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;
189 Cards in this Set
- Front
- Back
|
primary vs secondary sources |
- Primary Source: Original document by primary author
- Secondary SourceReport: review, summary of work in the field |
|
|
Six Steps in the Literature Search |
1. Write the problem statement. 2. Consult secondary sources. 3. Determine descriptors 4. Find primary sources 5. Read and record the literature. 6. Write the literature review.
|
|
|
Ethics versus Morality |
- morality from the latin “ manner, character, proper behaviour)”, philosphical belief
- ethics comes from greek which means “character”, different cultural beliefs |
|
|
Ethics Definition |
• The principles of conduct governing an individual or a group
- protect the rights of participants - ensure the scientific validity of the research study (ex: do the beliefs of the study to society outweigh the risks to the participants - three basic principles: autonomy, beneficence, justice |
|
|
autonomy & Justice
|
- ability of individuals to make decisions and act out those decisions, no controlling influence
- fairness - respectively |
|
|
Beneficence
|
: do good, well being of all individuals, vs nonmaleficence (do no harm), benefits should outweigh the harm
|
|
|
Research Practice
|
- safety- informed consent- debrief- confidentiality/privacy-opt out clause- responsible researcher- willingness to participate- coercion
|
|
|
Protecting Human Participants • What should human research participants expect? |
- right to privacy or non-participation - right to remain anonymous - right to confidentiality - right to experimenter responsibility |
|
|
Tuskegee Trial |
- 1932: “Tuskee study of untreated syphillis in the negro male” 1932- 399 with syphillis- 201 without- no informed consent- told they were being treated for bad blood - originally 6 months… went for 40 years- 1947: penicillin… not given to these men- Many people died for unnecessary reasons
|
|
|
Elements of informed consent |
Information: – Subjects informed fully – Information kept confidential and anonymous – Written in lay language – Researcher to answer questions Consent: – Voluntary – Special considerations-capacity to consent – Free to withdraw from the study
|
|
|
Conflict of Interest |
“conflicts of interest exists when an author (or the authors institution), reviewer, or editor has financial or personal relationships that inappropriately influence (bias) his or her actions
Potential Conflicts of Interest: • Financial • Personal relationships • Academic competition • Intellectual passion |
|
|
Areas of Scientific Dishonesty
|
1. Plagiarism 2. Fabrication and falsification: making up or altering data 3. Non publication of data, also called cooking data 4. Faulty data-gathering procedures 5. Poor data storage and retention 6. Misleading authorship 7. Unacceptable publication practices |
|
|
Scientific Method of Problem Solving Steps |
Step 1: Developing the problem (defining and delimiting it) Step 2: Formulating the hypotheses Step 3: Gathering the data Step 4: Analyzing and interpreting results
|
|
|
Independent variable |
– Variable that the researcher is manipulaKng – Experimental/treatment variable – “Grouping” variable |
|
|
Dependent variable |
– Effect of the independent variable – May also be called the “yield” |
|
|
operational definitions
|
detailed definition of a variable (a balance exercise….standing straight eyes closed etc)
|
|
|
Continuous Variable: |
• Takes on a value along a range of values • Fractions can occur - ex: height, weight, tempature |
|
|
Discrete Variable: |
• Whole units - number of children in family, male/female |
|
|
Nominal Scale |
• Categories according to criterion • Numbers represent the label of a category - gender, sport position, hair colour |
|
|
Ordinal Scale |
• Categories that are ranked • Many clinical measures use this • Distance between values is not necessarily equal • Relative position within a distribution - ex: manual muscle testing, Likret scales (agree, disagree) |
|
|
Interval Scale |
- equal distances between values - not related to a true zero - ex: year |
|
|
ratio scale |
• Highest level of measurement • Equal intervals • Interval scale that has an absolute zero • NegaKve values not possible • Can be transformed directly from one scale to another - ex: height (inches vs cm), weight (pounds vs kg) |
|
|
EXAMPLES OF SCALES |
- minutes: ratio (can change)- type of vehicle: nominal - volume of water: - movie ratings: ordinal (order, the more stars the better)- disease type: nominal- specific disease rating: ordinal - money (dollars): ratio
|
|
|
Step 2: Formulate Hypothesis |
- Anticipated outcome of the study/ experiment *Must be testable-study will support or refute |
|
|
Step 3: Gathering Data |
• Determine most appropriate methods to gather the data • Important that methods are “sound” • Internal validity • External validity |
|
|
Step 4: Analyze and Interpret Results |
• Statistical analysis • Supports or rejects hypothesis • Compare to relevant literature |
|
|
Descriptive Research |
• “Describe” events • No true comparison between groups - Example: Case study |
|
|
Analytical Research |
• Testing hypothesis • May be experimental or observational • Differences in outcome between groups - Example: Randomized controlled trial |
|
|
Observational Research |
• Observe and describe • May be descriptive or analytical • Exploratory • No intervention or manipulated variable - Example: Case series & Cohort study |
|
|
Experimental Research |
• Researcher manipulates variables and observes outcome - Example: Randomized controlled trial & Quasi experimental |
|
|
Qualitative Research |
• Open-ended questions, interviews, observations • Atempt to understand a phenomena • Describe, explore, develop theory or hypothesis |
|
|
Quantitative Research |
• Measure outcome of interest • Direct measurement • Subjective information into a questionnaire • Advantage: summarize data & statistical analysi |
|
|
Basic Research |
• “Bench” research • Typically done in lab • Test or clarify a theory |
|
|
Applied Research |
• Test theories in practice • Solving practical problems • Clinical Research |
|
|
Hierarchy of evidence |
1. Systematic Reviews 2. RCTs 3. Cohort Studies 4. Case Control Studies 5. Cross section surveys 6. Case reports |
|
|
Systematic Review/Meta-Analysis |
• Systematic search of literature – Identifiable methods of search – Data extracted and critiqued • Meta-Analysis – Statistical technique – Integrate multiple studies to determine overall estimate |
|
|
Experimental Studies |
• Randomized Controlled Trial (RCT) • Quasi-experimental • Pre-Experimental |
|
|
Experimental-RCT |
• Randomly allocated to treatment/control group • Control group: Typically standard of care • Intervention group: Intervention of interest Pros - considered highest level of evidence - provides strong evidence for cause and effect relationship Cons - $$ - may be very select group (ex: challenge with generalizability) |
|
|
Efficacy vs effectiveness |
- Intervention in controlled conditions - Intervention in “real world” situation |
|
|
Quasi Experimental |
• No randomization! • Assigned intervention • Potential for bias by way that individuals were grouped |
|
|
Pre-experimental |
• Only one group (everyone gets same intervention) • Independent variable is time • Pretest measures • Pos test measures |
|
|
Observational studies |
• Individual: Cohort, Case-Control, Cross-sectional • Population • Case series |
|
|
Observational Cohort Study |
• Prospective (typically) • Classified according to exposure • Measures over time to evaluate for disease pros - can measure multiple outcomes - good for rare exposures - data collected prospectively-better cause/effect cons - $$ and time consuming - selection bias (drop outs) - disease must be common or is inefficient |
|
|
Incidence Rate Ratio: |
– Rate of disease in Exposed/ Rate of disease in UnExposed - cohort |
|
|
Observational Case Control |
• Identify individuals with disease (D) • Look back for exposure (E) • Retrospective • Study prognostic factors Pros - rare diseases - efficient Cons - recall bias - selection bias - cause/effect relationship |
|
|
Odds ratos: |
– Odds of E in individuals with disease/ Odds of E in individuals without disease - case control |
|
|
Observational Cross Sectional |
• Exposure and outcome (disease) at same time • Example=survey Pros - quick - whole population or sample - prevalence estimates - gernalizability Cons - selection bias - recall bias - cannot establish cause and effect * incidence = all new cases * prevalence = cases at one point in time |
|
|
Observational Population |
• Also called ecological • Problem-ecological fallacy |
|
|
Observational/Experimental Case Study/Case Series |
• Uncontrolled • Descriptive • One (case study) or more (case series) patients |
|
|
Population vs. sample |
- “Larger group from which the sample is taken” - “Group of participants selected from a larger population |
|
|
Probability sampling |
1) Simple Random Sampling 2) Systematic Sampling 3) Simple Stratified Sampling 4) Cluster Sampling |
|
|
Non-probability sampling |
5) Convenience Sampling |
|
|
1) Simple Random Sampling |
• Use of a random numbers table • Computer generated random numbers |
|
|
2. Systematic Sampling |
- another method of sampling from a population - define the kth item - example: you want to obtain a sample of 500 individuals from a town of 10,000 people using systematic sampling |
|
|
3) Simple Stratified Sampling |
Use when population is too large • Divide population into “strata” • Randomly select from each strata • Proportionally select the number - Example: You want to obtain a sample of 40 from a population of 220. |
|
|
4) Cluster Sampling
|
• Divide population into “clusters” • Randomly select individuals from each cluster |
|
|
5) Convenience Sampling |
• Chosen on basis of availability • Most common = consecutive sampling - Example: Each consecutive patient attending a certain medical clinic is enrolled in the study. |
|
|
Background/Rationale |
1-2 sentences. Why does this study need to be done? (to find new information, what to reproduce study, make better)
|
|
|
Objective of the Study |
A clear concise statement of the purpose of the study |
|
|
Statistical Hypotheses |
– Null Hypothesis (H0): predicts no difference – Alternative Hypothesis (H1 or HA): states expected relationship |
|
|
Study Validity |
the degree to which the difference drawn from a study (including generalizability beyond the study sample) are warranted considering the study methods., representativeness of the study sample and the nature of the population from which the sample is drawn |
|
|
Internal Validity |
• Are the results true??? • Threats to validity: bias and random error |
|
|
Confounder: |
• A factor that obscures the relationship between the independent variable and the dependent variable • “Mixing” of effect • Common confounders: age, ethnicity, gender, social economic status |
|
|
Selection bias: |
• Systematic error in the way that subjects were selected for a study compared to those who are not that impacts the relationship between exposure and outcome. - volunteers: people who are interested in the study (ex: political beliefs) |
|
|
Measurement/Information bias: |
• Systematic error in measurement of information |
|
|
Random (Stochastic) Error |
• Variation in a measure • “Due to chance” • How long does it take to walk across the front of the room? |
|
|
External Validity |
external validity is the degree to which the conclusions in your study would hold for other persons in other places and at other times.
• Can the results of the study be applied? – Other popula4ons – Unbiased inferences |
|
|
Selection bias |
• Systematic error
• Those selected for a study are systematically different than those not selected for a study with respect to the outcome of interest. • Examples: Convenience sample & Drop outs |
|
|
Measurement Bias/Information Bias |
• Systematic error • Measurement of variables • Misclassification bias |
|
|
Misclassification bias |
Non-differential misclassification: • Same chance of being misclassified in each group • Ex. Measuring tool Differential misclassification: • Different chance of being misclassified in each group • Ex. Recall bias |
|
|
Hawthorne Effect: |
people act different when they know they are being watched |
|
|
3 Things A confounder must be: |
1. Independently associated with development of the outcome (smoking can cause CHD) 2. Associated with the exposure of interest (coffee and smoking have a relationship) 3. Not on the exposure-disease pathway (does drinking coffee cause smoking to cause CHD? No) |
|
|
Control for Confounding |
Research design • Randomization • Restriction • Matching Analysis • Modeling • Stratified analysis |
|
|
Effect modification |
• Not a source of bias! • Effect of a variable changes the outcome variable • Helps understand the nature of the variable • Example: Symptoms as risk factors for concussion |
|
|
Causation
|
relationship between cause effect - usually more than one cause component - component cause: multiple usually, some could be stronger than others |
|
|
Necessary Cause |
Always before an effect - has the be there - ex: HIV + AIDS |
|
|
Sufficient Cause |
- leads to/produces an event - hit play button to hear music, need speaker etc |
|
|
Sir Bradford Hill's Criteria (9) |
1. Strength of association: 95% more likely to fall wearing certain of shoes 2. Consistency: same cause has the same effect over and over 3. Specificity: 1 cause produces 1 result, single outcome 4. Temporality: cause should happen before effect, strongest 5. Biologic gradient: dose response relationship, the more you do the more difference result 6. Biologic plausibility: does it make sense? 7. Coherence: aligns with what is currently known 8. Experimental evidence: *RCT is best and systematic review to be sure 9. Analogy: assume that it would work in all populations * weakest |
|
|
Validity of Measurement tools |
• Is the measure measuring what it is intended to measure? Face Content Criterion Construct |
|
|
Face Validity |
• Logical Validity • Involves the performance being measured - ex: develop concussion tool, - does it measure what it’s intending to measure, does it look appropriated |
|
|
Content Validity |
• Test adequately measures all content • Upper Extremity Functional Scale |
|
|
Criterion Validity |
• Scores are similar to a criterion or standard • Pick standard carefully! • Two types: Concurrent (measure at the same time) & Predictive (predict future outcome) |
|
|
Construct Validity
|
• Scores on a test match a construct/theory • Balance |
|
|
Types of Validity |
Study – Internal – External Measurement Tool – Face – Content – Criterion – Construct |
|
|
5 Characteristics of Research |
1. Systematic: plan, identify, design, data, evaluate 2. Logical: examines procedures to evaluate conclusions 3. Empirical: decisions are based on data 4. Reductive: general relationships are stablished from data 5. Replicable: actions are recorded |
|
|
Inductive Reasoning |
- start with observation - develop hypothesis - test theory |
|
|
Deductive Reasoning |
- start with theory - test it - decide if hypothesis is correct |
|
|
Reliability |
• Consistency • Reproducibility • Dependability • Free of error |
|
|
Systematic Error |
• Predictable • One direction – over or underestimate true score • Can correct for or recalibrate • Problem with validity |
|
|
Random error |
• Due to chance • Unpredictable • Can occur for a variety of reasons (fatigue, inaccuracy) - With many trials, would eventually cancel out so average score is good es4mate of true score |
|
|
Sources of measurement error |
1. Individual taking measures 2. Instrument 3. Characteristic under measure |
|
|
Types of Reliability |
- test retest - intrarater -interrater |
|
|
Intrarater Reliability |
• Data recorded on one individual across trials • Self report measures? • Rater vs test? • What if the examiner remembers the first score? • Standardized protocol is ++ important |
|
|
Interrater reliability |
• Variation between raters • Same subject • All individuals observe the same performance of a task - time walking, walking with head motion |
|
|
Qualitative Research guest lecture |
How has Olympic medal designation shaped the sport of curling? |
|
|
Data Collection
|
Interviews Focus groups Textual analysis Autoethnography |
|
|
Data Analysis
|
Interpretation Thematic analysis Discourse analysis |
|
|
Thematic Analysis
|
Curling’s Olympic transition Delinking the competitor and the club Achieving credibility through athleticism Growing at the top at the expense of the bottom Striving to recognize (and embrace) a new vision for curling |
|
|
General Goals of Qualitative Research
|
Depth: What can we learn beyond numbers? Voice: Who can we represent in research and how? Outcomes Mapping Critique Social Change |
|
|
Validity in Qualitative
|
Credibility Transferability Dependability Trustworthiness Authenticity |
|
|
** |
Combining of qualitative and quantitative research is becoming more and more common. It is important to keep in mind that these are two different philosophies, not necessarily polar opposites. In fact, elements of both designs can be used together in mixed-methods studies. |
|
|
Objectives: |
-create student focussed curricula -teach student self monitoring -integrate PE in academic work and school activities -daily physical activity |
|
|
What is exergaming: |
-combining gaming and activity together -users use their body as the controller |
|
|
PhysicalLiteracy
|
–the motivation, confidence, physical competence, knowledge and understanding tovalue and take responsibility for engagement in physical activity for life
|
|
|
Fourcomponents of eastern philosophy to human movement:
|
-Visualization and observational skill -Bilateral body development -Peer to peer (P2P) learning -TASP – technique, accuracy, speed, power |
|
|
Reflections on article: |
-Free play vs structured play -Perceived exertion -Sharing and cooperation -Self monitoring and self reporting -Integrate in academic work and school activities |
|
|
What are Statistics |
An objective means of interpreting a collection of observations.”
|
|
|
What are statistics used for |
- Describe characteristics of data
- Test relationships between sets of data - Test differences among data sets |
|
|
Analysis using Stats Ex Height and standing long jump score in seventh graders |
• Describethe data: Mean
• Association: Measureassociation between height and long jump score • Measuredifferences: Randomizeinto two groups and put one group on a weight training program and evaluate ifthere is a difference in standing long jump after training |
|
|
Unit of Analysis
|
What can be considered the most basic unit from which data can be produced”
- Typically an individual participantGrade 7 student - ClusterGrade 7 class |
|
|
Measure of Central Tendency
|
- One number that can represent the “group”: MeanMedianMode
- Degree of difference between scores is the variability |
|
|
Mean |
M = ΣX/N
M=mean X=Value of variable of interest N=Number of values |
|
|
Median |
- A measure of central tendency representing the middle score in a group.
- (N+1)/2 - If median lies between two numbers then take the mean of the two numbers. |
|
|
Mode
|
The most frequently occurring score Example: 1, 3, 4, 5, 5, 5, 7, 8, 8, 8, 8, 10 |
|
|
Population + Parameters |
- Population is a complete set of individuals of interest
- Population parameters:Mean (mu) Standard deviation (sigma) |
|
|
Characteristics of normal curve
|
– mean = median = mode and sd
Mean + 1 SD ~ 68% of all scoresMean + 2 SD ~ 95 % of all scoresMean + 3 SD ~ 99% of all scores |
|
|
How is Variability Measure |
Degree of difference between scores is thevariability
– Range (b-a) – Standard Deviation – Confidence Intervals |
|
|
Standard Deviation |
Estimate of variability
s = √Σ(X-M)2/(N-1) X=Individual scores M=Mean N=Number of scores s^2 = variance |
|
|
Normal Distribution
|
• Also known as normal curve, Bell Curve or Gaussian Curve
• Characterized by a bilateral symmetrical bell shaped distribution of data • Mean, Median and Mode are the same |
|
|
Raw Score |
Which is better?
• 50 sit–ups in a minute or ( 1 raw score) • Throwing a ball at 60 Kilometers per hour ( 1 raw score) |
|
|
Standard Score
|
Allow us to evaluate raw scores and compare setsof data that are based on different units ofmeasurement
|
|
|
Central tendency use
|
- if you have normal distribution: you use mean - you use more or median for skewed
|
|
|
Standard Score Z test |
Compares means of one group to mean of the population Z score = (X -|X)/SD X = raw score X = mean SD = standard deviation |
|
|
Confidence Intervals |
• “Provides an expected upper and lower limit for a statistic at a specified probability level.”
• Typically 95% (1.96) • Provides an estimate of the variability of thedata. |
|
|
Confidence Intervals Are effected by: (3) |
– Size of sample– Homogeneity of values within the sample – Level of confidence selected
|
|
|
Confidence Interval Equation |
CI = mean +/- (standard error x confidence level)
standard error = s/√n |
|
|
Descriptive Statistics
|
• Describe data
• Characterize shape, central tendency and variability of data • Describe a population • Parameter = Population • Statistic = Sample |
|
|
Inferential Statistics
|
• Estimate population values from sample data
• Assumptions: sample represents the population• Assumptions based on:– Probability– Sampling error * branch of stats concerned with testing a hypothesis and using sample data |
|
|
What is probability "p" |
Probability is the likelihood that any one eventwill occur, given all possible outcomes
|
|
|
Sampling error
|
- the tendency for sample values to differ from population values - unpredictable and occurs due to chance - SE = sample mean (x) - population mean (mu) - greater sample error = less accurate estimate of pop mean |
|
|
Standard Error of the Mean
|
• Greater sample size = less sampling error
• Estimate of population standard deviation • Smaller n = greater sample variability • Bigger sample size = better representation of population mean (ie means are more likely to be close to population mean) sx= s/√n |
|
|
Purpose of Confidence Intervals |
- estimate how population behaves - provides an unexpected upper/lower limit for a stat at a specified probability level |
|
|
Define: Measurement, data, stats, evaluation |
- Measurement: process of comparing a value to a standard - Data: Result of a measurement- Statistics: Mathematical technique to organize and interpret data - Evaluation: Determine the worth of the data |
|
|
Is the data useful???
|
Reliable:– Reproducible, consistent
Valid:– Does it measure what it is intended to measure (Measuring tool)– Free of bias (internal validity of study) |
|
|
Why use descriptive stats? |
- summarize - get an idea of the distribution of data - Norms to compare to - allows to interpret |
|
|
How to describe data |
• Measures of central tendency
• Point estimate (mean, median, mode) • Variability • Graphs |
|
|
Purpose of Hypothesis Testing |
• estimate pop measures: Does this class have the same proportion of femalesas the other classes at the university?
• compare means: Do individuals who undergo a period of cognitive and physical rest have less time lost from sport than individuals who continue to be active? • relationship between variables |
|
|
Statistical Hypotheses |
- Null Hypothesis (H0):No difference in the groups
- Alternate Hypothesis (HA):True difference between groups |
|
|
Can you prove the Null Hypothesis |
- No, cannot prove or accept H0
- proper conventions: do not reject H0, fail to reject H0 |
|
|
Alternative Hypothesis |
- Likelihood that the difference is due to chance is small
- observed difference is not just due to chance - observed difference is real - reject H0 and accept HA |
|
|
Directional and Non-directional Hypothesis |
- uA does not equal uB (non directional, null) - uA > uB and uA < uB (directional and alternative) |
|
|
Type I Error (alpha) |
• Occurs when the H0 is rejected when it is fact true, false positive
– i.e. Conclude that there is a difference when there is not (difference is due to chance) • α and p |
|
|
• α:
|
probability of committing a Type I error of H0 = true•
|
|
|
p:
|
probability of difference occurring by chance
|
|
|
1. the ____ the p-value, the _____ the chance that you are actually seeing a true difference
2. the ___ the alpha value, the ____the Type I error rate |
1. smaller, greater 2. lower, lower |
|
|
5% significance level (α = 0.05)
|
• If p < α, then reject the null hypothesis
• if p=0.08: fail to reject the null hypothesis, p=0.002: reject the null hypothesis |
|
|
How do you interpret a p value? |
• p ≤ 0.01 very strong evidence against H0
• p < 0.05: strong evidence against H0 • p=0.05: inconclusive • p >0.05: little evidence against H0 |
|
|
Type II Error (Beta) |
• A difference does exist but was not found (ie results were not significant)
• Probability of failing to reject a false null hypothesis - false negative • Evidence not strong enough to reject null. • What could some implications of this be?: Assume treatment ineffective– Conflicting results in literature– Stop investigating in an area |
|
|
Beta |
• β = 0.20– 20% chance that fail to reject null hypothesiswhen it is really false
• β = likelihood that we are unable tostatistically identify a real difference |
|
|
Power |
• Power=1–β
• More power = less chance type II error - probability that a test will lead to rejection of the null hypothesis - probability of attaining statistical significance - size of your alpha effects your power |
|
|
Determinants of Power
|
• Significance criterion (α): Less power with smaller alpha
• Variance (s2): Greater power when less variance• Sample Size (n): Greater power with larger sample = decreases variance • Effect size (ES): larger effect size = greater difference between groups = greater power |
|
|
Effect Size |
“effect” of the experimental variable
– degree to which null hypothesis is false – Larger effect size=greater difference between groups – Large difference between groups=more likely to be significant |
|
|
Purpose of Z test |
- To determine if the mean of a sample is different from the population
- statistical hypothesis test: You conduct a study to see whether children in an athletic school program weigh less than children in the general population. |
|
|
H0: μ0=μ1
HA: μ0<μ1 |
** |
|
|
One Tailed vs Two Tailed |
• A two-tailed test is more conservative than a one-tailed test because a two-tailed test takes a more extreme test result to reject the null
• Unless you have a very good reason for a one- tailed test – use a two tailed! (e.g. The literature must show support for a one tailed test). |
|
|
Parametric Test
|
• Used to estimate population parameters
• Arithmetic manipulations • Valid if meets assumptions, Assumptions not met? Use nonparametric statistics • Generally considered more powerful |
|
|
Assumptions needed to be met by parametric test |
– Samples randomly drawn from normally distributedpopulation
– Variances are approximately equal – Interval/ratio data |
|
|
Non Parametric test
|
• Used when assumptions not met for parametric tests
• Can use if:– Data not normally distributed – Not homogeneity of variance • Small samples • Nominal and ordinal scales |
|
|
t-test
|
• Compares two means (Independent or Dependent )
• Parametric testAssumptions: |
|
|
Independent vs Dependent |
- independent: samples are independent of each other, 2 different groups - dependent: test same group of people before and after intervention |
|
|
T Test - DF is ___ for one group and ___ for 2 groups- high sample size we have a more ___curve |
- (n-1), (n-2) - narrow |
|
|
Tvalue ratio |
• Ratio = Difference between group means/ Variability within groups
= Treatment effect + error/ Error |
|
|
Independent samples t-test
|
• t=X1 –X2/ sX1-X2
Degrees of freedom: df= N-2 - less variability = bigger t score, more likely to be greater than your critical value |
|
|
Dependent/ Paired T test |
Pre-post treatment
• Uses a difference score d t = d/sd Where d = mean of the difference scores sd = standard error of the difference scores = sdv= sd/√n Degrees of freedom = n - 1 |
|
|
t-value
|
• The theoretical cut-off value is called t-critical and is found in the t distribution table
• This table is based:– sample size (degrees of freedom) – pre-determined error risk (alpha level) |
|
|
Mann Whitney U test
|
• Nonparametric version of independent t-test
• Ranks groups • H0 = mean of ranks are equal for both groups • U test statstc |
|
|
Sign Test
|
• Nonparametric version of dependent (paired) t-test
• Data analyzed using signs (+ or -) • If second score is greater then + is assigned • If second score is lower then – is assigned • No difference = 0 • H0 = half scores + and half - |
|
|
Analysis of Variance (ANOVA)
|
• Compare more than two means
• 3 or greater groups/conditions are compared – Independent groups– Repeated measures • Based on “F statistic” |
|
|
Incidence Rate Ratio (IRR):
|
Rate of disease in Exposed/ Rate of disease in UnExposed
*observational cohort study *** Incidence Rate (IR) = # new cases of disease/ # in sample |
|
|
Regression |
• Predict outcomes
• Powerful statistical approach • Regression towards the mean • Able to include multiple factors within a model • Equation to predict values of outcome (Y) • Y=a +bX |
|
|
Odds Ratio (OR) |
OR = Odds of E in individuals with disease/Odds of E in individuals without disease
* observation case control |
|
|
Chi-square test (x^2)
|
• Is there a difference in proportions in a categorycompared to what would be expected by chance?
• Nonparametric test • Is distribution of frequencies different from expected frequencies?: Ex 2 categories then would expect 50:50 distribution |
|
|
Chi Square Assumptions |
– Actual number of people/items (not percent or rank)
– Categories exhaustive and mutually exclusive (ex male/female) |
|
|
Chi square equation |
X2 = Σ(O-E)^2/ E
O: observed frequency E: expected frequency Significant if x2 greater than critical value |
|
|
1. Towards the real time monitoring of Achilles tendon strain and cumulative damage in basketball players using a wearable sensors
|
- tendiopathy: overuse, tendon strain is related to accumulation of damage
- Prevent: minimize tendon strain , Footwear modifications help - Objectives: look at relationship between strain and foot cushioning shock and strain - experimental bias: limited ability to blind subjects - narrow target populations - results of this study can be used by footwear companies as well as athletes and coaches |
|
|
2. The feasibility of PA program in Advanced Cancer Patients
|
- a lot of benefits in cancer patients
- Stage 3 or 4 cancerPA helps in every stage of cancer - objectives: Understand feasibility of intervention and look at quality of life and their function abilities, look at their feelings towards PA -18 >, M or F, Stage or III IV, ESASE score >3/10, cleared by physical for PA - group exercise program 1 x a week and an individual program for home - quantitative and qualitative - convenience sample at one cancer centre, they can be both on or off treatments |
|
|
3. Assessing Responsiveness of the Rotator Cuff Quality of Life Index
|
- internal consistency, content validity, responsiveness
- Index: disease specific health related PROM - Objectives: how valid is the index in detecting change after 24 month period, evaluating properties of responsiveness - 98, non surgical and surgical, 18 <, - convenience sample, misclassification bias (recall) |
|
|
Correlation
|
• Are A and B related to each other?
• describe with correlation coefficients • Scatterplot used to illustrate relationship |
|
|
Values of correlation coefficients:
|
• -1.0 = perfect negative relationship
• 0 = No correlation • 1.0 = perfect positive relationship • Closer to ±1.0 = stronger association |
|
|
Sign of correlation
|
– Positive = as X increases, Y also increases
– Negative = as X increases, Y decreases |
|
|
Pearson product-moment correlation coefficient (r)
|
– A measure of the strength of association between 2 or more variables – r for sample data – ρ for population data – Based on concept of covariance • Interval or ratio data • Scatterplot to illustrate relationship between data • df = n-2 • Critical value of r |
|
|
Pearson test of significance |
• Observed is greater than critical r? Reject null
• Observed is less than critical r? Fail to reject null• Coefficient of determination = r2 = fraction of variability in y explained by x |
|
|
Coefficient of Determination
|
• The squared correlation coefficient (r2); used in interpreting the meaningfulness of correlation
• represents the fraction of variability in y that can be explained by the variability in x. • r2 - the proportion of variance in common between the two variables |
|
|
Purpose of Correlational Research
|
1. Examining Relationships Between Variables
2. Predicting Outcomes |
|
|
Importance of Graphical Analysis
|
- Assists with giving an idea of the expected outcome
– Assists with detecting any irregularities in the data • e.g. Pearson Correlation may not be appropriate statistical test • Distinct groups • Non-Linear relationship • Data not normally distributed |
|
|
Correlation and Causation
|
Correlation between two variables represents the degree of observed linear association between two variables, not the extent of their causal relationship
|
|
|
Spearman’s rho Rank Correlation Coefficient (rs) |
(non parametric) • Measure of association that requires both variables be in at least an ordinal scale (ranked) – Class rank on mid-term and final – Scores between two judges |
|
|
Parametric vs Non Parametric * two related groups, two independent and correlation |
- Two related groups: Paired (Dependent samples) t-test (para) and Sign test (non)
- Two independent groups: Unpaired (Independent samples) t-test (para) and Mann-Whitney U test (non para) - Correlation: Pearson product moment correlation coefficient (para) & Spearmen rho (non para) |
|
|
Right skew, left skew and normal |
Normal: mean = median = mode Left Skew = -, mean, median, mode right skew = +, mode, median, mean |
