Final Flashcards
Research Statistics Final Review
- Independent groups
- Two levels of IV
- Has met assumptions for parametric tests
What test = ?
Purple
- Independent groups
- Two levels of IV
- Has met assumptions for parametric tests
What test = Unpaired T-Test
Research Statistics Final Review
- Independent groups
- Two levels of IV
- Has NOT met assumptions for parametric tests
What test = ?
Purple
- Independent groups
- Two levels of IV
- Has NOT met assumptions for parametric tests
What test = Man-Whitney U
Research Statistics Final Review
- Independent groups
- Three + levels of IV
- Has met assumptions for parametric tests.
What test = ?
Purple
- Independent groups
- Three + levels of IV
- Has met assumptions for parametric tests.
What test = IG NOVA
Research Statistics Final Review
- Independent groups
- Three + levels of IV
- Has NOT met assumptions for parametric tests.
What test = ?
Purple
- Independent groups
- Three + levels of IV
- Has NOT met assumptions for parametric tests.
What test = Kruskal-Wallis ANOVA
Research Statistics Final Review
- Repeated measures
- Two levels of IV
- Has met assumptions for parametric tests.
What test = ?
- Repeated measures
- Two levels of IV
- Has met assumptions for parametric tests.
What test = Paired T-Test
Research Statistics Final Review
- Repeated measures
- Two levels of IV
- Has NOT met assumptions for parametric tests.
What test = ?
Purple
- Repeated measures
- Two levels of IV
- Has NOT met assumptions for parametric tests.
What test = Wilcoxon Signe-Rank
Research Statistics Final Review
- Repeated measures
- Three + levels of IV
- Has met assumptions for parametric tests.
What test = ?
Purple
- Repeated measures
- Three + levels of IV
- Has met assumptions for parametric tests.
What test = RM ANOVA
Research Statistics Final Review
- Repeated measures
- Three + levels of IV
- Has NOT met assumptions for parametric tests.
What test = ?
Purple
- Repeated measures
- Three levels + of IV
- Has NOT met assumptions for parametric tests.
What test = Friedman’s ANOVA
Research Statistics Final Review
probability of Type I error = ?
Purple
alpha: probability of Type I error
- Set BEFORE the study
t-Test assumptions
Research Statistics Final Review
probability of Type II error = ?
Purple
beta: probability of Type II error
Research Statistics Final Review
calculated probability of Type I error = ?
p-value: calculated probability of Type I error
- AFTER the study
Research Statistics Final Review
calculated value for t-test = ?
Purple
t Statistic: calculated value for t-test
Research Statistics Final Review
calculated value for ANOVA = ?
Purple
F Statistic: calculated value for ANOVA
Research Statistics Final Review
equal variances for RM ANOVA = ?
Mauchly’s W: equal variances for RM ANOVA
Research Statistics Final Review
correlation coefficient = ?
(what letter?)
r: correlation coefficient
Research Statistics Final Review
effect size for t-test = ?
Cohen’s d: effect size for t-test
Research Statistics Final Review
effect size for ANOVA = ?
Eta squared: effect size for ANOVA
Research Statistics Final Review
reliability for continuous data = ?
ICC: reliability for continuous data
- Unitless
Research Statistics Final Review
reliability for categorical data = ?
Kappa: reliability for categorical data
- Unitless
Research Statistics Final Review
measure of internal consistency = ?
Cronbach’s alpha: measure of internal consistency
Research Statistics Final Review
Point Estimate
VS.
Confidence Interval
Point Estimate: a single value that represents the best estimate of the population value
Confidence Interval: a range of values that we are confident contains the population parameter
- Width concerns the precision of the estimate
Research Statistics Final Review
Correct interpretation of a 95% CI = ?
Correct interpretation of a 95% CI?
- If we were to repeat sampling many times, 95% of the time our confidence interval would contain the true population mean.
Incorrect interpretation of a 95% CI
- There is a 95% probability that a given measurement falls within a confidence interval.
Research Statistics Final Review
Independent Groups:
t = Difference between means / Variability within groups
What does this mean?
Independent Groups
t = Difference between means / Variability within groups
Difference between means:
- Represents all the possible reasons groups could be different, including treatment effects and error.
Variability within groups:
- Difference explained by error alone
Error = all sources of variability that CANNOT be explained by the IV.
Research Statistics Final Review
Repeated Measures:
t = Mean of difference between pairs / Std. error of the difference scores
What does this mean?
Repeated Measures
t = Mean of difference between pairs / Std. error of the difference scores
Mean of difference between pairs:
- Represents all the possible reasons groups could be different, including treatment effects and error.
Std. error of the difference scores:
- SD (of mean differences) divided by square root of sample size. Variance assumed equal.
Research Statistics Final Review
Null value = 0
- = ?
Confidence Interval Null Values
Null value = 0 = there can be negative values
- Between group or within group differences
- Correlations
Research Statistics Final Review
Null value = 1
- = ?
Confidence Interval Null Values
Null value = 1, there will be no negative values
- Relative Risk
- Odds Ratio
- Likelihood Ratios
Research Statistics Final Review
Normal distribution, Yes = ?
Normal distribution, No = ?
t Test Assumptions
Normal distribution, Yes = parametric
Normal distribution, No = Non-parametric
No multiple comparison for t-Test
Research Statistics Final Review
Equal variances, Yes = ?
Equal variances, No = ?
t Test Assumptions
Equal variances, Yes =
- Independent Groups (levene’s is NOT significant p>.05) = interpret p-value
- Repeated measures = N/A = interpret p-value
Equal variances, No =
- Independent Groups (levene’s IS significant p<.05) = adjust and interpret p-value
- Repeated measures = N/A = interpret p-value
No multiple comparison for t-Test
Research Statistics Final Review
Interval / Ratio DV, Yes = ?
Interval / Ratio DV, No = ?
t Test Assumptions
Interval / Ratio DV, Yes = Parametric
Interval / Ratio DV, No = Non-parametric
No multiple comparison for t-Test
Research Statistics Final Review
Normal distribution, Yes = ?
Normal distribution, No = ?
ANOVA Assumptions
Normal distribution, Yes = Parametric
Normal distribution, No = Non-parametric
Research Statistics Final Review
Equal Variances, Yes = ?
Equal Variances, No = ?
ANOVA Assumptions
Research Statistics Final Review
Interval/Ratio DV, Yes = ?
Interval/Ratio DV, No = ?
ANOVA Assumptions
Research Statistics Final Review
IndependenT Groups
- (a) Fisher’s Least Significant Difference
- (b) Tukey’s honestly Significant Difference
- (c) Bonferroni t-Test
IndependenT Groups
- Tukey’s honestly Significant Difference
Research Statistics Final Review
Repeated MeasureS
- (a) Fisher’s Least Significant Difference
- (b) Sidak’s Multiple Comparison
- (c) Bonferroni t-Test
Repeated MeasureS
- Sidak’s Multiple Comparison
Research Statistics Final Review
Examples of Different Variances
- A: No variance within groups, only between groups
- B: Equal variance within groups; the groups are different
- C: Greater variance within groups; the groups appear less different
- D: Unequal variance within groups
Research Statistics Final Review
Equal variances in independent groups designs = _ test ?
Equal variances in independent groups designs = Levene’s test
- Tests the null hypothesis that “there is no difference in variance between groups”
- Variance is the spread of the scores
- Levene’s test p<.05 we reject the null hypothesis for equal variances
- Variances are NOT equal
- Levene’s test p>.05 we fail to reject the null hypothesis for equal variances
- Variances are equal
Research Statistics Final Review
When to Use (or Not Use) Odds Ratio = ?
When to Use (or Not Use) Odds Ratio
- Used when we want to study risk OR used in case control studies
- Classify participants by outcome
- Odds of exposure by disease (yes/no)
- Can be used in interventional studies to look at the odds improvement after treatment
Research Statistics Final Review
When to Use (or Not Use) Relative Risk = ?
When to Use (or Not Use) Relative Risk
- Used when we want to study risk RR is for cohort studies
- You can study interventions using a cohort study design
- We cannot use RR in case control because outcome has been determine and that is how we are grouping our participants
- RR looks at “incidence”
- Interventional studies can use OR
- Outcome is known and used for grouping, so similar logic to case control/OR applies
Research Statistics Final Review
Best Way to Remember When to Use RR/OR
- Cohort Study = ?
- Case Control = ?
Best Way to Remember When to Use RR/OR
- Cohort Study = The “o’s” in cohort study are really close, kind of like family (i.e. relatives) = Relative Risk
- Case Control = Here we have a, e and o’s. These are different and “at odds” with each other = Odds Ratio
Research Statistics Final Review
RR/OR = 1 = ?
RR/OR = 1 = there is no risk increase or decrease
Research Statistics Final Review
RR = 1.2 = ?
RR = 1.2, those with exposure are 1.2x more likely to develop the disease or exposed people are 20% more likely to develop the disease.
Research Statistics Final Review
RR = 1.6 = ?
RR = 1.6, those with exposure are 1.6x (or 60% more likely) to develop the disease.
Research Statistics Final Review
RR/OR = 1= ?
RR/OR = 1 = there is no risk increase or decrease.
Research Statistics Final Review
OR = 1.2 = ?
OR = 1.2 = the odds of disease are 1.2 times higher (or 20% higher) in those with the exposure
Research Statistics Final Review
OR = 1.6 = ?
OR = 1.6 = the odds of disease are 1.6x (or 60% more likely) in those with the exposure
Research Statistics Final Review
I want to determine if there is a correlation between statistics grades (percentage) and the number of studies published after PT school. = ?
I want to determine if there is a correlation between statistics grades (percentage) and the number of studies published after PT school. = Pearson’s (Parametric)
Research Statistics Final Review
I want to use a 5 point Likert scale to see if there is a correlation between attitudes about research and the number of research studies published after PT school = ?
I want to use a 5 point Likert scale to see if there is a correlation between attitudes about research and the number of research studies published after PT school = Spearman’s (Nonparametric)
Research Statistics Final Review
- Both variables are interval/ratio
- Data meet assumptions
What test = ?
- Both variables are interval/ratio
- Data meet assumptions
What test = Pearson’s r
Research Statistics Final Review
- Both variables are interval/ratio
- Data does NOT meet assumptions
What test = ?
- Both variables are interval/ratio
- Data does NOT meet assumptions
What test = Spearman’s rho (non-parametric)
Research Statistics Final Review
- One or both variables are ordinal
What test = ?
- One or both variables are ordinal
What test = Spearman’s rho (non parametric)
Research Statistics Final Review
- Both variables are nominal
- Measuring risk, yes
- Classified by exposure
What test = ?
- Both variables are nominal
- Measuring risk, yes
- Classified by exposure
What test = relative risk
Research Statistics Final Review
- Both variables are nominal
- Measuring risk, yes
- Classified by outcome
What test = ?
- Both variables are nominal
- Measuring risk, yes
- Classified by outcome
What test = Odds Ratio
Research Statistics Final Review
- Both variables are nominal
- Measuring risk, yes
- Classified by outcome
What test = ?
- Both variables are nominal
- Measuring risk, yes
- Classified by outcome
What test = Odds Ratio
Research Statistics Final Review
- Both variables are nominal
- Measuring risk, no
- Classified by outcome
What test = ?
- Both variables are nominal
- Measuring risk, no
- Data meet assumptions, yes
What test = Chi Square
Research Statistics Final Review
ICC Models
- Model 1 = ?
- Model 2 = ?
- Model 3 = ?
ICC Models:
Model 1
- Raters are chosen from a larger population; some subjects are assessed by different raters (rarely used)
Model 2
- Each subject assessed by the same set of raters, and we want to generalize to other raters. Used for test-retest and inter-rater reliability.
Model 3
- Each subject is assessed by the same set of raters, but the raters represent the only raters of interest.
- Used for intra-rater reliability or inter-rater when you do not wish to generalize the scores to other raters.