Week 9-13 Flashcards
What non-parametric test do you use for independent samples?
Mann-Whitney U test =Z
What non-parametric test do you use for independent samples?
Mann-Whitney U test
What non-parametric test do you use for paired samples?
Wilcoxon-Rank Sum
What non-parametric test do you use for a one way ANOVA?
Kruskal-Wallis test
What non-parametric test do you use for a repeated measures ANOVA?
Firedman’s test
What non-parametric test do you use for a mixed factorial ANOVA?
There is currently no know test for this design
What non-parametric test do you use for paired samples?
Wilcoxon-Rank Sum =Z
What non-parametric test do you use for a one way ANOVA?
Kruskal-Wallis test =H
What non-parametric test do you use for a repeated measures ANOVA?
Firedman’s test X2f
What non-parametric test do you use for a mixed factorial ANOVA?
There is currently no know test for this design
What two reasons would make you use a non-parametric test?
When you have ordinal or nominal data,
When your data is not normally distributed
What do you need to calculate effect size?
desired alpha, desired effect size, level of Power
Can the statistics of effect size and power be applied to all statistical tests that involve significance testing processes?
Yes
Is there ALWAYS a chance of us as researchers making an error?
yes always.
1 - a is the ____ ?
probability of you being right
Beta is the ___ ?
probability that you incorrectly retained the null hypothesis
Historically, what type of error was ignored ?
Type II error
The chance of not finding a difference even though one is really there is? (you said there was no difference, but there there really was)
A type II error
What is beta normally estimated to be?
.2 (20%)
What is power?
The probability that a statistical test will correctly reject the false null hypothesis (1 - b)
Do we want power to be big or small?
big as possible
Are 1 - a and 1 - b related?
Yes
What affects power? (4 things)
a -alpha, The alternative hypothesis, Sample size, population variance,
What two ways can you increase your power in a study?
A larger sample size, and less variance between the sample and the population
The measure of how big the effect that we’re measuring actually is, is?
Effect size
Does effect size take into account the number of people in your study?
Nope
Increasing your standard deviation (decreases/increases) your d (effect size) and increase in the difference between group one and two (decreases/increases) your effect size?
Decreases; Increases
Can you estimate the effect size before a study is conducted?
Yes (because if we estimate effect size, we can also estimate power)
Looking at prior research, Personal assessment, and use existing conventions, are all ways to estimate ___ ?
Effect size
Partial eta squared is the effect size reported when you have ___ ___ 1 IV
More than
What is the universally accepted power estimate needed in a study?
80%
Statistical ___ is the likelihood that the research study will detect an effect in a sample when one exists in reality?
power
___ Analysis is used during the planning stage of a research project?
POWER
What three components do you need to estimate required sample size?
Alpha level you wish to use; level of power you wish the study to have; the likely Effect Size
What is more powerful, a parametric or non-parametric test (which is also more likely to detect a difference if there is actually one there)?
Parametric tests, (Non-parametric tests also have less precision)
When you have a very very small sample size, should you use a non-parametric test, or a parametric test?
Non-parametric test
Is naturalistic observation High-Constraint or Low-Constraint?
Low-Constraint
Is Archival research High-Constraint or Low-Constraint?
Low-Constraint
Are Case studies High-Constraint or Low-Constraint?
Low-Constraint
Are surveys High-Constraint or Low-Constraint?
High-Constraint
Are program evaluations High-Constraint or Low-Constraint?
High-Constraint
Are field experiments High-Constraint or Low-Constraint?
High-Constraint
The Observation and systematic recording of naturally occurring events is called?
Naturalistic Observation
Cohen’s ___ is the measure of agreement among rater’s in the measures of observational research
Kappa
Cohen’s Kappa over (.20, .50, .70, .90) is a generally suitable level of agreement
.70
Cohen’s Kappa is commonly used to calculate inter-___ ____
inter-related reliablility
Cohen’s Kappa is ranged from 0 - _?
1
What type of research is more constrained. Case studies or naturalistic observation?
Case Studies
Can you test hypotheses in case studies?
Nope
Time sampling, Event sampling, and Trait Ratings are all techniques of ___ ___ techniques
Structured Observational Techniques
Time sampling, ___ sampling, and Trait Ratings are all techniques of Structured Observational techniques
Event Sampling
Time sampling, Event sampling, and ___ Ratings are all techniques of Structured Observational techniques
Trait Ratings
In multiple baseline across participants techniques: ___ is a term used when they introduce the IV at different times for different people
Staggering
Explain what the a and b stand for in a reversal design of ABA design?
A = baseline measure, B = the introduction or reintroduction of a IV
What two types of Single-Case Research Designs are there?
Reversal designs and Multiple-Baseline Designs
The elimination of ___ ___ is a positive advantage of single-case research designs
Error Variance
What type of design is conceptualised as a “within-subjects repeated measures design of only one participant” ?
A single-case Experimental Design
Responsibility to ensure the ___ and dignity of the individuals who participate in research studies; and the responsibility to ensure the ___ reports of research are accurate and honest are the two categories of ethical responsibiltiy
Welfare; public
Responsibility to ensure the ___ and dignity of the individuals who participate in research studies; and the responsibility to ensure the ___ reports of research are accurate and honest are the two categories of ethical responsibiltiy
Welfare; public
Responsibility to ensure the ___ and dignity of the individuals who participate in research studies; and the responsibility to ensure the ___ reports of research are accurate and honest are the two categories of ethical responsibiltiy
Welfare; public
The ____ code is a set of 10 ethical guidelines for treatment of participants in research that was created in 1947
Nuremberg Code
In 1964 the Declaration of _____ was established as guidelines for medial research involving humans
Helsinki
University of Canberra _____ for Responsible Practice in Research
Guidelines
University of Canberra Guidelines for Responsible _____ in Research
Practice
- Avoid Harm
- Informed Consent
- Deception
- Privacy and ______
- Institutional Approval
- _____
- Record Keeping
- Inducements for Participating in Research
- Debriefing
Confidentiality; Competence
- Avoid Harm
- Informed Consent
- Deception
- Privacy and Confidentiality
- Institutional Approval
- Competence
- _____ Keeping
- Inducements for Participating in Research
- _______
Record;Debriefing
- Avoid ____
- Informed Consent
- Deception
- Privacy and Confidentiality
- Institutional Approval
- Competence
- Record Keeping
- _____ for Participating in Research
- Debriefing
Harm;Inducements
Researchers must not ___ participants, either physically or psychologically
Harm
Particularly with harm, participants must be informed of the ___
Risks
Psychological harm can be created in the study you conduct but also self-____ by the participant
generated
A ____ ____ sheet is usually filled out before prospective participants are involved in a study
Participant information sheet
Telling participants at the end of a study the complete complete explanation and justification of the research is called ____
Debriefing
A ___ ___ sheet and ____ are used in every research study
participant sheet and debriefing
Clinicians have a responsibility to provide the best possible treatment for their patients; this is known as Clinical ____
Equipoise
Clinical Equipoise is the responsibility to provide ….
The best possible treatment for their patients
it is ethically irresponsible to compare two treatments in a research study where one is known to be less effective; this refers to clinical _____
equipoise
The idea that it is sometimes difficult to provide complete information is known as the problem of ____
Infomation
The problem of _____, problem of ____ and problem of _____ participation are all components to think about when thinking of informed consent
Information; Understanding; Voluntary
If the goal of the research is to study ‘normal behaviour’, often _____ is used
deception
___ deception involves withholding information from the participants
Passive
____ Deception involves presenting false or misleading information to participants
Active
Is active or passive deception more common
Passive
If deception is used, it needs to be ____
Justified
After deception is used, _____ ALWAYS needs to occur
Debriefing
information obtained from a research participant will be kept secret and private is known as _____
Confidentiality
What is a mainstay of ethical research?
Confidentiality
What supports continued research by continuing trust?
Confidentiality
Making participants identified by a code number is known as ____
Anonymity
All research projects conducted in organisations need to be reviewed by an ___ ___ Board
Institutional Review
Researchers have appropriate training and experience conduct a study is the ethical issue of _____
Competence
What are the two big issues in Scientific Intergrity?
Fraud and Plagiarism
The explicit effort of a researcher to falsify and misrepresent data is known as ____
Fraud
Usually involves creating data or changing it to support a hypothesis; this is known as ____
Fraud
Is fraud different from error?
Yes of course
____ occurs because of Competition within academia
Fraud
The result of an existing bias towards publication of significant findings is ____
Fraud
What are the safeguards against fraud?
Replication and Peer Review
____ and ____ review are the safeguards against fraud
Replication; Peer
1 - a is the chance that you are ____
Right/Correct
1 - b is the probability of you ……
correctly retaining the null hypothesis
Power is …
the probability that a statistical test will CORRECTLY reject a null hypothesis
What affects power?
Alpha, the alternative hypothesis, (how big the difference between u0 and u1), sample size, and population variance
Alpha, the alternative hypothesis, (how big the difference between u0 and u1), sample size, and population variance affect ____
power
If we are looking for a really big difference, it is normally ___ to find
easier
The larger the difference between the null and alternative hypothesis the larger the ____
power
If we change our number of participants, or the population variance, we change _____
power
if you increase your standard deviation, you decrease your ___ ___
effect size
___ ___ is not affected by N
effect size
How can we estimate effect size before a study? (3 things)
Prior research, Personal Assessment, existing conventions
It is possible to calculate the sample size required in order to have a specific ___ level
power
What do you need to know to estimate the required sample size?
Alpha level, Level of power you wish to achieve, the relevant effect size
What is the difference between A priori power and Retrospective power
A priori is estimate before, Retrospective is after
What is a good economical/logistical reason for estimating power?
It can help you optimise your sample size and not waste time or effort trying to get to many people
No strict distribution assumptions; wider range or variable type can be used (nominal, ordinal); simpler calculations are the advantages for ….
non-parametric tests
You are less likely to find a difference in a …. than a ….
Non-parametric test; parametric test
There is loss of precision in a ….
non-parametric tests
Is there a non-parametric test to test interactions?
no
The ___-___ _ non-parametric test ranks all the values, then compares ranks between the two groups)
Mann-Whitney U
The absolute difference between the scores for each individual are calculated, and then these values are ranked. The sum of all the negative and positive ranks is calculated separately and evaluated against a table of critical values: this process is used in what calculation?
Wilcoxon Signed Rank Test
All data is ranked from lowest to highest (ignoring groups). Ranks for the different groups are then seperated and summed. A value (H) is then calculated representing how big the differences between the groups are. This value is compared against the Chi-square distribution
Kruskal-Wallis test
What is the non-parametric test for a one way ANOVA? also what would you use to follow it up?
A Kruskal-Wallis test, and post hoc tests of Mann-Whitney U with the Bonferonni Correction
The value for each individual are ranked. These ranks are then compared summed for each sample, and compared against the chi distribution
A Friedman’s Test
A way of measuring the degree of agreement among raters - the value is a score representing how much consensus there is between the ratings of different judgements. This is known as?
Inter-rater reliability
Data obtained reflects behaviours in the participant’s natural environment; Allows collection of data where experimental techniques are unethical or impractical; and Allows collection of a large amount of data of a whole variety of behaviours are the three advantages of _____ ____
Naturalistic observation
“Obtaining informed consent, would defeat the purpose of _____ _____”
naturalistic observation
Allows collection of a large amount of data for a small group of people; Allows researches to study rare behaviours or disorders; and allows researchers to investigate complex issues. These are the advantages of what type of low constraint research?
Case study research
Inability to test hypotheses; Inability to draw conclusions about causality; and inability to generalise findings are the three disadvantages of what type of low-constraint research method?
Case study research
A vast amount of archival data is available; Data exists in a variety of forms; Data covers a vast range of different behaviours and issues; Often economical! are the 4 advantages of what type of low-constraint research method
Archival Measures
Researcher has no control of data collection processes; Selective deposit; selective survival; and high probability of missing data; are the disadvantages of what type of low constraint research method?
Archival Measures
Problem statements and hypothesis testing evolves in ___-Constriaint research methods
Low-Constraint
Data gathering process is less formal and more fluid in ___-constraint research methods
Low-constraint
When using low constraint research methods you can only generalise your findings to similar ____ ?
samples
When evaluating low constraint research data you first need to go home and ____ it
Code
Poor representativeness; Poor replicability; Causal inference; limitations of the observer; going beyond the data; are all limitations of what type of research methods?
Low-Constraint Research Methods
What are the two types of Single-Case designs?
Multiple base lines design; reversal design
What are the sub-types of the Multiple-Baseline design?
Multiple baselines across participants ; Multiple baselines across behaviours; Multiple baselines across situations
A stable _____, only introduce one ____ at a time, and stable ____ with IV are the methodological considerations of what type of single-case design?
baseline; IV; behaviour; Multiple-Baselines
Difficulties with generalizability and external validity; Poor experimental control; Intensive for subject; Observational issues; these are all the disadvantages of what type of design?
Single-Case Designs
Structured ____ techniques are those that use a clearly defined system for recording behaviour
observation
____ sampling; ____ sampling; and ____ ratings are the three common structured observational techniques
Time; Event; Trait
A record of a pre determined length of time is called ___ sampling
Time
A record of every time a particular behaviour occurs is called _____ sampling
Event
Use of a rating scale to indicate how well a behavioural description reflects the behaviour of an individual is called ____ ratings
Trait
The use of ______ is used to convey what a study was about, minimise any negative effect, convey the objective of the research, and explain the nature and justify and deception used, and answer any questions the participant might have.
Debriefing
Clinical equipoise is under the subheading of which ethical consideration?
Avoid Harm
Is power and effect size related?
Yes
What does effect size not take into account?
Sample size
The likelihood that the research study will detect an effect in a sample when one exists in reality is called ___ ____
statistical power
Cohens d small medium and large are?
.20; .50; .80;
Cohens f small medium and large are?
.1; .25; .40
Cohens eta squared small medium and large are?
.01; .06; .14
The exact significance value in non parametric tests is given when samples are normally under ___ people
20
What is the name of the difference between the sample data and population data?
Sampling Error
_____ _____ is the difference between the parameter and the statistic
sampling error
Will probability or non-probability probably lead to larger sampling error?
non-probability (non-random)
What is the standard error?
It is the standard deviation of the sampling distribution of a statistic
Confidence Intervals provide an _____ estimate of a parameter ( a range in which the parameter is likely to fall)
interval
A z score is also called a _____ score
Standardised
MS____ is the pooled sample variance,
error
MS_____ is the variance of sample means
treatment
