cos 0 30 45 60 90
sin 0 30 45 60 90
1
- SI units:
- Femto
- Kilo
- Deci
Femto
- 10^{-15}
Kilo
- 10^{3}
Deci
- 10^{-1}
SI Units:
- Mega
- Pico
- Hecto
10^{6}
10^{-12}
10^{2}
SI Units:
- Tera
- Nano
- Centi
10^{12}
10^{-9}
10^{-2}
SI Units:
- Deca
- Giga
- Micro
- Milli
10^{1}
10^{9}
10^{-6}
10^{-3}
Convert 650 nm to SI units
650 x 10^{9}
Const. force (not velocity!) causes WHAT acceleration?
constant acceleration
When you throw a baseball, when is the only time it is accelerating?
Only when it is in contact with your hand
Increasing something by 25% is the same as multiplying it by what?
x 5/4
Define “mass”
the measure of an object’s inertia
Define “inertia”
- the ability of an object to RESIST its change in velocity
Where on a mass/object is its “Center of Gravity?”
is at the center of the mass/object
Where is the “Center of buoyancy?”
- at center of mass of the FLUID displaced by the submerged object
Scalar or Vector?
Mass
scalar
Scalar or Vector?
temperature
scalar
Scalar or Vector?
velocity
vector
Scalar or Vector?
speed
scalar
Scalar or Vector?
displacement
vector
Scalar or Vector?
acceleration
vector
Scalar or Vector?
force
vector
Scalar or Vector?
work
scalar
Scalar or Vector?
energy
scalar
Scalar or Vector?
weight
vector
Scalar or Vector?
charge
scalar
Scalar or Vector?
electric field
vector
Scalar or Vector?
magnetic field, B
vector
Scalar or Vector?
time
scalar
Scalar or Vector?
momentum
vector
Scalar or Vector?
impulse
vector
Scalar or Vector?
density
scalar
Scalar or Vector?
torque
vector
What are the 4 questions that test conceptual understanding?
- Can I visualize it?
- Can I draw a picture/graph/diagram of it?
- Can i explain it to someone in layman’s terms?
- Can i think of and describe real-life examples?
- Area of triangle
- formula=?
A_{tri}=1/2 bh
Vol of sphere formula
V_{sphere}=4/3 πr^{3}
SA of sphere formula
4pir^2
Manipulating equations mnemonics
SSISDDODIOSD
sqrt 2=
1.4
sqrt 3=
1.7
sqrt 2/2=
.7
sqrt 3/2=
.9
product of [H][OH] always equals:
1x10^-14
Doppler effect formula
deltaf/fs=v/c
Be careful of S.N.E.W.L
Qualifiers (Write these down!) “StrengthensNot ExceptWeakensLeast”
tan0=
sin0/cos0
sin^2x+cos^2x=
1
11^2
121
12^2
144
13^2
169
14^2
196
15^2
225
Multiplying 2 vectors: If answer is scalar, (ie work), also mult by what?
cos0
multiplying 2 vectors: if answer is a 3rd vector (ie torque), also multiply by what?
sin0T=Frsin0
Decimal equivalent:1/5
.2
Decimal equivalent:1/8
.125
Decimal equivalent:1/9
.11
For fractions where numerator is higher (like 13/5), what should you do to solve?
Create compound fraction: 5x2=10, left w/ 3/5 13/5 becomes 2x (3/5)
For fractions where denominator is larger, what should you do to solve?
“High/Low” methodChange denominator to 1 digit higher or 1 digit lowerex:3/7 changed to 3/6 and 3/8. Must be a little less than .5
When multiplying with scientific notation, what happens to the exponents?
You add them together
When dividing with scientific notation, what do you do with the exponents?
You subtract them
- Estimating Fractions
29/4 =?
=7.25
4 x 7=28
¼ is left behind
- Estimating Square Roots
- “High/Low” Method
What is the square root of 72?
- √81=9
- √64=8
∴ the square root of 72 is somewhere in between, so about 8.5
- Surface Area of a Sphere
- Formula=?
SA_{sphere}= 4πr^{2}
- Trigonometry
- All of the angles in any triangle must add up to?
180°
Trigonomic Relationships
- sinθ=
- cosθ=
- tanθ=
sinθ=Opp/Hyp
cosθ=Adj/Hyp
tanθ=Opp/Adj
”SOHCAHTOA”
Converting from DEGREES to RADIANS
- π radians=?
- 2π radians=?
- How many radians are in ONE CIRCLE?
- ∴, if something is turning at 12 rad/sec, it is making approximately ___ revolutions/sec
- π radians=180°
- 2π radians=360°
There are approximately 6 radians in ONE CIRCLE
- ∴, if something is rotating at 12 rad/sec, it is making 2 revolutions/ sec
Trigonomic Relations
- What are the INVERSES of sin, cos, and tan?
- sin^{-1}
- cosecant
- cos^{-1}
- secant
- tan^{-1}
- cotangent
Trigonomic Relationships
- tanθ= ?
tanθ=sinθ / cosθ
Linear & Non-Linear Graphs
- What does a graph look like for:
y=x
Linear & Non-Linear Graphs
- What does a graph look like for:
y = 1/x
Linear & Non-Linear Graphs
- What does a graph look like for:
y=x^{2}
Linear & Non-Linear Graphs
- What does a graph look like for:
y = |x|
(Absolute Value)
Linear & Non-Linear Graphs
- What does a graph look like for:
y = x^{3}
^{<em><span>(Cubic)</span></em>}
Linear & Non-Linear Graphs
- What does a graph look like for:
y = √x
Linear & Non-Linear Graphs
- What does a graph look like for:
y= ^{3}√x
(Cube Root)
Linear & Non-Linear Graphs
- What does a graph look like for:
y = ln x
(Logarithmic)
How does it look different than y= logx?
Linear & Non-Linear Graphs
- What does a graph look like for:
y = sinx
Linear & Non-Linear Graphs
- What does a graph look like for:
y = cosx
Linear & Non-Linear Graphs
- What does a graph look like for:
y = a^{x }
^{<span>(<em>Exponential)</em></span>}
Linear & Non-Linear Graphs
- What does a graph look like for:
y = 1/x
(Reciprocal)
Linear & Non-Linear Graphs
- For the equation X= ½at^{2}
- Which of the following relationships will be LINEAR?
- Which will be NON-LINEAR?
- X vs. t (or t vs. X)
- X vs. a (or a vs. X)
- a vs. t (or t vs. a)
X=½at^{2}
- NON-linear
- Linear
- NON-linear
Linear & Non-Linear Graphs
What does a graph look like for:
y = tan x
Graphs used as Answer Choices
- If a variable is changing exponentially, will be linear or non-linear on:
- a semi-log graph?
- a log-log graph?
The log of an exponential function becomes linear, so exponential functions graphed on semi-log axes end up being linear
- …AND a linear function would be non-linear
If the exponential function is graphed on log-log axes, the function will be non-linear
- The table below may be helpful*
- The semilog graph described is for a linear X-axis and a logarithmic Y-axis (log-lin type not lin-log)
Graphs used as Answer Choices
- Describe:
- a semi-log graph
- A semi-log graph has a logarithmic scale on one axis, but a linear scale on the other axis
Graphs used as Answer Choices
- Describe:
- a log-log graph
A log-log graph has a logarithmic scale on BOTH axes
Manipulating Equations
ay=vx^{2}/cq
- How are a and y related to each other?
inversely
Manipulating Equations
ay=vx^{2}/cq
How are x and q related to each other?
directly
Manipulating Equations
ay=vx^{2}/cq
How are a and q related to each other?
directly
Manipulating Equations
ay=vx^{2}/cq
How are a and c related to each other?
inversely
Manipulating Equations
X=½at^{2}
What will happen to time if the distance (x) is tripled?
x∝t^{2}
- If the distance traveled (x) increases by a factor of three (3), then time must increase by some factor that, when squared, will also equal a factor of three
- This number is the square root of three, so time will increase by a factor of 1.7—*
- or be 1.7 TIMES larger than it was originally*
√3=
1.7
Well-designed research must have a hypothesis that is….?
a Testable hypothesis
- basically, it can be used to verify a clear YES or NO answer
Types of Research
- Decribe Experimental or “Basic Science” Research
- What kind of environment is it conducted in?
- Is it conducted on Human subjects?
- What does this type of research allow investigators to have?
- The main goal of this type of research is to indicate…?
- What kind of environment is it conducted in?
Experimental or “Basic Science” Research
- Laboratory research conducted in a highly controlled environment;
- NOT on human subjects!
- This type of research allows investigators to have the strictest level of control over all possible variables and conditions
- Experimental research is therefore thought to be the most reliable way to indicate CAUSATION
Types of Research
- Human Subjects Research
- Where is it conducted?
- What is a common example of this type of research?
- How does it compare (in efficiency and level of control) to “Basic Science” research?
- Are its conclusions more or less definitive?
What are the 2 Types of Human Subjects Research?
Human Subjects Research
2 Types: Experimental & Observational
- Research conducted outside the laboratory, often on human subjects
- Drug trials are a common and familiar example
- Has less control over conditions than in “Basic Science” research*
- making conclusions less definitive*
- For example, a study drug may fail to show decreased blood pressure in some subjects
- However, it is later discovered that all participants did not follow the strict low-sodium diet required by the study guidelines
- This was thought to influence blood pressure measurements in those patients
- However, it is later discovered that all participants did not follow the strict low-sodium diet required by the study guidelines
Human Subjects Research
- Describe “Experimental” Research
- Research involves a specific ________ controlled by the ________.
- Subjects are separated into _____ and ________ groups
- Research involves a specific intervention controlled by the investigator
- Subjects are separated into control and treatment groups.
- e.g., To test the efficacy of a new drug, patients with allergies are separated into groups randomly and given either the drug (treatment) or placebo (control)
Human Subjects Research
- Describe “Observational” Research
- Investigator _______ data WITHOUT….? (2)
- Investigator observes data WITHOUT:
- Direct control over the variables, OR
- Implementation of interventions
Example:
An investigator reviews case studies from COPD [chronic obstructive pulmonary disease] patients and examines demographic information and lifestyle choices in an attempt to identify risk factors associated with COPD)
Medical Ethics
- Describe (in general) “Beneficence”
- One important aspect of this is ending a study because of…?
DO GOOD.
- Doctors and researchers have an obligation to promote the welfare of patients or study participants
- Patient welfare should always be a primary consideration in study design and execution
ENDING A STUDY BECAUSE OF POSITIVE (“GOOD”) RESULTS:
- One classical application of beneficence in human subjects research is the obligation to END an experimental study when it is clear a drug or intervention results in obvious benefit
- This may sound counterintuitive, but remember that each study must have control groups
- If a drug is found to save the lives of dying cancer patients…..
- it is NOT ethical to continue the study long-term and thereby save the lives of those in the treatment groups
- …while those in the control groups are withheld from taking a drug researchers know could help them
Medical Ethics
- Describe (in general) “Nonmaleficence”
- One important aspect of this is ending a study because of…?
DO NO HARM
- This is the physician’s oath, but applies equally to researchers
- Doctors and researchers have an obligation to not harm their patients or study participants
ENDING A STUDY BECAUSE OF NEGATIVE (“BAD”) RESULTS:
- Researchers are similarly obligated to end a research study as soon as it is verified that a treatment harms the subjects
- Most early research that is criticized today is drawn into question because it violated this principle
- The famous psychology study involving “Little Albert,” for example, while revealing evidence about conditioning, is now thought to have had an unethical impact on the young child involved :(
- Albert was conditioned to have severe generalized phobias of animals and the study involved obvious emotional trauma to Albert.
Medical Ethics
- Describe (in general) “Autonomy”
Patient autonomy and informed consent
- Physicians and researchers have an obligation to:
- Inform patients or study participants, and
- Allow them to make decisions about their own health and treatment
At times, some deception (e.g., placebo) is necessary to effect research
- However, this should be the minimal amount possible, the truth should be revealed as soon as possible, and deception in general should be approved by an Internal Review Board (IRB)
Medical Ethics
Describe (in general) “Justice”
- Hint: “Equal…”
Equal treatment of all people
- Equal allocation of resources, to the extent possible
- ..without bias, prejudice or discrimination.
Observational Research Study Types
- Describe a “Cohort Study”
- Cohort Studies usually employ _______s to demonstrate a ______
- Give an example
Hint:
A “cohort” is an ancient Roman military unit, comprising six centuries, equal to one tenth of a legion
THINK: Cohort = GROUP
Cohort Study
- is a longitudinal study observing characteristics (usually risk factors) of members of a cohort across time
Cohort studies usually employ correlations to demonstrate a relationship
EXAMPLES:
- Smokers (Cohort A) were three times more likely to develop lung cancer before the age of 50 than non-smokers (Cohort B)
- A statistically significant correlation (r = 0.84, p < 0.01) exists between socioeconomic status and frequency of pre-term births
Notice that this is NOT the manipulation of an independent variable, THEN observing outcomes
- Scientists are simply observing data about people or populations as it exists and looking for relationships
Observational Research Study Types
- Describe a “Cross-Sectional Study”
- Give some examples
Cross-Sectional study
- is the analysis of data collected from a population or sample AT ONE SPECIFIC TIME
- …compared to across a time period for cohort studies
EXAMPLES:
- A survey of the U.S. population to determine the current prevalence of a disease
- A study examines blood pressure among those with incomes above $100K/year and those below $100K.
Observational Research Study Types
- Describe a “Case-Control Study”
- By design, a Case-Control Study is always ________
- Give an example
Case-Control Study
By design, a Case-Control Study is always RETROSPECTIVE
- Is an observational study of individuals in the population WITH a condition present
- …and comparison of that group to a control group of persons WITHOUT the disease IN THAT SAME POPULATION (i.e., “reference group”)
Most famous Case-Control Study is the one is the study that linked lung cancer to smoking
EXAMPLE:
- There is a suspicion that zinc oxide (the white non-absorbent sunscreen traditionally worn by lifeguards) is more effective at preventing sunburns that lead to skin cancer than absorbent sunscreen lotions
- A case-control study was conducted to investigate if exposure to zinc oxide is a more effective skin cancer prevention measure
- The study involved comparing a group of former lifeguards that HAD developed cancer on their cheeks and noses (cases) to a group of lifeguards (same population) WITHOUT this type of cancer (controls)
- …and assess their prior exposure to:
- zinc oxide OR
- absorbent sunscreen lotions
- …and assess their prior exposure to:
- This study would be retrospective in that the former lifeguards would be asked to recall (“Recall Bias”) which type of sunscreen they used on their face and approximately how often
Observational Research Study Types
- What are some Pros & Cons of:
- Cross-Sectional Studies
Observational Research Study Types
- What are some Pros & Cons of:
- COHORT Studies
Observational Research Study Types
- What are some Pros & Cons of:
- Case-Control Studies
Also:
What kind of outcome are these studies useful for?
Useful for RARE outcomes
Observational Research Study Types
- Differentiate b/t:
Prospective & Retrospective Cohort Studies
- What’s going on wrt the Exposure & Outcome?
Independent vs. Dependent variables
- Differentiate b/t the two
- For each:
- “Also called the ____ variable”
- “It can be thought of as the….?”
- What axis of a graph does each go on?
- For each:
HINT: “DRYMIX” or “I’M a DR.”
INDEPENDENT VARIABLE
- The variable that gets:
- MANIPULATED (or rather…)
changed BY THE INVESTIGATOR!!
- Also called the “predictor variable.”
It can be thought of as the “CAUSE”
- Always goes on the X-axis
DEPENDENT VARIABLE
- The variable MEASURED
- ….as a response to changes in the INdependent variable
- Also called the “outcome variable”
It can be thought of as the “EFFECT”
- You can’t “Add” a Dependent Variable
- i.e., you can “add” an outcome…you can only “add” something that will change the outcome
- Always goes on the Y-axis
Independent vs. Dependent variables
- Identify the independent and dependent variables in the following scenarios:
- Time spent studying and test score
- Gas mileage and octane rating of the gas used
- Dosage of medication used and lab rat survival rate
- Level of aggression and amount of exposure to violent video games
- Independent: Time Spent studying
- Dependent: Test score
- Independent: Octane rating of gas
- Dependent: Gas mileage
- Independent: Dosage of medication
- Dependent: Survival Rate
- Independent: Exposure to violent VG’s
- Dependent: Lvl of aggression
Study Methods
- Define a “Control Group”
Control Group
A group or trial in which ALL conditions and environmental factors are IDENTICAL to the treatment group–
EXCEPT for the treatment itself!!
Control Groups
- Define a “POSITIVE Control”
- What purpose do they serve in an experimental study?
- An experiment results in the ability of bacteria to grow on a petri plate containing antibiotic (POSITIVE control=?)
POSITIVE control
- Is a group given a treatment with a KNOWN OUTCOME
PURPOSE:
- The positive control can be compared to the unknown outcome of whatever treatment is being studied
EXAMPLE:
- If your experiment results in the ability of bacteria to grow on a petri plate containing antibiotic, your positive control will be bacteria that are known to carry the appropriate drug resistance marker
Even if none of your experimental bacteria grow, as long as there is growth of the positive control
- ….you know that growth was at least POSSIBLE
Control Groups
- Define a “NEGATIVE Control”
- What purpose do they serve in an experimental study?
- HINT: What can they help expose?
- An experiment results in the ability of bacteria to grow on a petri plate containing antibiotic (NEGATIVE control=?)
- What purpose do they serve in an experimental study?
NEGATIVE control
- is a group that does NOT receive any condition or treatment
- …and for which NO outcome is expected
PURPOSE:
- To help expose CONFOUNDING variables
EXAMPLE:
- An experiment results in the ability of bacteria to grow on a petri plate containing antibiotic
NEGATIVE Control:
- Bacteria which do NOT carry a drug resistance marker should NOT be able to grow on a petri plate containing antibiotic
∴ if growth IS observed, it is a red flag that something is WRONG with the experiment!
- “What could be one reason for growth?”
- aka what could be a confounding variable here?
Evaluating Research
- Sources of Experimental Bias or Error
- Define “Selection Bias”
Selection Bias
The method used to select participants is
NOT TRULY RANDOM
- ∴
- the results are not representative of the whole population
- There are “systematic BASELINE differences” among participants
….because true randomization was not achieved
Evaluating Research
- Sources of Experimental Bias or Error
Define “Specific Real Area Bias”
- Give an example
- Will having a specific location for a study always mean this type of bias will exist?
- Give an example why/why not
- Will having a specific location for a study always mean this type of bias will exist?
SPECIFIC REAL AREA BIAS:
- is bias introduced by conducting the study in a specific area that does NOT include a representative sampling of the population being studied
For example:
- A study of U.S. eating habits (population= U.S. citizens) conducted at the gym leaves out all those who do not attend the gym—particularly bad because:
- being a member of a gym is likely to confound with eating habits
Having a specific location for a study does NOT always mean “specific real area bias” will exist!
For example:
- a study investigating the drug-using habits of Maricopa County High School students (population=HS students) would not be biased if conducted only at the high school
- because the sample would be representative of the population
Evaluating Research
- Sources of Experimental Bias or Error
Define “Self-Selection Bias”
- What type of research shows SIGNIFICANT self-selection bias?
- Give an example of SSB
Self-Selection bias
- is bias introduced when participants in the study have the ability to:
- choose to participate or not to participate, OR ,
- to determine their level of involvement
Surveys show significant self-selection bias
For example:
- Feedback boxes at fast-food establishments are far from representative of customer opinions
- Generally because those with negative opinions have been shown to be far more likely to fill out such surveys
- (as opposed to those with positive experiences)
- Generally because those with negative opinions have been shown to be far more likely to fill out such surveys
Evaluating Research
- Sources of Experimental Bias or Error
Define “Pre-Screening,” or “Advertising Bias”
- Give an example wrt advertisements looking for volunteers for a study being done on obesity
Pre-Screening or Advertising Bias:
- Occurs when the screening or advertising *PROCESS* *ITSELF* results in
- …an unrepresentative sample
For example:
- Advertisements asking for volunteers FOR THE SAME STUDY that are worded:
- “Volunteers needed for an obesity study”
OR
- “Volunteers needed for a weightloss study”
…will likely elicit different volunteers
(Obesity=BAD! Weightloss=GOOD!)
Evaluating Research
- Sources of Experimental Bias or Error
Define “Exclusion Bias”
- Give an example
Exclusion Bias
- Refers to the exclusion of an entire group from the population
Example:
- A study about childhood education that does not survey any homeschooled children
Evaluating Research
- Sources of Experimental Bias or Error
Define “Healthy User Bias”
- Give an example
Healthy User Bias
- When the persons included in the study are likely to be healthier
- ….than the general population
Example:
- Studying cardiac atherosclerosis among participants in a triathalon
Evaluating Research
- Sources of Experimental Bias or Error
Define “Berkson’s Fallacy”
- How does it compare to Healthy User Bias?
Berkson’s Fallacy
- Is the selection of participants from hospitals–
- where the participants are likely to be LESS healthy than the general population
Is the OPPOSITE EFFECT of the Healthy-User Bias
Evaluating Research
- Sources of Experimental Bias or Error
Define “Overmatching”
- What happens if a Confounding variable is matched for?
- What kind of study can Overmatching be FATAL in?
OVERmatching
Matching on variables other than those that are risk factors for the disease under study
- Is a negative outcome resulting from what is normally a good practice:
- matching for potentially confounding variables*
- i.e., age-matched
EXAMPLE:
- If you use neighborhood controls in a study on NUTRITION and TUBERCULOSIS…
- You are inadvertedly matching for SES (and thus, nutrition)*
Evaluating Research
- Sources of Experimental Bias or Error
Define “Observer Bias”
- Give an example using the Foot Study you’re in
OBSERVER BIAS:
Observers (or researchers) KNOW THE GOALS of the study or the hypotheses
- …and allow this knowledge to INFLUENCE their observations during the study
Foot Study Example:
- We want statistically significant data, so we “observe” things in a way that will help us get statistically significant data
Evaluating Research
- Sources of Experimental Bias or Error
Define “Demand Characteristics”
- Give an example using the Foot Study you’re in
Demand Characteristics
- PARTICIPANTS form an interpretation of the experiment’s purpose
- …and unconsciously change their behavior*
- to FIT that interpretation*
Example:
- During doming trials, we remark how subject’s numbers were the “highest we’ve seen”
- This makes them think we WANT to see higher numbers
- They then strive to get the highest numbers possible
(In reality, we want NATURAL doming strengths)
Evaluating Research
- Sources of Experimental Bias or Error
Define “Information Bias”
- This definition is split up into 2 different definitions–
- depending on whether you’re dealing with Continuous or Categorical Variables
- What is it known by with each variable?
- depending on whether you’re dealing with Continuous or Categorical Variables
- Give an example using the Foot Study you’re in
- For both Continuous AND categorical variables
Information Bias
=Wrong or inexact recording of variables or data
- With CONTINUOUS Variables (such as blood pressure):
- this is referred to as “Measurement Error”
- With CATEGORICAL Variables (such as tumor stage):
- this is known as “Misclassification”
Example:
- Continuous (measurements)
- recording them wrong
- Categorical (shoe size on form)
- you put down the wrong shoe size for them
Evaluating Research
- Sources of Experimental Bias or Error
Define a “Confounding Variable”
- Give a hypothetical example of a confounding variable (using Variable A,B, and C where C is the confounding variable)
- What is the “Placebo Effect?”
Confounding Variables (a.k.a. confounding factor, confounder)
- An extraneous variable that INFLUENCES the variables being studied
- ….but is NOT part of the _expected_ correlation or causal pathway being investigated
EXAMPLE:
- Suppose the relationship between A and B is being investigated
- A is hypothesized to increase B
- Experimental results support this hypothesis
- A is hypothesized to increase B
- It is later discovered that a third variable (i.e., extraneous variable) influences both A and B,
- Decreasing A, and
- Increasing B
- In fact, it was C that created the observed result in B that was originally attributed to A
Variable C is a confounding variable
Placebo Effect:
- One example of a confounding variable
- The placebo effect occurs when participants given a placebo (i.e., sham treatment) during a study experience REAL or PERCEIVED health benefits–
- due to their belief that they ARE being treated
Evaluating Research
- Sources of Experimental Bias or Error
Define a “Detection Bias”
- Give an example
Detection Bias
- Systematic differences between groups…
- …caused by inconsistency in*
- the method of detection or diagnosis*
EXAMPLE:
- A study reports that inner city children suffer from ADHD at twice the rate of suburban children
It is later discovered that systematic differences existed between:
- diagnostic tools, and
- training available
- …at inner city hospitals included in the study
- …vs. suburban hospitals included in the study
- Basically, they’re detecting ADHD rates inconsistently, because suburban hospitals are able to more accurately diagnose those who REALLY have ADHD
- Inner city hospitals are quicker to just “slap a label” on ‘em
Evaluating Research
Sources of Experimental Bias or Error
Define “Performance Bias”
- Give an example of this in a hospital scenario
Performance Bias
- Systematic differences between groups…
- in terms of the*
ACTUAL CARE OR TREATMENT PROVIDED
EXAMPLE:
- A physician unconsciously pays closer attention to and conducts more followup with patients the doctor knows to be enrolled in a heart study
This results in differences in care for those individuals not accounted for in the study
Evaluating Research
- Sources of Experimental Bias or Error
Define “Experimenter Bias,” a.k.a., “_______ Bias”
- What are the 2 types of Experimenter Bias?
Apart from those 2 types, what else could be considered “Experimenter Bias?”
Experimenter Bias (a.k.a., Researcher Bias)
- ERRORS introduced into a study
due to the expectations of the investigator
- 2 Types:
- Confirmation Bias
- Reporting Bias
Experimenter Bias also can include unconscious communication of expected results to the participants
- …thereby influencing their behavior
Evaluating Research
- Sources of Experimental Bias or Error
Experimenter Bias (2 types)
- Describe “Confirmation Bias”
Confirmation Bias:
- The tendency to favor information that confirms one’s hypothesis or preconceived notions
- and to dismiss information that discredits them
(This sort of thing also happens psychologically)
Evaluating Research
- Sources of Experimental Bias or Error
Experimenter Bias (2 types)
- Describe “Reporting Bias”
Reporting Bias:
- Systematic differences resulting from some findings being reported and other findings NOT being reported
- Investigators may withhold or ignore data that does not support their hypothesis (if conscious and intentional this would be a clear violation of ethics)
Statistically-significant results are usually far more likely to be reported than are statistically insignificant results
…although BOTH are important to
an unbiased determination
Research Design & Execution
- Measurement
Compare Accuracy vs. Precision
Accuracy
- is a measure of the degree to which a value represents the true or correct value
Precision
- is a measure of the degree to which repeatedly measured values show the same or reproducible values
- The greater the range or scatter of measurements, the less precision
Research Design & Execution
- Measurement
Compare “Reliability”with“Validity”
- What results in LOW internal Validity?
RELIABILITY
- Results are consistent and repeatable
VALIDITY
- The test or experiment measures what it purports to measure
- …and uses methods that meet scientific standards
Failure to adhere to the standards of the scientific method or other accepted experimental best practices results in LOW internal validity
Research Design & Execution
- Measurement
Reliability
- Compare “Test-Retest” reliability with “Inter-Rater” reliability
In both cases, a high value of reliability is around _._ and shows good _____ ______
“Test-Retest” reliability
- is a measure of the degree of consistency between one administration of a test and a subsequent administration of that same test (i.e. the “retest”)
“Inter-Rater” reliability
- is a measure of consistency between multiple raters or evaluators
that are assigning the same values or making
the same observations
In both cases, a high value of reliability is around 1.0 and shows good internal validity
DECREASING“x” by a factor of 4 is the same as multiplying the variable (x) by?
1/4x
Pv=nRT
- If the volume goes down by 80%, it has lost ___ of its value—
- which is the same as being multiplied by___
- or going down by a factor of _
- If the volume goes down by 80%, it has lost 4/5 of its value—
- which is the same as being multiplied by 1/5
- or going down by a factor of 5
To say that something is 225% AS dense as something else is to say that it is ___% MORE dense
125%
Same as multiplying it by 5/4
Correlation vs. Causation
- Linear Regression Analysis
Correlation Coefficient, r^{2}
- Describe how the correlation coefficient relates to SCATTER PLOTS, aka Linear Regression Analysis (shown below)
- If the data points are CLOSER, as a whole, to the TREND LINE:
- Will you have a higher/lower r^{2} value?
- Correlation coefficients relate to linear regression analysis
- A large number of trials
- involving changes in the independent variable
- are graphed against the dependent variable
resulting in a SCATTER PLOT
- A “Least Squares” or “Best Fit” line is drawn
- that best approximates the trend of the data points
_The correlation coefficient, r^{2} is a measure of how tightly the DATA fit to THIS line!_
The CLOSER the data are, as a whole, to the trend line:
- the HIGHER the r^{2} value
Correlation vs. Causation
- Linear Regression Analysis
Correlation Coefficient, r^{2}
- Correlation coefficients vary from __ to __
- an r^{2} value of ___ would be a PERFECT correlation
- The correlation coefficient is said to explain the amount of variance in __ ACCOUNTED FOR by __
Correlation Coefficients vary from 0 to 1
- An r^{2} value of 1.0 would be a PERFECT correlation!
- The correlation coefficient is said to explain the amount of variance in y ACCOUNTED FOR by x*
Research Methods
- Describe HILL’S CRITERIA
- What are they used to EVALUATE in statistics?
- List the 9 Criteria
- Which is considered to be the MOST IMPORTANT of the 9?
HINT: The Crazy DRagon Caleb Poops The Silly String Always!
Hill’s Criteria is a SET OF GUIDELINES used to evaluate whether or not a CAUSAL RELATIONSHIP exists
- Temporality
- Consistency
- Dose-Reponse Relationship
- Coherence
- Plausibility
- Testable by Experiment
- Specificity
- Strength
- Analogy
Research Methods
HILL’S CRITERIA
- Describe “Temporality”
Temporality:
In time, the exposure or treatment
MUST PRECEDE THE OUTCOME
- If A causes B, A must occur in time prior to B
This is often considered the most important Hill Criteria
- If the outcome is found to occur before the exposure in even a few cases, the potential of causality is generally REJECTED
Research Methods
HILL’S CRITERIA
- Describe “Strength”
- A larger ___ and _ -value, or a smaller _ -value
…. would all SUPPORT causation
Strength
- The MAGNITUDE of the CORRELATION or association
- including statistical measures of significance
A larger correlation, larger r^{2} value, or a smaller p value would all support causation
Research Methods
HILL’S CRITERIA
- Describe “Consistency”
- “The correlative relationship is __________”
Consistency:
The correlation or association:
- continues across MULTIPLE trials, across time, or when the study is REPLICATED by others*
- In other words, the relationship is REPRODUCIBLE*
Research Methods
HILL’S CRITERIA
- Describe “Specificity”
Specificity:
One of the weakest Hill Criteria
- High specificity means the potential cause results in only ONE specific effect
- If present, it can provide additional support for causation–
- but the lack of specificity is common in MANY established cause-effect relationships
Research Methods
HILL’S CRITERIA
- Describe “Plausibility”
Plausibility:
The association fits LOGICALLY within our current understanding of how a process works
The association has a logical theoretical basis
Research Methods
HILL’S CRITERIA
- Describe “Dose-Response Relationship”
- The observed _____ is proportional TO?
Does meeting this criteria determine CAUSATION?
Dose-Response Relationship:
The observed response is proportional to:
the dosage or degree of the exposure or treatment
- This is considered STRONG evidence of causation*
- …but is NOT necessary*
Research Methods
HILL’S CRITERIA
- Describe “Testable by Experiment”
- What IS it that needs to be “testable?”
Testable by Experiment:
The ASSOCIATION can be:
REPRODUCED VIA EXPERIMENT
Research Methods
HILL’S CRITERIA
- Describe “Coherence”
- Association is compatible with…?
Coherence:
The ASSOCIATION is COMPATIBLE with:
existing or previously-established science
Research Methods
HILL’S CRITERIA
- Describe “Analogy”
SIMILAR associations are shown
or known to exist that are:
analogous to the
association being considered
Research Methods
- Define “INTERNAL Validity”
- “The extent to which you are able to say that…?”*
- What is internal validity a function of wrt the study itself?
Internal Validity
The extent to which you are able to say that:
No OTHER variables (except the one you’re studying) caused the result
…which is a function of the scientific RIGOR of the study
Research Methods
- Differentiate b/t a CONDITION that is “Necessary” vs one that is merely “Sufficient”
Necessary:
A condition that MUST be satisfied
in order for an event to occur
Sufficient:
A condition that, if satisfied,
GUARANTEES that an event WILL occur!
Research Methods
- Define “EXTERNAL Validity”
- External Validity= ___________bility*
- What does external validity DEPEND heavily on?
- Give an example of a study with POOR external validity
External Validity = GENERALIZABILITY!
=The degree to which the findings can be
extrapolated to the general population
This depends a lot on:
the SUBJECTS being tested– are the representative of the general population?
Example:
- Sarah is a psychologist who teaches and does research at an expensive, private college.She’s interested in studying whether offering specific praise after a task will boost people’s self-esteem. If her hypothesis is correct, then giving someone a specific compliment on a job well done after a task will make them feel better about themselves. And if she can show that specific praise post-task boosts self-esteem, then managers at companies everywhere will be able to boost their employees’ self-esteem by offering them specific praise.*
- But here’s a problem: the volunteers that Sarah gets for her study are all college students, most of them are white, and most of them are from privileged backgrounds*
Sarah worries that her results might not be applicable to people who are NOT in their late teens or early 20s, white, and rich
Research Methods
- Define a SINGLE-BLIND test
- What all information gets witheld (3), and from whom?
Information about:
- The study itself
- Who is in the…
- Control groups
- Treatment groups
- other potentially biasing details
are concealed from:
the person DOING the assessment
Research Methods
- Define a DOUBLE-BLIND test
- What all information gets witheld (3), and from whom?
Information about:
- The study itself
- Who is in the…
- Control groups
- Treatment groups
- other potentially biasing details
are concealed from BOTH:
- The SUBJECT
- The person DOING the assessment (researcher)
*
Research Methods
- Type I vs Type II Errors
Type I Error: REJECTED H0** but **SHOULDN’T HAVE
CLAIMED difference between groups
…when NONE existed
Type II ErrorSHOULD** **HAVE** rejected H0 but **DIDN’T
Did NOT claim a difference between groups
…when one DID exist
Research Methods
- Define the NULL HYPOTHESIS, H_{0}
- What is the Null hypothesis when testing for:
- Group differences
- Correlation or Causation
- What is the Null hypothesis when testing for:
The Null Hypothesis:
H_{0} is always the LACK OF A RELATIONSHIP
OR GROUP DIFFERENCE
In testing for GROUP DIFFERENCES:
- H_{<strong>0</strong> }= there are NO statistically significant differences between groups
When testing for CORRELATION or CAUSATION:
H_{0}= there is NO RELATIONSHIP
Research Methods
- Define the ALTERNATIVE HYPOTHESIS
- What is the Alt. Hypothesis when testing for:
- Group differences
- Correlation or Causation
- What is the Alt. Hypothesis when testing for:
The Alternative Hypothesis:
The PRESENCE OF A RELATIONSHIP OR GROUP DIFFERENCE
The OPPOSITE of the H_{0}!
In testing for group differences:
- Alt. Hypothesis=there IS a difference between groups
When testing for correlation or causation:
- Alt. Hypothesis= there IS a relationship.
Statistics
- Compare & Contrast “Sample” vs. “Population”
Sample =
- that portion of a population included in the data
- [e.g., U.S. veterans called in the phone survey]
SMALLER than the POPULATION
which is ALL the members in the group/category being sampled
[e.g., ALL U.S. veterans]
Statistics
- Define a STATISTIC
- A statistic is a measure or data point*
- (such as mean, median, etc.)*
- that is calculated for the sample*
[e.g., average income among U.S. veterans in a sample]
Statistics
- Define “Parameter”
- How does it compare with a Statistic?
Parameter =
a measure (such as mean, median, etc.) that is calculated for the ENTIRE population,
NOT merely the sample of the population
[e.g., average income among ALL U.S. veterans]
Statistics
- Compare Mean, Median, & Mode
- Also, define Range
MEAN=
- the average value
MEDIAN =
- the middle of the data
MODE=
- the most frequently occurring data point
RANGE=
- the minimum data point ⇒maximum data point
Statistics
- Define STANDARD DEVIATION (SD)
- What does a SMALL SD indicate?
- What does a LARGE SD indicate?
STANDARD DEVIATION (SD) =
- how tightly associated the data are TO THE MEAN
A small SD indicates:
a narrow set of data for which most values are close to the mean
A large SD indicates:
a greater spread, or wider distribution,
of the data around the mean
Standard Deviation (SD)
- Define “Normal Distribution”
- 1 SD=__% of the population
- 2 SD=__% of the population
- 3 SD=__% of the population
Normal Distribution=
Hypothetical perfect BELL-SHAPED curve f_or which the following is true:_
- 1 SD = 68% of the population
- 2 SD = 95% of the population
- 3 SD = 99% of the population
Research Design
When it comes to Probability:
- The ASSUMPTION is that the OUTCOMES are _____ and _____ ________
Assumption = Outcomes are:
- INDEPENDENT
- Do NOT influence one another
- MUTUALLY EXCLUSIVE
- They CANNOT occur together
Probability
- “And” vs. “Or”
- To solve for each:
- “______ the probabilities of individual events to get the overall probability of ____ events occurring”
- To solve for each:
AND vs. OR
AND =
- MULTIPLY the probabilities of individual events to get the overall probability of BOTH events occurring
OR =
- ADD the probabilities of each individual event together to get the overall probability of EITHER event occurring
Hypothesis Testing
- t-test or z-test
- What is the t-test or z-test (once calculated) COMPARED to?
- What does this comparison give us?
- What is the t-test or z-test (once calculated) COMPARED to?
t-Test or z-Test ⇒p value
The “test statistic” (either a t-value or a z-value) is calculated
- This result is COMPARED to:
- a TABLE of t-values or z-values
The table gives us:
the SIGNIFICANCE LEVEL, “α,” associated with that test statistic
Hypothesis Testing
- t-test or z-test
- For a z-test,“n” (sample size) must be GREATER THAN what number?
- Once this criteria is met, what can we assume about the distribution?
- When would we have to use a t-test?
- What does THIS mean wrt its distribution?
- For a z-test,“n” (sample size) must be GREATER THAN what number?
z-test
- For a z-test, n>30
- This will mean it has a NORMAL (BELL-SHAPED) DISTRIBUTION
- We’d use a t-test when our sample size is SMALL
- This will mean our graph will NOT be normally distributed (only approximately)
Hypothesis Testing
Significance Level, “α”
- What are the 3 α values we’ll see?
When comparing p and α values:
- what does a p value > .005 indicate wrt level of confidence?
Significance Level (α)
- α = 0.05, 0.01, 0.001*
- p < 0.05 means we can be 95% confident that the results are actual/real
- …RATHER than the result of random chance*
- p < 0.01 means we can be 99% confident, and so forth