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;
84 Cards in this Set
- Front
- Back
Inter-rater reliability
|
variation between 2 or more raters who measure the same group of subjects
|
|
Intra-rater reliability
|
stability of data recorded by one rater across 2 or more trials
|
|
Variance
|
Differences among scores
|
|
Reliability Coefficient
|
True Score/True Score + Error Variance = RC
* 1 = best RC * 0 = worst RC |
|
Classic Reliability Theory
|
Single score made up of true score + random error gives the best estimate of actual value.
|
|
Generalizable Theory
|
Not all variations from trial to trial should be attributed only to random error. Single score made up of true score + various types of error. Must consider specific measure when considering reliability of measure.
|
|
Regression toward mean
|
Single tests potentially extreme (high or low) score, multiple tests reveal score closer to group average
|
|
Systematic Measurement Error
|
form of measurement error where error is consistent across trials.
|
|
Random Measurement Error
|
Change that causes unpredictable measurements from trial to trail
|
|
test-retest reliability
|
no rater, changes based on re-test.
|
|
Validity
|
accuracy, measuring what you intended to measure, coming up with the right answer, average values near target. Minimizes systematic error.
applies to measure and study |
|
Snowball Sampling
|
TYPE OF NON PROBABILITY(CONVENIENCE) SAMPLING
Have current subjects "tell a friend" to join study |
|
Purposive Sampling
|
TYPE OF NON PROBABILITY (CONVENIENCE) SAMPLING
Handpicking based on subective judgement |
|
Quota Sampling
|
TYPE OF NON PROBABILITY (CONVENIENCE) SAMPLING
Consecutive sampling - get specific number of subjects with various characteristics. • 1st 50 pt's with left TKR • 1st 50 pt's with right TKR |
|
Consecutive Sampling
|
TYPE OF NON PROBABILITY (CONVENIENCE) SAMPLING
Recruiting all subjects that meet study requirements as they become available • 1st 100 pt's with TKR |
|
Cluster Sampling
|
TYPE OF PROBABILITY SAMPLING
For a population too large to make complete sampling list. Initial selection must elicit equal chance of each person in population to be selected. |
|
Disproportionate Sampling
|
TYPE OF PROBABILITY SAMPLING
example: - 100 female PT students have average height 64" - 100 male PT students have average heigh 68" 75% of PT students are female so the average height of pt students with respect to gender is: (.75)64 + (.25)68 = 65" |
|
Stratified Random Sampling
|
TYPE OF PROBABILITY SAMPLING
use to get proportional representation of males to females |
|
Systematic Sampling
|
TYPE OF PROBABILITY SAMPLING
Divide total # of elements in accessible population by the number of elements to be selected. Randomly select a starting point, select each element at the sampling interval. ex: 100 = pop, need 10 subjects. Randomly select number between 1 & 10. Say we picked 7. Select the 7th element, then select every 10th element from #7 (10 = sampling interval) |
|
Simple Random Sampling
|
TYPE OF PROBABILITY SAMPLING
Randomly select subjects using random number generation |
|
Probability Sampling
|
Every member of population has an equal chance of being selected for study
|
|
Target Population
|
entire population with characteristics of interest
|
|
Accessible Population
|
portion of target population that has a chance of being selected for study participation
|
|
Sample
|
subgroup of population
|
|
Generalization
|
what is true for the sample is true for the population
|
|
Sampling Bias
|
sample characteristics do not equal target population characteristics. Generalization of study conclusion from sample to population may be incorrect.
|
|
Population
|
group that meets criteria
|
|
Inclusion Criteria
|
used to identify target population
|
|
Exclusion Criteria
|
exclude based on ethical reasons, finances, inadequate data, or potential to skew data.
|
|
PICO
|
PROBLEM - PT Diagnosis
INTERVENTION - treatment COMPARISON OUTCOME - (desired result) • represents the components of a good clinical question |
|
Probability Sampling
|
Every member of population has an equal chance of being selected for study
|
|
Target Population
|
entire population with characteristics of interest
|
|
Accessible Population
|
portion of target population that has a chance of being selected for study participation
|
|
Sample
|
subgroup of population
|
|
Generalization
|
what is true for the sample is true for the population
|
|
Sampling Bias
|
sample characteristics do not equal target population characteristics. Generalization of study conclusion from sample to population may be incorrect.
|
|
Population
|
group that meets criteria
|
|
Inclusion Criteria
|
used to identify target population
|
|
Exclusion Criteria
|
exclude based on ethical reasons, finances, inadequate data, or potential to skew data.
|
|
PICO
|
PROBLEM - PT Diagnosis
INTERVENTION - treatment COMPARISON OUTCOME - (desired result) • represents the components of a good clinical question |
|
Methodological Studies
|
Use correlative methods to demonstrate reliability and validity of measurement instruments
|
|
Descriptive Research
|
documentation of observations of one or more groups of people
|
|
Experimental Research
|
Establishing cause & effect
|
|
Exploratory/Correlational Studies
|
Establishing relationships between factors (weaker levels of cause & effect)
|
|
Evidence Based Practice
|
method to examine evidence to anser patient care questions
|
|
Source of Knowledge
|
methods of collection patient care related information. Ranges from traditional to scientific
|
|
Reliability
|
Precision, repeatability of a measure, values similar
minimizes random error |
|
Correlation
|
Relationship
|
|
Agreement
|
recoding same actual values
|
|
Continuous Data: Correlation Coefficient
|
used to quantitatively describe the strength and direction of a relationship between 2 variables
examples: Pearson & Spearman's Correlation Coefficient & Intraclass (ICC) Correlation Coefficient |
|
Categorical data: kappa statistic
|
chance corrected measure of agreement, in addition to looking at proportion of observed agreements, kappa also considers proportion of agreements expected by chance
|
|
Categorical Data: percent agreement
|
how often raters agree on scores given to individual subjects
|
|
responsive
|
changes with change, doesn't change with no change
|
|
predictive
|
inclusive of all possible outcomes (an answer for every possible situation)
|
|
Measurement Construct (Tools):
Straight forward tools |
to measure length - ruler, tape measure
to measure speed - distance/time |
|
Measurement Construct (Tools):
Less Straight forward tools |
Ex. Balance - rate basted on test.
|
|
Face Validity
|
knowledgeable individuals agree that the measure is measuring what is intended. WEAKEST TYPE OF VALIDITY
|
|
Content Validity
|
In the development stage of measure, knowledgable individuals identify all aspects of the measure that should be included/excluded. Adds expert opinion while developing measurement tool.
|
|
Criterion related validity
|
predictive ability of a test
|
|
Criterion related validity:
validity of criterion |
must be reliable, free from bias & relevant to target test. Use gold standard test to test a new way to test the same thing
|
|
Criterion related validity:
Concurrent Validity |
Criterion measure and target test scores taken at same time. Use gold standard test and new test at same time
|
|
Criterion related validity:
Predictive Validity |
target test taken to predict the future criterion score/event. Measurement that can predict and outcome. EXAMPLE: measuring TKR pt's knee flexion prior to surgery to predict what their range will be 4 weeks post surgery.
|
|
Construct Validity
|
Does a developed construct appropriately measure the variable to be measured?
|
|
Construct Validity:
Factor Analysis |
identifies different components of a construct
|
|
Construct Validity:
Hypothesis Testing |
Construct behaves in measuring subjects as one would hypothesize
|
|
Construct Validity:
concurrent techniques |
Known group methods
Convergence/discriminant |
|
Minimal Clinically Important Difference
|
a measure of responsiveness. smallest difference in a measure that signifies and important difference in a patient's condition. Smallest difference that patient would perceive as "beneficial". allows determination of better vs not better.
|
|
Statistical Conclusion Validity
|
Is the statistical relationship between 2 or more variables under consideration determined by the appropriate statistical procedures?
|
|
Internal Validity
|
Did the measures of the investigation truly establish causality between variables under study?
|
|
4 Threats to internal validity
|
History - one group has different experience than other
Maturation - subjects change due to time not treatment Attrition - loss of subjects in non-random fashion Testing Effects - ability on follow-up test effected by learning how to take the test on initial measurement. |
|
Hawthorne Effect
|
when people behave differently when they know they are being studied
|
|
External Validity
|
appropriateness of generalizing findings to larger population, outside of experimental situation
|
|
3 Threats to External Validity
|
non-representative samples
setting-specific characteristics that may not generalize time in history when study was done |
|
Measurement Scales: Ratio
|
equal intervals, true zero, meaningful proportions. Example: distance measurement
|
|
Measurement Scales: Interval
|
Equal intervals, nota true zero, proportions not meaningful.
Examples: time on calendar (days/months are equal intervals, 0 is arbitrary) Temperature - for C & F 0 is arbitrarily set a freezing point |
|
Measurement Scales: Ordinal
|
unequal intervals, numbers indicate rank, hierarchal order.
Example: Olympic medalists (unequal intervals but hierarchal) |
|
Measurement Scales: Nominal
|
descriptive categories only, frequency counts per category.
Example: Hair colors in the room |
|
Continuous Measures
|
scale that can be defined into ever diminishing increments (ration and interval scales)
|
|
Categorical Measures
|
measures that have categories, measured by frequency of occurrence
|
|
Dependent Variable
|
outcome or response variable - what is being measured for comparison or what is being predicted (continuous variable)
|
|
Independent variable
|
predictor or treatment variable - variable that is being manipulated by the reseacher (categorical variable)
|
|
Parameter
|
population's characteristic
|
|
Statistic
|
estimate of populations characteristic based on sample from population
|
|
Distribution
|
total set of variable scores and the shape of those scores
|