Module 5 - Research in Communication Disorders: Group Designs

Group Designs

Designs used in true experimental and some quasi-experimental research designs where you have a group of individuals that you investigate, and they respond to some sort of independent variable

Performance

With group scores, what are we worried about?

Single-subject design

What type of research design is good to see how individual people score?

Manipulated

In experimental research, is the IV manipulated or selected to determine the relationship between the IV and the DV?

Relationship to the DV

In descriptive research, is the IV manipulated or selected to determine the relationship between the IV and the DV?

Experimental

In what type of research design is the IV manipulated and the DV is affected / impacted and measured / observed?

Descriptive

In what type of research design is the IV is selected because of it’s relationship to the DV?

• Experimental

• Descriptive

What are 2 types of research designs?

Experimental Research Designs

Examines causality under controlled conditions

Research Design

A plan that happens before the study starts that identifies the population of interest

• Age limitations

• Disabilities

• Geographical area

What are things you identify the population by discussing

Research Design

What will answer your research question?

• Having a control group

• Doing pre-testing and post-testing to see if there is change over time

• Pre-testing to ensure initial group equivalency between both

• Blinding

• Everyone is trained to collect data or administer IV

• Make study shorter

• History effect

What are some ways to establish how the extraneous variables / threats will be controlled?

Faulty

Do poor or less rigorous research designs lead to strong or faulty results?

• Answer the research question

• Control extraneous variables and threats

• Be valid and reliable

The worth of a research designs is judged by its ability to do what?

Internal Validity

The degree to which the relationship between the IV is observed without the influence of extraneous variables

Extraneous / confounding variables

Weak research designs struggle to control what?

• How the study will unfold

• The results of the study

• How the IV will affect the DV

• The best guess that we make at the beginning of the study prior to picking the sample

• Thing you test and see whether you are right and wrong

What are some things that constitute the hypothesis?

• Theory

• Best available research

What is the hypothesis based off of?

Hypothesis

The thing that is necessary before you even are able to develop the research designs

Research question

What is the thing that is going to answer the hypothesis?

Hypothesis

What is the thing that the research question is based off of?

• Does it link the independent and dependent variables?

• Is it testable?

• Does the theory or logic behind the hypothesis make sense?

What are questions to ask to evaluate the hypothesis?

Sampling Protocol

Identify the population of interest (target population)

Small amount of observations from the population

What is the goal when extracting and creating the sampling protocol?

Simple Random Sampling

Every individual in the population has an equal change of being in the sample and is selected by chance

Stratified Sampling

Divides the target population into strata, such as geographical regions

Strata

Subpopulations

Random sampling occurring with each strata

What happens after the strata is selected?

Convenience Sampling

Sampling participants from a pool of individuals that are available because of their close geographic proximity or other reasons

The ability to generalize results to larger populations

What does convenience sampling severely limit?

Representative Sample

Sample represents all members in the population who have or might be at risk for having

Individuals

What impacts the power of the study?

Thorough description

What increases the believability of a research results?

Power

Measurement of design sensitivity and it detects the change in the dependent variables in response to the introduction of the independent variable

Likelihood of finding the true treatment effect

If the power is low, what else is low?

• Plan

• Implementation

• Statistical analysis

What does sensitivity of research design depend on?

Good research plan

What is power tied to?

Research design

The better of THIS you have, the better the sensitivity it is and how well it was implemented

False conclusions

What do poorly selected research designs cause?

• Internal validity issues

• Measurement reliability

• Choice of statistical analyses

• Sample size

Power is affected by a number of issues. What are some reasons?

A Priori Power Analysis

Analyses that are used before the study that starts to determine adequate sample size and ensure increase statistical power

Alpha level

What is tied to statistical significance?

80%

What is the recommended power?

Effect size

This is part of power that works with practical significance

Alpha level

This is the part of power that works with statistical significance

Priori

Effect size should be determined by what?

• Determine adequate sample size a priori

• Control threats for validity

• Tighten up procedures for data collection, administration of the IV, and valid and reliable measurement of DV

What are some ways to avoid issues with low power?