Inferential Statistics and Experimental Research Flashcards
INFERENTIAL STATISTICS
INFERENTIAL STATISTICS
Descriptive statistics are used to describe ______ while Inferential statistics are used to make decision about the __________ group based on the information of the ________ group.
- data
- population, sample
In descriptive statistics you use mean (SD) for _________ data and count (percent) for __________ data.
- continuous
- categorical
What is the difference between population and sample?
Population
- contains all subjects of interest
- impractical
- numerical property = parameter (pop mean, proportion)
Sample
- part of selection that s required to be random
- best approach available
- quantity from it = statistic (sample mean, proportion)
Sample selection is required to be ______ but might not be good because random selection doesn’t gaurantee proportional representation of all parts of the population.
random
- The standard deviation of sample means is called what?
- What is the formula for this?
- SEM (standard error of the mean)
- s/sqrt (n) where s=the sample SD and n=sample size
What does the Central Limit Theorem state?
Sampling distribution of the sample means approaches a normal distribution as the sample size gets larger.
The SEM for a smaller sample size is typically _______ than that of a larger sample size. What does this mean?
larger, means there is more variability
What are the 2 common methodologies for inferential statistics?
- Statistical Hypothesis Testing (SHT)
- Confidence Interval (CI)
What are some common uses of inferential statistics?
- estimate population parameters
- compare effects between groups
What are the 4 steps of the procedure for SHT?
- ) State the statistical hypothesis (null and alternative hypothesis)
- ) Select a level of significance (α)
- ) Decide which test to use (t-test, analysis of variance, Mann-Whitney U Test, chi-square test, McNemar test)
- ) Make a decision to reject or retain the null hypothesis based on resulting quantity called p-value
What is the difference between null and alternative hypothesis?
Null Hypothesis
-group means are not different
Alternative Hypothesis
-there is a true difference between groups, and the treatment was effective
What is a p-value?
Major resulting value of running a statistical hypothesis test and quantifies how consistent your sample values are with the null hypothesis.
The p value ranges from 0-1. What does a large p-value mean? What does a small p-value mean?
- Large p-value- your sample values are consistent with the null hypothesis
- Small p-value- your sample values are not consistent with the null hypothesis
For a p-value (α) level <0.05 you _______ the null hypothesis. What does this mean?
- Reject
- This means that the observed difference shows significant effect
For a p-value (α) level >0.05 you _______ the null hypothesis. What does this mean?
- Retain
- This means the observed difference is probably due to chance and is not significant
Difference between a Type I and Type II error?
Type I = rejection of a true null hypothesis
Type II = non-rejection of a false null hypothesis
What is a confidence interval?
A range of values so defined that there is a specified probability that the value of a parameter lies within it.
Confidence Interval Example:
Fitzgerald et al estimated the population mean for lumbar spinal extension for 30-39 year old individuals. Based on the random sample of 42 individuals, they determined that mean=40.0 degrees and sd=8.8 degrees. How can you tell how accurate this estimate is?
- Find the SEM.
- Assess the 95% confidence interval.
SEM
=sd/sqrt (n) = 8.8/sqrt(42) = 1.36
CI
=(40-21.36),(40+21.36)
=37.3 - 42.7
- A narrower CI implies higher ________ with less ___________.
- A wider CI implies lower __________ with increased ___________.
- precision, variability
- precision, coverage (variability)
CI can be used to estimate population parameters, but it can also be used to compare effects between ________.
groups
CI can be used to compare between groups based on the __________ or the ______.
- mean difference
- ratio
The effect is not significant if the CI contains the ___________. The effect is significant if the CI doesn’t contain the _________.
- null value
- null value
- Null Value for the Mean Difference = _____
- Null Value for the Ratio = ______
- 0
- 1
The mean difference of SBP was 10 and its 95% CI was (5.2, 15.7). Is the difference significant?
Yes, CI does not contain null value.
The odds ratio of having not been physically active for patients with breast cancer over the controls was 2 and its 95% CI was (0.5, 4.5). Is the difference significant?
No, Ci contains the null value.
Summary for Inferential Statistics:
- The purpose of using inferential statistics is to make decision about the _________ group based on the information of the _______ group
- Sampling errors are the differences between the ________ values and the _______ values
- Sample mean together with its standard error can picture what the sampling distribution looks like and can provide an interval _______ for the population mean
- ________________ is performed to make a decision about the population by rejecting or retaining the null hypothesis based on the resulting quantity called p-value
- ______ as a result of running a statistical hypothesis testing is used to make decision
- In a statistical hypothesis testing are the potential errors including type _ and type __ errors
- Confidence interval for the mean difference is checked for inclusion of its null value of __ for the significant effect
- Confidence interval for the ratio is checked for including of its null value of __ for the significant effect
- population, sample
- sample, population
- estimate
- SHT (statistical hypothesis testing)
- p-value
- Type I (reject true null), Type II (non-rej of false null)
- 0
- 1
EXPERIMENTAL RESEARCH
EXPERIMENTAL RESEARCH
What is the purpose of experimental research?
Provide a structure for evaluating the cause-and-effect relationship between a set of independent and dependent variables.
The term ________ _____ is often used to describe experimental studies that examine the effect of intervention on human subjects.
Clinical Trial
What are the 4 phases of Clinical Trials?
Phase I
-Show safety on small sample of subjects, help to understand mechanism of action.
Phase II
-Once shown to be safe, this phase demonstrates effectiveness on relatively small samples.
Phase III
-Randomized, usually blind, experiment comparing against a placebo. Large scale studies with 100s or 1000s of subjects.
Phase IV
-Once approved by FDA, this phase involves continual investigation on its effects.
What is considered the gold standard of true experimental designs?
RCT
What is the difference between a True-Experimental Design and a Quasi-Experimental Design?
True-Experimental Design
- subjects are randomly assigned
- at least 2 comparison groups
Quasi-Experimental Design
- does not meet the requirements of a true experiment
- lacking random assignment or comparison groups, or both
- used when a true-experimental design is not feasible
- accommodates for the limitations of natural settings
What is the difference between Between-Subjects Design and Within-Subject Design?
Between-Subjects Design
- subjects randomly assigned to independent groups
- completely randomized designs
Within-Subjects Design
- subjects act as their own control
- repeated measures design
Difference between Single-Factor Design and Multi-Factor Design?
Single-Factor Design (one-way design)
- structures the investigation of one independent variable
- easy to develop
- can’t account for simultaneous interactions of several variables
Multi-Factor Design (multi-way design)
- incorporates two or more independent variables
- can name variously
Pretest-Posttest Control Group Design is the basic structure of a _________ and compares two or more groups that are formed by _______ assignment.
- RCT
- random
What is considered the scientific standard in clinical research for establishing a cause-and-effect relationship?
Pretest-Posttest Control Group Design
When is a Posttest Only Control Group Design used?
When a pretest is impractical or potentially reactive.
_____________ design involves the ability to control the potential influence of individual differences and uses subjects as their own control.
Repeated Measures Design
Repeated Measures Design has a potential for _________ effects and ________ effects. This design is only used when the outcome measure will revert back to baseline between interventions.
- practice
- carryover
What is a Crossovere Design?
Half the subjects receive treatment A followed by B; half receive treatment B followed by A.
Crossover Design is only used when patient’s condition will not change over time and is not a reasonable approach when treatment effects are _____. _______ period is essential.
- slow
- Washout
Mixed Design has both the _______-subjects and _________-subjects factors.
within-subjects and between-subjects
Summary For Experimental Research:
- Experimental research provides a structure for evaluating the ______________ relationship between a set of independent and dependent variables
- In true experimental design subjects are ________ assigned to at least two comparison groups
- When a true-experimental design is not feasible, ____-experimental design is useful lacking random assignment or comparison groups, or both
- In between-subjects design subjects are _______ assigned to independent groups
- In within-subjects design subjects act as their own ________
- By the number of independent variables (=factors), design can be _______-factor or ______-factor
- _______-_______ design compares the changes from pre- to post- test for two or more groups that are formed by random assignment
- When a pretest is either impractical or potentially reactive, _________ only design is useful
- Repeated measures design and crossover design are the types of _______-subject design
- Mixed design is available when you have _____ the within-subjects factors and between-subjects factors
- cause-and-effect
- randomly
- quasi
- randomly
- control
- single-factor or multi-factor
- pretest-posttest
- posttest
- within
- both