BIOSTATS FML #1
Card Set Information
BIOSTATS FML #1
BORING ASS SHIT
What is science?
- A method of inquiry.
- Explore patterns of interrelationships
- Explore cause and effect relationships
- Theory + research methods + statistics
What is human inquiry?
Causal and probablistic reasoning (variability)
Second hand knowledge
What are the types of second hand knowledge?
Traditional - friends/family
Authority - professors/president
T/F - Do todays actions cause tomorrows effects?
T/F - Effects are more likely to occur if associated with a cause?
6 errors in human inquiry?
Made up Information
What does the research process tell us?
How and why things happen.
T/F - Can research be disorganized?
False - organized systemic sequential
7 components of research process
Identify a problem
Clear statement of problem
Sub divide problem
Create a study design
Whats the most important factor in writing a clear statement of problem?
Written to imply measurement
What happens when you sub divide a problem?
You develop a hypothesis
What makes a study design good?
Controls for extraneous problems as much as possible to ensure findings.
Wanna look at design, validity, reliability, sample section processes, inclusion criteria
What occurs when you collect data?
You are applying the methodology into action and making sure no contamination occurs during the process
What two types of data can you interpret?
Either supports or refutes your hypothesis
What is the hourglass notion of research?
Narrow it down/focus in
Generalize back to questions
What is a variable? Types?
Outcome that can take on more than one value.
Is it best to measure a variable as accurately as possible?
Yes, the more precise the more useful the measurement is
What are independent variables?
Treatment/conditions controlled by researcher
Need at least 2 values of IV
What are dependent variables?
Outscomes of research study
Is an independent variable confounded?
Does the dependent variable affected by independent variables?
Yes, it is sensitive to changes of IV
What are theories?
systematic explanations of observed facts
What are hypothesis?
expectations about empirical reality
What are laws?
universal generalization about facts
Why is a research hypothesis needed?
To compare to the null hypothesis and to be tested during research process
Does the hypothesis need to reflect the general problem of the study?
Yes, so you can test the general problem and hypothesis should be made into a mathematical statement
What is a Null Hypothesis?
Starting point for analysis and used for comparison.
States there's no relationship between independent and dependent variables.
: m1 = m2
What is a research hypothesis?
States a relationship exists between IV and DV
: X1 not = X2
T/F Research hypothesis is equal.
FALSE - statement of inequality H1: x1 not= x2 always.
What is a nondirectional research hypothsis?
groups are different, but no direction specified (H1: x1 not= x2)
What is directional research hypothesis?
group is different, and direction is specified (H1: x1 >/< x2)
What makes a good hypothesis?
Brief and to the point
Reflects theory or literature
Posits a relationship between variables
What does the population represent?
The entire group of interest
What consists of a sample?
The representative portion of a population
Should results of a sample be generalized to a population?
Yes, if sample was chosen randomly
Do observed differences result from the treatment or from chance?
What is significance level?
Risk associated with not being 100% certain the null hypothesis is wrong
What makes a good research paper?
Research conducted and sequence it was carried out.
Narrative story in third person
What are the 7 parts of a research study?
What do you want in a good intro?
Capture readers attention
Purpose of study
Background to justify study
Objective of Study
Identify limitations and assumptions
What is in a literature review?
Establishes need for study.
Other relevant studies
Historical perspective of teh problem
How do you review literature?
Define idea in general terms using general sources.
Search through secondary sources
Search through primary sources
Write the proposal
What are considered to be general sources?
What are considered secondary sources?
Books on specific subjects
Reviews of research
"once removed" from the original work
What are considered primary sources?
Educational resources info center (ERIC)
Is the internet a good source to use?
NO, its unregulated
What is the best source?
Journals because its written in specific formats and are peer reviewed
What is an abstract?
1-2 paragraph summary of what journal is about
What programs are out there to search for abstracts?
What is the best thing to use for searching for research?
Computers, but its EXPENSIVE
What are the 5 main points in methodology section?
Study design & protocol
Measurement and Observation
What is a good study design and protocol?
Study and control grp
What section do you talk about blindness in the methodology section?
Measurement and Observation
What does the statistical analysis section in methodology talk about?
what programs/tests did u use and why
What do you talk about in the discussion section of a research paper?
The answer to the question and the evidence and counter evidence to it.
What 3 sections does a proposal consist of?
Intro, Lit Review, Methodology
What are the 6 basic principles of ethical research?
Secure personal info
Don't coerce to be participant
Explain study to them after its complete
Split equally all beneficial outcomes ($$)
Know what is involved in the study
How to ensure HIGH ethical standards?
Report info as grp not individual
Start with small sample if possible if side effects
If harmful find participants who have been naturally treated
What does IRB stand for?
Institutional Review Board
What do the Institutional Review Board do?
Look at proposed methods and study designs whether it is ethical.
Looks out for the participants safety
Only approve proposals if benefits > risks.
AHHHH what is the Declaration of Helsinki?
Statement of Ethical principles
Guildlines to conduct research
US research rules?
Guided by federal government Title 45 CFR part 46
Established IRB's to review and monitor research
What is the structure of research that is the "glue" that holds all of the elements in a research project together?
How is a one grp study represented? control grp?
1 line and has no control grp
R O X O or R X O
How is a 2 grp study represented?
2 lines one control one test grp
R O X O = test
R O X O = control
What each letter R O X O, represent in a study design?
R = randomly assigned
O = pretest/posttest measurements
X = treatment
~ = no
~X = no treatment
What is another way of saying R O X O? R X O?
R O X O = 1 grp before and after only design = 1GBAO
R X O = 1 grp after only design = 1GAO
2 lines of ^^ = 2 grp after only design = 2GAO
What are the 6 single grp threats to measure outcomes?
If a relationship is shown does it mean it's a causal one?
No, other variables can cause the same outcome.
What is validity?
Accuracy, everything is correct
What are the 4 validities?
What is internal validity?
Evidence to prove what you did caused what you observed
What are the threats to internal validity?
alternative causes that can cause differences in the posttest
What is history threat?
A historical event that caused the outcome not the treatment.
What is maturation threat?
Same exact outcome even without treatment
What is testing threat? What design does it only occur in? How to prevent it?
Pre-post test design.
The pretest prepared participants for program.
Give a posttest only but history threat becomes problem.
What is instrumentation threat? What design does it occur in?
Pre/Post test design.
Change in instrumentation leads to the outcome.
What is mortality threat?
Participants dropping out of study.
Only problem in a pre-post design can skew results.
Ex. someone with low score drops out
Whats the only way that you can say generalization?
Did random sampling/selection
What does it mean when the grps were not comparable before the study?
You had selection bias/threat
What is selection threat?
Any factor other than the program that leads to posttest differences between grps
What is a HUMAN activity?
Applied social research
The 3 types of social threats?
Diffusion or Imitation of treatment
What is Diffusion or imitation of treatment?
The control grp learns about the program from the test grp and decreases chance of seeing any program effect.
What is compensatory rivalry?
The control grp knows that the program grp is getting and becomes competitive with them.
Increases chance of seeing no program effect even if program is effective
What is resentful demoralization?
The control grp know what program grp gets
Get discouraged and angry.
Get exaggerated posttest differences so program looks more effective than it actually is.
What are pitfalls you want to avoid?
Loving idea too much.
Try something new not trivial.
What is inferential method based on?
Inferring sample to a population
What is probability sampling?
The likelihood of any member of the population being selected is known.
What is nonprobability sampling?
Likelihood of any member of the population being selected is unknown.
What does simple random sampling do?
Makes sure each member of a population has an equal and independent chance of being chosen.
How do you chose a simple random sample?
Know your population.
List all members in it.
Assign them a number.
Use a criteria to select sample.
Keys to success in simple random sampling?
Selection criteria should not be related to factor of interest!
What is systematic sampling?
Divide population by size of desired sample ex. 50/10 = 5
Start at a random # and choose participant by every other 5.
What is stratified sampling?
Random sample is selected that reflects the proportions in the popultion ex. 40% male 60% female
When do you use stratified sampling?
When people in the population differ along some characteristic that relates to factors being studied.
What is cluster sampling?
Randomly select grps of individuals.
NO RANDOM INDIVIDUAL SELECTION
How do you avoid BIAS in cluster sampling?
Grps have to be homogenous.
What are the 2 nonprobablity sampling strategies?
What is convenience sampling?
Easy or captive sampled population
What is quota sampling?
Select people of interest until quota is met
What is sampling error?
Difference between sample and population characteristics.
Want this to be low by increaseing sample size
Does larger samples ignore the power of scientific inference?
How big is too big for sample size?
Larger the better more representative but more expensive.
When do you need a large sample size?
Large variability within each grp
Small differences between each grp
Do all measurement data contain errors?
What is reliability?
Consistent, stable, dependable, predictable, accurate.
Low amt of errors.
What is the true score model?
X = T +or- E
An observed score = true score plus an unknown amt of error
Always some type of error
Ave. observed score good estimate of true score
What assumptions can you assume about true score errors?
Random errors not systemic
Can be + or -
No correlation with magnitude of true score
T/F Is error the foundation of unreliabiltiy?
What are the 2 types of measurement error?
What is systemic error?
Predictable and not part of true score model.
Has to do more with validity of a measuring instrument.
What is random error?
Due to chance. Ex. person moved while being measured.
Can effect observed score.
What causes random measurement error?
The person taking the measurement
The instrument itself
Variablility of characteristic being measured because its not stable
Can the tester or subject cause error?
What are the three types of reliability?
What is test-retest reliability?
Doing a self-report, physical, physiological test then doing the same test again at a different time.
Minimize testing effects.
Have sufficient intervals before next test.
What is rater reliability?
The rater/s is consistent in applying criteria.
Make them practice, take tests, training, grade/score them to be more reliable.
Do you check intra-rater reliability or inter-rater reliability first?
What is alternate forms reliability?
Paper tests, psychological tests.
Parallel forms of tests
Equivalent forms of tests
What do you want in internal consistency with tests?
Measure a single thing and nothing else
What is split-half reliability?
Split test into two halves.
Score each separately and correlate the scores together.
If internally consistent then have high correlation.
Prefer method to establish reliability since there is no time lag.
What is the range of reliability coefficients?
What is reliability a pre-requisite to?