Psych Stats Quiz #3 4/17/23 Flashcards
probability
statistical inference asserts probability that a particular assertation is true
always state conclusions in terms of probability not proof
probability vs relative frequency
future vs past tense
relative frequency is used to express probability
probability ex
cards - the probability of picking a red card from a deck of randomly shuffled cards = 1/2 or 50%
a spade = 1/4 or 25%
random sampling with replacement
after picking an observation you put it back into the pool because if you don’t then the scores are considered independent
Ex. if cards are not replaced the probability changes on subsequent samples
distribution of sample means
establish a set of rules that allows us to make inferences from a sample about the population
- this distribution is the answer to why we can generalize about population based on samples
sampling distribution
a distribution of statistics obtained by selecting all samples of a specific size from the population (usually mean)
mean sampling distribution
a collection of sample means for all possible random samples of a particular size that can be obtained from a particular population
sample mean characteristics
cluster around the population mean
distribution is normal (central limit theorem)
can use distribution to answer probability questions
the expected value of M
the mean of the distribution of sample means which is equal to the population
standard error of sample means
the standard deviation of the distribution of sample means
standard deviation/square root of sample size
σ/√n
standard error characteristics
standard or typical distance from the mean
specifies precisely how well a sample mean estimates the population
hypothesis testing - z test steps
1) state hypothesis about an unknown population mean
2) set criteria for decision between hypotheses
3) collect sample data
4) evaluate null hypothesis
criterion - alpha level
called the level of significance
probability value that is used to define “very unlikely”
p < 0.05
single sample z test: directional
directional - upper and lower 2 1/2 % of normal distribution = -1.96 and 1.96
single sample z test: one-tailed test
1.64, 0.05 = rejection region
reject the null
when same data is substantially different from what the null predicts
fail to reject the null
no evidence was found that the null is wrong, but can never be 100%
type 1 error
null reject when treatment really does not have an effect
false positive
type 2 error
null not reject when treatment does really have an effect
false negative
infant affection holding example
μ = 26 pounds, n = 16, σ = 4
H₀: μinfants held = 26 pounds
H₁ = μinfants held≠ 26 pounds
standard error of mean (σm) = σ/√n = 4/√16 = 1
critical region test:
X = μH₀ (+/-)Z (σm)
X = 26 + 1.96(1) = 27.96
X = 26 - 1.96(1) = 24.04
24.04 ≥ X ≥ 27.96
fail to reject the null because 26 falls in the critical region, meaning that there is no effect on holding infants and their weight
critical region test
X = μH₀ (+/-)Z (σm)
z score
Z(i) = (X - μ)/σ(m)
statistically significant
the null hypothesis was rejected
p-value
the smallest critical region that would allow rejection of the null hypothesis - 0.05
making 0.05 smaller
the rejection region moves out farther
making 0.05 smaller
the rejection region moves out farther