Research Flashcards
1
Q
Clinical Trial
A
- Tests how well methods of screening, prevention, dx or tx of a disease work in people
2
Q
Completely Randomized Design
A
- Subjects are randomly assigned to different groups and each group receives a unique intervention
3
Q
Crossover Design
A
- Subjects receive BOTH treatments in random ordered separated by period of NO TREATMENT
- Subjects serve as their OWN CONTROL
4
Q
Factorial Design
A
- 2 or more independent variables are investigated with different subjects assigned to the various combinations of levels of the independent variables
5
Q
Pretest-posttest Control Group Design
A
- Compares outcomes of 2+ groups formed by random assignment by testing all groups BEFORE and AFTER treatment
6
Q
Posttest Only Control Group Design
A
- Compares outcomes of 2+ groups formed by random assignment by testing all groups ONLY AFTER treatment
7
Q
RCT
A
- Clinical trial where individual is randomly assigned to an experimental group or control group
8
Q
Repeated Measure Design
A
- Subjects are tested under ALL CONDITIONS
- Act as their OWN CONTROL
9
Q
Sequential Clinical Trial
A
- Data is analyzed as it becomes available so the trial can be stopped as soon as evidence is sufficient
10
Q
Single-subject Design
A
- Drawing conclusions about the effects of treatment based on the responses of a single patient
- Repeated measurements of a response over time
- At least 2 test period phases: (1) baseline before treatment + (2) after treatment
- Can be replicated
- —– Ex: A-B, A-B-A, A-B-A-B, etc
11
Q
Quasi Experimental Design: One-group Pretest-posttest Design
A
- Measurements are made on ONE group of subjects BEFORE and AFTER treatment
- TIME is INDEPENDENT VARIABLE
12
Q
Quasi Experimental Design: One-way Repeated Measures Design Over Time
A
- Extension of One-group Pretest-posttest design
- Measurements are made on one group of subjects at MULTIPLE, PRESCRIBED TIME INTERVALS
13
Q
Quasi Experimental Design: Time Series Design
A
Multiple Measurements are made BEFORE and AFTER treatment to observe patterns or trends during the pretreatment and posttreatment periods
14
Q
Nominal
A
- Aka classification scale
- Values of the variable are mutually exclusive and exhaustive categories
- Can only be in ONE category
- Qualitative
- Ex: blood type, type of breath sound, type of arthritis
15
Q
Ordinal
A
- AKA Ranking scale
- Data ranked on basis of a property of a variable, intervals between may not be equal or known
- Ex: MMT, levels of assistance, pain, joint laxity grades
16
Q
Interval
A
- Intervals between adjacent values are EQUAL, but NO TRUE ZERO
- Ex: temperature
17
Q
Ratio
A
- Intervals between adjacent values are EQUAL, AND there IS A TRUE ZERO
- Ex: ROM, distance walked, time to complete an activity, nerve conduction velocity
18
Q
Reliability
A
- Reproducibility or repeatability of measurements
19
Q
Reliability: Alternate Forms Reliability
A
- AKA Parallel Form Reliability
- Assesses the consistency or agreement of measurements obtained with different forms of a test
- NECESSARY if different forms of test are to be used INTERCHANGEABLY
- ——Ex: NPTE
20
Q
Reliability: Internal Consistency
A
- Extent to which items or elements that contribute to a measure reflect one basic phenomenon or dimension
21
Q
Reliability: InTRArater Reliability
A
- Same person over time
22
Q
Reliability: InTERrater Reliability
A
- Different people
23
Q
Reliability: Test-retest Reliability
A
- Same individual on separate occasions
24
Q
Validity
A
- Degree to which a useful of meaningful interpretation can be inferred from a measurement
25
Q
Validity: Face Validity
A
- Degree to which a measurement appears to test what it is supposed to
26
Q
Validity: Content Validity
A
- Degree to which a measurement reflects the meaningful elements of a construct and the items in a test adequately reflect the content domain of interest and not extraneous elements
- Whoever wrote this should burn in hell
27
Q
Validity: Construct Validity
A
- Degree to which a theoretical construct is measured by a test or measurement
28
Q
Validity: Criterion-related Validity
A
- Validity of the measurement is established by comparing it to either a different measurement (gold standard) or data obtained by different forms of testing
29
Q
Validity: Criterion-related Validity: Concurrent Validity
A
- Comparing a measurement to a “gold standard” at APPROXIMATELY THE SAME TIME
30
Q
Validity: Criterion-related Validity: Predictive Validity
A
- Measurement is considered to be valid because it is predictive of a future behavior or event
- Ex: High school GPA to predict performance in college
31
Q
Validity: Criterion-related Validity: Prescriptive Validity
A
- Measurement suggests the form of treatment a person should receive
32
Q
External Validity
A
- Degree to which results of the research study are generalizable to populations or circumstances beyond those included in the study
- Threats: interaction of tx with specific type of subjects tested, setting/place and time/history in which the experiment is performed
33
Q
Internal Validity
A
- Degree to which an intervention being evaluated is the cause of the outcome measured in the study
- Threats: history, maturation, attrition, testing, instrumentation, regression toward the mean
34
Q
Type I Error (Alpha Error)
A
- Aka false positive
- Concluding there is a relationship when there is NOT
- Wrongly rejecting the null hypothesis
35
Q
Type II Error (Beta Error)
A
- Aka false negative
- Concluding there is NOT a relationship when there is one
- Wrongly deciding not to reject the null hypothesis
36
Q
Effect Size
A
- Measure of the magnitude of the difference between 2 treatments or the magnitude of the relationship between 2 variables
- LARGER ES = more likely it is statistically significant
37
Q
Effect Size Index
A
- Represents effect size using a standardized value
- Calculation: Mean control group - (subtract) mean experimental group / (divided by) std deviation of one of the groups
- —— < 0.1 = trivial effect
- —— 0.1 - 0.3 = small effect
- —— 0.3 - 0.5 = mod effect
- —— > 0.5 = large effect
38
Q
Minimal Clinically Important Difference (MCID)
A
- Smallest difference in condition that the patient or PT consider WORTHWHILE
39
Q
Minimal Detectable Change (MDC)
A
- Smallest difference that is NOT DUE TO CHANCE = that it is statistically significant
40
Q
Standard Deviations
A
- 1 std deviation (above + below) = 68%
- 2 std deviations (above + below) = 95%
- 3 std deviations (above + below) = 99%
41
Q
Coefficient of Variation (CV)
A
- Ratio of the std deviation of a distribution to the mean
- Expressed as a %
- Equation: CV = (s / mean) x 100
42
Q
Parametric Stats
A
- Assume samples come from NORMAL DISTRIBUTION
43
Q
Parametric Stats: ONE-Way Analysis of Variance (ANOVA)
A
- Tests the equality of means between 2 or more populations by analyzing sample variances
- Similar to t-test but accommodates 2+ population means
- Separated into different groups on the basis of ONE characteristic/factor/independent variable
44
Q
Parametric Stats: TWO-Way Analysis of Variance (ANOVA)
A
- Tests the equality of means between 2 or more populations by analyzing sample variances
- Compare 2+ population means with 2 independent variables
45
Q
Parametric Stats: Repeated Measure ANOVA
A
- All individuals are measured under a number of different experimental conditions
- Used ONLY if there is limited learning/carryover
46
Q
Pearson Product Moment Correlation
A
- Measures the magnitude + direction of a linear relationship between 2 variables
- Range -1.0 to +1.0
- —- -1.0 = inverse relationship in a straight line
- —- +1.0 = both increase/decrease at the same time
- —- 0 = no relationship
47
Q
Sensitivity
A
- The % of people who test positive for a specific disease who actually have the disease (i.e. testing positive when the person ACTUALLY has it)
- SNOUT - used to rule out a dx because if the test is NEGATIVE the person does NOT have it (because if it were positive the test would be positive)
48
Q
Specificity
A
- The % of people who test negative for a specific disease who do NOT have the disease (i.e. testing negative when you do NOT actually have it)
- SPIN - used to rule IN a dx because if it is POSITIVE the person HAS it (because if it was negative the test would be negative)
