Research
In Experimental Research, the researcher manipulates an IV (while controlling for potential EV) to examine what effect the manipulation has on DV. To establish a functional relation between the IV and DV.
Hypothesis
- Null Hypothesis - no relationship between variables
- no difference in IQ between people between with X and Y
- Directional Hypothesis - we state the direction of the relationship between two variables
- X will have higher scores than Y on IQ tests
- Non-directional hypothesis - we don't know and won't speculate about the direction of the relationship between two variables
- There will be a difference in IQ scores between X and Y
IRB
- ethical research when using humans as subjects
- types of research that need IRB
- biomedical
- biology
- drugs
- devices
- behavioral
- surveys
- experiment
- questionnaires
- all other types of human subject research
- biomedical
- key elements of informed consent
- explanation of procedurees
- risks or discomforts
- benefits
- confidentiality
- termination of participation
- costs/compensation
- questions/contact information
- signatures and copy
Qualitative vs Quantitative
- Qualitiative
- open ended observation to obtain naturalistic data
- DESCRIPTIVE overview of data/observation
- SMALL sample
- in-depth description of data/observation
- finding themes, patterns and relationships
- understand perspective: WHY a particular outcome
- SUBJECTIVE
- Quantitative
- time tested research method
- observations are MEASURED and EXPRESSED in NUMERICAL values
- objective
- systematic process
- numerical dat is used to test the hypothesis
- LARGE sample
- statistical analyses of numerical data
- understand cause-effect relation; WHAT causes an outcome
- OBJECTIVE
Questions
- What is the statement of the general problem? Is it justified with current literature?
- Was the rationale for study logical and convincing? Give one reason why you think so
- Did authors proved current, thorough, and accurate literature review?
- State the aim/purpose and research questions/hypotheses of the study. What is your opinion of the research question?
- Was the introduction clearly written and well organized? What would you change?
Research Design
- specific plan for selecting the IV(s) & measures to assess the DV(s) in order to answer specific research questions about their relations
- Two purposes of a design
- To permit investigator to answer research questions
- To control extraneous variance/noise (variables that could affect the outcome of the study)
- Major Classification of Research Designs
- Single Subject Designs (SSD) - purpose is to demonstrate a functional relationship between an INTERVENTION (IV) and a change in target behavior (DV)
- behavior of one or few individuals is studied
- each person's performance evaluated in detail to examine affects of IV on DV
- repeated, systematic measurement of DV
- before, during, after manipulation of IV
- DV is usually a human characteristic/behavior
- IV usually involves application of intervention
- Types of SSD
- A-B design
- used to quickly assess the effects of a treatment
- phase 1 (A)
- measure baseline response
- phase 2 (B)
- introduce treatment while measuring response
- Baseline - Intervention
- get a baseline for trial 1 - 5 and then see what happens at interventional stage
- like monitoring, trials at baseline and then a measured response after a change is made
- can have a delayed effect
- effect happens after many trials
- effect happens after many trials
- an immediate effect
- happens right after baseline trials
- happens right after baseline trials
- small effect or large effect
- variable baseline and intervention
- jagged, all over the place
- jagged, all over the place
- stable baseline and intervention
- large jump to intervention then stable
- large jump to intervention then stable
- combined effects
- baseline trend = intervention trend
- basically linear; x = y, y = x
- baseline trend does not = intervention trend
- can have a delayed effect
- phase 1 (A)
- used to quickly assess the effects of a treatment
- Withdrawal Designs
- A-B-A
- simplest of single subject designs
- repeatedly introduces and withdraws treatment
- A - Baseline phase
- B - Treatment phase
- A - Withdrawal phase
- Objective
- A-B-A-B
- most powerful demonstration of causality and functional relationship
- withdrawal phase allows more reliable assessment of intervention effects
- A-B-A-C
- A-B-A
- Multiple Baseline Designs
- Multiple Treatment Design
- A-B design
- Group Designs
- Study one or more "large" subject groups to examine affect of IV on DV
- Compare "average" performance
- Between - Subjects Design - Treatment Based Design
- primary purpose is to examine the effect of the IV on two or more groups of subjects
- with intent to determine whether groups perform similarly or not
- treatment research with a control group
- experimental (target) group receives treatment
- control group does not receive treatment
- experimental (target) group receives treatment
- Experimental
- experimental research comparing the performance of two or more groups on one experimental task or condition
- example: do typically developing children and children with hearing impairment with hearing aids differ in their speech recognition
- example: comparing young vs old in some aspects of hearing/listening skills
- example: do typically developing children and children with hearing impairment with hearing aids differ in their speech recognition
- experimental research comparing the performance of two or more groups on one experimental task or condition
- Within - Subjects Design
- purpose: compare the performance of a single group on 2 or more conditions
- all subjects in the study receive all treatments or levels of a treatment
- also called repeated measures design
- the main advantage over the between subject design is that it requires fewer participants, making the process much more streamlined and less resource heavy
- example: does the speech intelligibility or pre-school hearing impaired kids improve comparably after receiving phonological treatment one and phonological treatment two
- example: does speech perception better in BTE than CIC in the presence of wind noise?
- experimental design
- all subjects receive all experimental tasks and/or all levels of one or more of the experimental tasks
- example 1: development changes in ABR/OAE in children longitudinal or cross-sectional study
- example 2: does the comprehension of oral directions differ as a function of input speech rate (slow, normal, fast)
- Mixed - Subjects Design
- studies that combine both between-subjects design component (have 2 or more groups) and within-subjects design component (all subjects in all groups receive all conditions or tasks in the experiment)
- make comparisons between subjects/groups (is there difference between groups on task conditions)
- make comparisons within subjects/groups (how do subjects within each group perform on each level of the IVs)
- Mixed Subjects Studies can be
- transmission studies
- experimental (non-treatment) studies
- studies that combine both between-subjects design component (have 2 or more groups) and within-subjects design component (all subjects in all groups receive all conditions or tasks in the experiment)
- Single Subject Designs (SSD) - purpose is to demonstrate a functional relationship between an INTERVENTION (IV) and a change in target behavior (DV)
- Validity in Research Design
- internal validity
- internal validity is the confidence that we can place in the cause and effect relationship in a study
- could there be an alternative cause that may explain my observations
- external validity
- external validity is the ability of the study to generalize
- whether results obtained from a small sample group, often in laboratory surroundings can be extended to make predictions about the entire population
- internal validity
- Counter balance
- to minimize carryover effect/order effect
- order in which the treatment or experimental conditions are administered
- not practice if you have > 5 conditions: expensive and very time consuming
Research Strategies
- Descriptive research
- naturalistic/observational
- no intervention in study, does not manipulate or control anything
- simple observation
- no manipulation of independent variable
- survey research
- asking rather than observing
- obtain information about opinions
- questionnaire
- retrospective research
- obtaining data from already existing clinical data/records
- no direct experimental manipulation
- naturalistic/observational
- Experimental research
- bivalent experiments
- study the effects of two values on one IV on DV
- multivalent experiments
- study the effects of several values of IV on DV
- parametric experiments
- simultaneous effect of more than on IV on DV
- example: study the speech of hearing impaired listener would have to consider
- level of background noise
- distance between talker and listener
- acoustic character of talkers speech
- type and severity of hearing loss
- amplification properties of HA
- bivalent experiments
- Combined experimental & descriptive
- Correlation research
- purpose is to investigate relations between/among 2 or more variables by measuring how changes in one variable (X) corresponds to another (Y)
- three possible outcomes
- positive correlations
- both variables increase or decrease at the same time
- negative correlations
- indicates that as the amount of one variable increases, the other decreases (and vice versa)
- no correlation
- indicates no relationship between the two variables
- positive correlations
- three possible outcomes
- CAUTION 1 - Correlation does not imply causation
- if 2 variables are correlated it means there is an association, not necessarily a causal association
- Home Computer -> Child's vocabulary
- if 2 variables are correlated it means there is an association, not necessarily a causal association
- CAUTION 2 - "significant" correlations may be due to uncontrolled variables
- association between X and Y may be spurious (lucky) because another variable may be responsible for moderating the association between X and Y
- association between home computer use and vocabulary may be due to various other factors like parent education
- association between X and Y may be spurious (lucky) because another variable may be responsible for moderating the association between X and Y
- purpose is to investigate relations between/among 2 or more variables by measuring how changes in one variable (X) corresponds to another (Y)
Scale
- Nominal
- Ordinal
- Interval
- Ratio
Variables
- anything that can be varied
- Independent Variable - variables manipulated by the experimenter
- experimental group (drug) vs control group (placebo)
- hearing aids vs no hearing aids
- slow stimulation rate vs fast stimulation rate
- CAUSE
- Dependent Variable - variables that respond to experimental manipulation
- static admittance
- speech in noise performance
- wave V latency
- EFFECT
- Extraneous Variable - aka nuisance variable
- Testing environment
- Instructions
- Monaural vs Binaural
- Age
- Motivation