Module 2
No description yet. This curated deck is ready for daily study.
Card preview (24)
Define “Null Hypothesis” What does it mean in terms of deviation if data?
Specific statement, interested to reject. The hypothesis being tested. -Any deviation from precited data is due to sampling error
Define “Alternative hypothesis” What does it mean in terms of deviation from data?
Non-specific statement, anything but the null hypothesis outcome. -Anyd aviation from data is not due to sampling error alone
Define “ Two-tailed (sided) hypothesis”
Hypothesis that allows for two possibilities (left or right movement in the distribution)
Define “one-tailed (sided) hypothesis”
Hypothesis that has only one possibility.
What are the steps for hypothesis testing?
• State hypotheses • Compute the tests statistic • Determine P value • Draw conclusion
Define “Test statistic”
Number calculated from data using a statistical test. Allows for null H to be rejected/ fail to be rejected.
Define “Null distribution”
The sampling distribution under the null hypothesis
What is the significance level?
Probability value sd in combination with the p-value to decide whether or not the hypothesis can be rejected (Value under 0.05 in this course)
Define a “Type I error”. What would be a solution? What is the problem with said solution?
Rejecting a true null hypothesis • Solution = reduce P from 0.05 to 0.01 • Problem = increases chances of type II error. • alpha
Define a “Type II error”. What is the relationship between this type of error and power?
Failing to reject a false null hypothesis. • Lower probability of type II error = Higher power • Beta
Define “Power” of a study. What does it tell us about a study?
The probability of rejecting a false null hypothesis. • More power means a more credible study. • more power, higher probability of committing a type 1 error. **Look at formula
Define “P”
Probability of obtaining the null hypothesis aligning data/ probability of obtaining values as extreme/more extreme than the expected values. Probability of obtaining the expected data or more extreme values . P = (sum of observed and more extreme values) x 2 **If bell shaped (null) distribution.
What are some factors that increase the power of a study?
Higher sample size, low variability.
Define a “nonsignificant result”
When the null hypothesis is not rejected. Treat experiment as default.
Define a “Clinical trial”
an experimental study involving humans and at least 2 treatments
Define “Confounding variables”
Variable that masks or distorts the casual relationship between measured variables in the study. ex. Increase in temp (confounding variable) increasing both ice cream consumptions and crime rate, making it seem like ice cram consumption increases crime rates.
How is the correlation caused by confounding variables torn apart? What effect remains?
Random assignment dilutes the correlation. Does not eliminate variation introduced by confounding variable.
Define “experimental artifacts”
Bias introduced by unintended consequences of unnatural conditions in the experimental procedure
How can we deal with experimental artifacts?
Conduct experiments under conditions as natural to the subject as possible
What is an experimental (simultaneous) control group?
Group of subjects subjected to the same technique for assigning treatment as experimental subjects, except they do not receive he actual treatment.
Define “Pseudo-replication” . What are the 2 types?
occurs when independent replicates are not identified correctly. -2 types: time autocorrelated and spatial autocorrelated
What are the 3 elements that reduce bias in an experiment?
• Simultaneous control group • Randomization • Blinding
Define “blinding” and the 2 different types
Concealing information from participants and researchers. types are single blind and double blind.
What is a single blind experiment?
Participants are unaware of their treatment assignment.