Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
121 Cards in this Set
- Front
- Back
what are the 4 components that go into evidence based practice? |
clinical expertise
patient values and preferences
best research evidence
clinical circumstances and setting |
|
what does PICO stand for? |
patient, intervention, comparison, outcomes |
|
what are the four sources of best available evidence for EBP? |
tradition (lowest), authority, trail and error, scientific method/applicaiton of logical reasoning (highest) |
|
source of best available evidence for EBP that says we've always used __ at our facility and have experiences good outcomes |
tradition |
|
source of best available evidence for EBP that Dr. __ states that this leads to the best outcomes |
authority |
|
source of best available evidence for EBP that trys a lot of different measures to see if they work |
trial and error |
|
source of best available evidence for EBP that uses the systematic thought process of logical reasoning to answer questions and aqurie new knowledge |
scientific method |
|
what are the 5 steps to research? |
identify the question
design tehs tudy
methods (carry out the design)
analyze and interpret the data
communicate results (leads back to identifying the question) |
|
describe basic vs applied research |
basic is used to gain knowledge for its own sake (may lead to applied research questions)
applied is used to solve a clinical problem |
|
what is the important of translation research? |
so basic and applied science communicate as well as applied and oragnizational/health systems. |
|
explain experimental vs non-experimental research |
experiemental looks for cause and effect relationships (manipulates and controls variables to see effects on outcomes)
nonexperimental uses observation of variables to describe things |
|
describe quantatative vs qualitative research |
quantitative: uses objective, numerical data
qualitative: helps you know what the data means |
|
describe the research continuum |
descriptive research (describes populations) leads to exploratory research (finds relationships) which leads to experimental research (cause and effect) |
|
what is the difference between primary and secondary literature? |
primary: the actual original research article
secondary: summary article, meta-analysis |
|
oversee all research involving human subjects |
intstituational review board (IRB) |
|
what are the two governing bodies control IRB's |
FDA, DHHS |
|
Which of these are considered "human subjects"?
A: collection of data or specimins that have identifiable information that could be linked back to a specific person
B: collection of data or specimens from individuals that are no longer living
C: collection of data or specimens through interaction or intervention of living animals |
A and C are human subjects |
|
which of these would be considered human subject research?
A: research consisting of a clinical trial where the results would be published in teh NEJM
B: planning to hand out evaluations at the end of this session to assess the strengths and weaknesses of my presentation
C: do a research project that i will present to my psychology 101 class only
D: assess parental opinions about breastfeeding which I will present at a national conference |
A and D |
|
systematic inversigation designed to develop to contribute to generalizable knowledge |
research |
|
what is the definition of a human subject? |
a living individual about whom an investigator conducting research obtains data thorugh:
intervention or interaction with an individual collection of private health information |
|
nazi experimentation, research in US prisions, and the tuskegee experiments all showed a need for __ |
experimental (human) ethics |
|
what are the three basic principles of the belmont report? |
human research must have: respect for persons (autonomy) beneficence (do good) justice (utility) |
|
what does it mean for a study to have autonomy? |
consent process
treated as autonomous agents
diminished autonomy requires increased protectoin |
|
when you gather systematic and comprehensive information about proposed research, maximize possible benefits an minimize risks |
beneficience |
|
the probability to harm and the severity (magnitude) or harm |
risk |
|
what does it mean if a subject has "justice" |
individual (fairness in exclusion and inclusion)
social (class burden) |
|
when you weight the risks vs the benefits of the study you are incorporating __ |
justice |
|
what study looked at whether people would follow authority figures to the point of inflicting pain on people |
milgrams study |
|
the probabilty and magnitude of harm or discomfort anticipated in the reserach are not greater in and of themselves than those ordinarily encoutered in daily life or during the performance of routine physical or psychological examinations or test |
minimal risks |
|
what are three reasons we use animals in reserach? |
allows for use of true 'experimental design'
human safety
cant replicate complex responses that occur with intact multisystem organisms |
|
characterized by controlled manipulation of variables by teh reserachers
reserachers manipulate and control one or more variables and observe the resultant variation in other varables
prmary purpose is to suggest 'cause and effect' relatinships between variables |
experimental research |
|
considered the 'gold standard' of resreach providing the greatest degree of confidence in validity of outcome |
experimental design |
|
what are the three essential components of experimental research? |
randomization, control/intervention groups, mainpulation |
|
US public health service, USDA, OLAW, AAALAC, AVMA, and IACUC all oversee what kind of research? |
animal research |
|
equivalent to an IRB for animal research |
institutional animal care and use committee (IACUC) |
|
what are the three R's of the IACUC? |
reduction (reduce teh number of animals used in each study)
replacement (replace lab animals with non-animal models whenever possible)
refinement (refine the tests to ensure the most comfortable and humane conditions possible |
|
the process of assigning a number or category to a variable to represent quantities of qualities of characteristics |
measurement |
|
the property or characteristic that distinguishes among members of a group |
variable |
|
a kind of variable with any value along a continum
(distance walked, size of a wound) |
continuous |
|
a kind of varaible that uses whole units only, can also be discrete qualitative categoreis
e.g. number of exercise reps, blood type |
discrete |
|
kind of variable that has 2 possible values
e.g. answer to yes/no question |
dichotomous |
|
BMI and TEMP are examples of what kind of variable? |
continuous |
|
MMT grade, diagnosis, and HR are all examples of __ variables |
discrete |
|
what is the hierarchy of levels of measurement? |
nominal<ordinal<interval<ratio |
|
kind of level of measurement where there is no order, it is purely cartegorical. examples: gender, type of arthritis. qualitative, indentifies differences, mutually exclusive |
nominal |
|
kind of level of measurement that is cateogorical but with an order. it is qualitative and can indentify differences in direction and difference. it cannot tell the magnitude of the difference (numbers don't repreent quantities, intervals aren't necessarily equal) |
ordinal |
|
MMT grades, level of assistance, or stages of cancer would represent what kind of measurement? |
ordinal |
|
kind of level of measurement that is quantitative, has equal intervals between values, has a zero point which does not mean an absence of a variable, and can negative values. |
interval |
|
temperature is an example of what kind of level of measurement? |
interval |
|
kind of level of measurement that is quantitative, has equal intervals between values, has a difference between values that has meaning, has a true zero that means the absence of a variable, negative numbers aren't possible, and can do any mathematical operation |
ratio |
|
ROM measurments, HR, and time to complete activcity are all examples of what kind of level of measurement |
raito |
|
__ and __ levels of measurements are more sensitive to change |
interval and ratio |
|
what are three kinds of descriptive statistics |
frequency distributions
measures of central tenedency
measures of variability |
|
a table or graph of rank ordered scores taht shows the number of times each score appears in a data set |
frequqncy distributions |
|
what 3 things do frequency distributions make easier to interpret |
low and high scores
which scores occurred most often
where scores cluster (shape) |
|
a quantative summary of a group of score that reflects the 'center' of the distrubution |
measures of central tendency |
|
the arithmetic average |
mean |
|
the mean can be biased by __ |
extreme scores |
|
if data is noramally distributed, the __ is approximately the same as the __ |
median, mean |
|
the middle score |
median |
|
the __ is unaffected by extreme scores and good for ordinal data |
median |
|
the measure of central tendency which is the score that occurs most often, is useful for categorical data, can have multiple modes in a distrubution. it is best for ordinal data |
mode |
|
what three things should you consider when choosing a measure of central tendency? |
distribution shape, application of data, level of measurement |
|
characterize dispersion of scores in a data set |
measures of variabilty |
|
which measureso f variability express the same units of measuremnet as teh rest of the data? |
standard deviation and range |
|
__ is teh square of standard devation and is in squared units |
varience |
|
the difference between the lowest and highest scores. reported as single value or as low and high values, affected by extreme score, is not effective when comparing samples of differnt sizes |
range |
|
why is range not effective when comparing samples of different sizes? |
if you ahve a large set of data, you are more likely to find an outlier |
|
measure of variablity where distribution is ivded into 100 parts or 4 equal parts. this describes a single scores position relative to other socres |
percentile/quartiles |
|
a type of graph used to show data sets when data is skewed. the R side has teh dependetn variable, the center of teh box is teh media, teh top bar is teh 75th percentile, the bottom bar is the 25th percentile, whiskers represent 10th and 90th percentile and dots are outliers |
box plots |
|
affected by every score in teh distribution, able to compare variabilty of samples of different sizes, based on how different each score is from the mean |
variance |
|
the valueof teh variance will be __ if teh scores are similar and __ if the scors are dissimilar |
small, large |
|
brings variance back to original units of measurement, rather than squared units. gives the average deviation on either side of the mean |
standard devation |
|
given in independent units to express standard deviation as a proportion of teh mean. allows comparison of distributions with different units |
coefficient of variation |
|
is the standard deviation divided by the mean |
coefficient of variation |
|
what is the difference between descriptive stats and inferential stats? |
descrpitve: cannot infer anything beyond data sample
inferential: estimates population characteristics based on sample data (answers questions about comparisons/relationships |
|
people realistically in your study |
accessible population |
|
everyone who could be in your study |
estimate population |
|
what determines critical value |
alpha, decrees of freedom, 2vs1 sided test |
|
what are the two big keys to estimating population characteristics from our sample data? |
probability and smapling error |
|
what is probabilty? |
# of times an outcome can occur/total number of outcomes |
|
Probability is predictive of what __ happen in the long run, not what __ happen |
should, will |
|
what is a graphical representation of a normal distribution? |
symmetrical bell curve |
|
what are the levels of standard devation? |
0-1=34.13% 1-2=13.59% 2-3=2.14% 3-on=.13% |
|
both standard deviation and standard error are measures of __ |
variaiblity |
|
degrees of freedom are based on __ |
sample size |
|
an estimate of the varability of the sample (individual observations) |
standard deviation |
|
an estimate of the varability of the sample stastic (if repeated sampling was done) |
standard error |
|
reflects how well our sample represents the population |
standard error |
|
what two things will decrease standard error? |
decreasing standard devation or increasing sample size |
|
what does standard error equal? |
s(standard deviation)/square root of n (sample size) |
|
provides a range of values that are likely to include the true population mean |
confidence interval |
|
what does confidence interval equal? |
mean +/-confidence level X standard error |
|
the hypothesis of "no effect", means there is no difference between groups |
null hypothesis |
|
the hypothesis stating that there is a difference between groups |
alternative hypothesis |
|
what are the two directions of alternative hypothesis's? |
non-directional: there is a difference
directional: there is a difference and its in this way |
|
the probability of data having arsen by chance when the null is actually true |
P |
|
a __ P gives evidence that effect is not zero (evidence for alternative) |
small |
|
a __ p gives evidence that effect is zero (null) |
large |
|
how do you decide what is a large or small p value? |
level of significance |
|
compares calculated p values with alpha to make decisions during hypothesis testing. usually .05 or .01 |
level of significance (alpha) |
|
what is the difference between a two sided nondirectional test vs a one sided directional test? |
two sided nondirectional: results can occur equally often in either direction
one sided directional: used when a real difference can only occur in one direction |
|
you measure blood pressure in 50 women and 50 men. you want to determine if there is a difference in BP between genders. whould you do a one sided or two sided test? |
two sided |
|
you measure blood pressure in 50 women and 50 men. you want to determine if there is a difference in BP between genders. the statement "there is no diffrence between BP between teh wto genders is an example of __ |
the null hypothesis |
|
you measure blood pressure in 50 women and 50 men. you want to determine if there is a difference in BP between genders. the p=.025 and teh alpha=.01. should you reject the null hypothesis |
reject |
|
when the null is rejected but it is actually true, false positive. |
type 1 error |
|
when the null is not rejected but is false, false negative |
type 2 error |
|
what is alpha equal to? |
the probability of type 1 error |
|
what is beta equal to? |
the probabilty of type II error |
|
what does power equal? |
1-B, the probability of rejecting null when it is really false |
|
this kind of error can lead to using treatments that aren't really effective, treating populations that don't need it |
type 1 error |
|
this kind of error can lead to abandonment of effective treatment, not treating a population who needs it |
type II error |
|
1: In reality, there is no benefit to manual therapy. 2: In reality, there is a benefit to manual therapy. 3: if they committed an error, they must not know how to conduct research. Researchers conduct a study and conclude that the use of manual therapy is effective in reducing pain in persons with lateral epicondylitis. The null hypothesis is: Pain is not reduced with the use of manual therapy. |
1 |
|
1: In reality, there is no benefit to manual therapy. 2: In reality, there is a benefit to manual therapy. 3: The researchers still have no clue how to conduct research. Suppose the same researchers reached a different conclusion: Manual therapy does NOT reduce pain in persons with lateral epicondylitis. The null hypothesis remains: Pain is not reduced with the use of manual therapy. |
2 |
|
Assuming α = 0.05, what if p is close to .05 but slightly higher? Whatever difference in value that exists is likely due to chance (the difference is not real)....OR.... The test was not powerful enough to detect a difference. |
2 |
|
The complement of a Type II error The more powerful a statistical test, the less likely we are to make a Type II error and miss a significant result Probability of finding a significant difference when it really exists. Probability of rejecting Ho when it is really false |
stastiical power |
|
what 4 things affect power? |
effect size, variance, alpha, sample size |
|
what is effect size? |
how big an effect does an intervention have.
how far apart are the means |
|
how does alpha relate to power? |
teh larger the type I error rate and teh samller the type II error rate.
as alpha increases, beta decreases, power increases |
|
how does variance affect power? |
the lower the varaince, the more power |
|
how does sample size effect power? |
the larger your sample size, the greater the power |