Basic Research

• answers a question
• adds to existing knowledge
• not meant to solve problems
• no direct practical application
• can be used to develop theories

Applied Research

• Purpose is to solve a problem
• or to provide practical application

Experimental

• random assignment of particpants to groups
• conditions are manipulated by researcher

Quasi-experimental

• no randomization
• control group or multiple measurements (on the same group/person)
• assess differences across groups
• across time
• across different tasks

Non-experimental

• no comparisons or
• no manipulation of conditions
• can be qualitative

Types of non-experimental research

• case studies
• ethnographic studies (groups of people)
• Historical design - use existing data
• Correlational designs - look for relationships
• Surveys
• Developmental - change over time
• longitunidal
• cross-sectional

Qualitative Research

• form of non-experimental
• "variety of analytic procedures designed to systematically collect and describe authentic, contextualized social phenomena with the goal of interpretive adequacy"

Quantitative Research

• Test of specific ideas, usually experimental
• numbers, statistical analysis important
• used to answer specific questions
• Hypotheses guide the research design

Qualitative Research

(characteristics)

• exploration of topics, ideas, insights
• Numbers/stats not as important
• used to answer general questions, in-depth exploration
• findings lead to hypotheses

List of types of research

• Quantitative
• Qualitative
• Basic
• Applied
• Experimental
• Quasi-experimental
• Non-experimental

Variable

Concept with quantitaive or qualitative value (not a constant)

Ex.

• age
• sex
• diagnosis
• score of Goldman-Fristoe
• IQ score
• Reaction Time

Independent Variable

• selected or manipulated by investigator
• fixed: what the participant brings (age, education, severity, type, etc)
• Manipulated: type of treatment, assignment to groups
• Planned by the investigator

Dependent Variable

• Focus of scientific observation
• values depend on the experimental manipulation
• Perceent intelligibility (DV) using a letter board vs. no letter board (IV)
• Number of requests (DV) by children with an without autism (IV)
• DV is what the particpant does

Extraneous Variable

• Everything else that could possibly affect the results of an experiment
• Can be nuisance variables

Examples of extraneous variables

• Demographics
• age, education, gender, SES, family environment
• Clinical
• severity of SLI, time since stroke
• Methodological
• order of test presentation
• testing site (at home, in lab, at school)

Not Variables

• Inclusion criteria
• based on the research question
• no variability across participants

Operational definitions

• Required to descirbe measurement and manipulation of variables
• identifying particiants
• defining categorical variables
• defining treatment

Data Collection

• data - information collected and coded
• reliability of data collected
• must be objective and repeatable

Ways to make collected data objective/repeatable

• Blinding
• minimize patient and researcher bias
• participant doesn't know if s/he's getting treatment or not
• researcher assessing results doesn't know which participants are getting treatment
• Inter-observer reliability
• measure consistency of observations
• different people assess data
• score tests, code transcripts, rate intelligibility

Research Design

Definition

What it involves

• blueprint for study
• participants
• variables
• implementation (methods)
• collecting data
• ethics of project
• Developed in concert with research question and hypothesis

Good Research Designs

• will answer the research question/s
• control for extraneous variables

Population

• group of interest

Types of Sampling

• random sampling
• convenience sampling
• representative sample
• large enough to answer the question (power)

Types of Designs

• Single group design
• 2-group design
• Complex design
• Factorial design

Single group designs

• one group, multiple conditions/tests
• pre-test/post-test

Two group designs

• 2 groups compared on same condition/treatment
• parallel - groups always the same
• cross-over design - groups participate in each condition (groups serve as own controls)

Complex Designs

• 3+ group design
• 3 or more groups
• compared on multiple conditions

Group Designs

• Equivalence of groups
• random assignment
• block-randomization
• stratified randomization
• matching

Single case designs (not case studies!)

• manipulate conditions
• compare participant to him/herself in different conditions
• no control group
• baseline phase is critical

Types of single case designs

• A-B-A Design
• A-B-A-C design

A-B-A design

• A - baseline
• B - intervention/treatment
• A - second baseline (withdrawal)

A-B-A-C

• A - baseline
• B - first intevertion/treatment
• A - second baseline (withdrawal)
• C - second intervention/treatment

Multiple baseline single case study

• multiple baselines - for each different variable
• Treat different variables in succession

Internal Validity

Definition

• "the degree to which the relationship between the IV and DV is observed without the influence of extraneous variables"
• data collected reflects the influence of the experimental manipulation, not something else

Types of threats to internal validity

• Ambiguous temporal precedence
• Differential Selection
• History
• Maturation effects
• Statistical regression
• Attrition
• Multiple test effects
• Instrumentation effects
• Additive and interaction effects

Threats to internal validity:

Ambiguous temporal precedence

Threats to internal validity:

Differential Selection

• Groups aren't really equivalent
• extraneous variable might be influencing the DV

Threats to internal validity:

History

OUtside events occurring during the study that might affect the DV

• Specific historical event or chain of events

Threats to internal validity:

Maturation effects

• changes in a person over time that might affect DV
• general changes in a person; specific to that person

Threats to internal validity:

Statistical regression

• over time, extreme scores will move (regress) towards the mean
• Statistical - affects groups that are non-randomly selected and with imperfect correlations

Threats to internal validity:

Attrition

• over time, participants will drop out of the study
• characteristics of drop-outs may be important
• differential attrition may be important

Threats to internal validity:

Multiple test effects

• repeated administration of same test
• practice effect - scores increase
• novelty effect - wears off; scores decrease
• test anxiety - may increase or decrease

Threats to internal validity:

Instrumentation effects

• changes in instruments (devices) over time
• wear and tear
• audiometers get out of calibration
• Changes in human (researcher) performance
• becomes faster, more efficient with data collection over time
• become more accurate with transcribing or coding over time
• change definition of codes over time

Types of External Validity

• Accessible vs. target population
• "convience sample"
• Description of independent variable
• multiple treatment interference effect
• novelty
• disruption effect
• experimenter effects
• pre-test and post-test sensitizing effects
• measurement of DV

External Validity

Definition

• Population validity: generalization to other people (generalize from sample to population)
• Ecological validity: generalization to other settings/environments (outside the lab)

Threats to external validity:

Accessible vs. target populations

• Convenience sample may not be representative

Threats to external validity:

Description of IV

• enough detail for replication is necessary

Threats to external validity:

Multiple treatment interference effects

• second treatment/task may be influenced by first
• ex. Participants receive 2 weeks of cycles treatment followed by 2 weeks of minimal pairs treatment

Threats to external validity:

Novelty effect

• a new treatment appears to work better simply because it is new (exciting and interesting)

Threats to external validity:

Disruption effect

• new treatment doesn't work as well because the experimenter is not well versed in administering it

Threats to external validity:

Experimenter effects

• behaviors differ across participants/testing sessions
• ex. more positive, more feedback, more interactive
• appearance influences performance
• male/female, dress, age
• observations or data recording differs

Threats to external validity:

Pre-test and post-test sensitizing effects

• exposure to the tests affects performance

Threats to external validity:

Measurement of DV

• operationalization of DV
• ex. DV - improved expressive langauge - what does this mean?
• fluency?
• words per minute? content units per minute?
• syntactic structure?
• sentence length?