

The research is an adequate representation of participants perspectives

Uses triangulation (of sources, methods, investigators)

Trustworthiness: Transferability

The researchers ask whether findings hold in other contexts (i.e., future participants)

Can results be "transferred" to your clients and/or setting

Trustworthiness: Dependability

The researchers report the methods such that they can be replicated

Trustworthiness: Confirmability

The researchers attempted to void bias and remain as neutral as possible

Team of researchers check each other

Respect for Persons

treating persons as autonomous agents and protecting those with diminished autonomy

Beneficence

Investigators should design research studies to maximize benefits and minimize risks to individuals

Justice

The burdens and benefits of research should be fairly distributed among individuals, groups, societies, etc.

Institutional Review Board (IRB)

A panel of people who review research proposals with respect to ethical implications and decide whether additional actions need to be taken to assure the safety and rights of participants.