Hypothesis and different stats Flashcards
What does it mean when we say that the sample mean is plausible (logical/credible)?
It means that it lies within a normal distribution/ normally distributed. If not plausible, then the null hypothesis is false.
What is central limit theorem?
Sampling distribution of the sample means approaches a normal distribution as the same gets larger and these means will be distributed around population mean
Estimation of where the sample mean falls relative to population mean depends on?
- Sample size
2. an estimate of the population variance
The smaller the p-value, the more statistical evidence exists to support the alternative hypothesis.
Please name the range.
- less than 1 % - overwhelming evidence
- 1- 5% - strong evidence
- 5 - 10% - a trend / weak evidence
- Over 10% - no evidence to support H1
What is Type 1 and 2 error?
Type 1 = accepting H1 (rejecting H0) wrongly : false positive
Type 2 = accepting H0 (rejecting H1) wrongly: false negative
Z value and rejection region
The region where u can use to reject H0
What are the criticisms to H0?
Significant effect does not mean big effect size
ADvantages of effect size?
Allows researchers to come to a common metric for evaluating a divers experiment.
How does cohen’s D measure the mean difference?
using SD
Size of effect for Cohen’s D? and for Pearson’s correlation?
- Cohen’s D : 2, 5, 8
2. 1, 3, 5
What is power of a hypothesis test?
The power of a hypothesis test is defined as the probability that the test will reject the null hypothesis when the treatment
does have an effect (rule of thumb: minimum of 80%).
What does power of test depends on?
Alpha level, sample size and effect size
When H0 is true and u reject = alpha
when H1 is true and u accept H0 = beta
when H1 is true and u reject H0 = 1 - beta
NIL
What is SE (and therefore SEM)
SE shows how accurate ur sample mean is compared to population mean.
Principles of Central limit theorem?
- Sample mean (s) = population mean
- SD of sample mean = SE
- For large sample , mean distribution is normal even if the distribution of samples drawn from this population is not normal.
How do you calculate the probability of observation 1 and observation 2?
P (A and B) = P(A) x P (B)
Imagine all 4 cups are right
• What is the probability of this outcome?
• What is the probability of getting the first cup right?
• P (4 correct guesses) =
4/8 x 3/7 x 2/6 x 1/5 = 0.014
P values assume that H0 is true and it allows statement of the data and not theory
NIL
a p value of < 0.5 doesn’t mean no effect, it could be that that the effect is too small to be observed
NIL
What is power of 80% means?
80% of the chances to detect a true effect or 80% chances p value falls under 0.05
What does p value distribution depends on?
Statistical power (power to detect true effect) and power depends on sample size
Sensitivity and specificity in terms of H0 and H1?
Sensitivity - Reject H0 when H0 is false
Specificity - Accept H0 when H0 is true
What is probability as propensity?
Comparing different likelihoods of outcome given a current data OR
tendency of a given type of physical situation to yield an outcome of a certain kind, or to yield a long run relative frequency of such an outcome
what is probability in Bayesian analyses?
Strong prior beliefs, see new things update belief but new data is always actively being compared to prior belief.
The more tests we do with only ONE data set (not replication), more errors and have to correct using Multiple Comparison
NIL
