exam 4 Flashcards
inferential statistics
used to determine whether results match what would happen if we were to construct experiment again and again
(basically testing to see if sample means reflect a true difference in population means)
formal term for null hypothesis
Ho
goal of inferential statistics
determine chance (probability) of getting a particular outcome (sample) if our assumption about the world (null hypothesis) is true.
what do you do if probability is too low
reject null hypothesis
how to calculate probability
p=frequency/all outcomes-
null hypothesis
population means are equal-observed differences is due to random error
“there is no change, difference or relationship in the general population”
research hypothesis
population means are not equal
binomial distribution
likelihood that a value will take one of two independent values under a given set of parameters or assumptions. The underlying assumptions of the binomial distribution are that there is only one outcome for each trial, that each trial has the same probability of success, and that each trial is mutually exclusive, or independent of each other.
in a coinflip test, the binomial distribution represents the..
null hypothesis
possible outcomes=
population
probability of getting a particular outcome
sample
if the probability of your “sample” is high
enough given your null hypothesis…(“population”),
“fail to reject the null hypothesis”
if sample is too unlikely you..
“reject null hypothesis”
why do we say fail to reject null
Hypotheses can be rejected with certainty,
but the correct single one may never be identified…
research hypothesis in formal terms
H1
statistical significance
a significant result when the outcome has a very low probability of occurring if the population means were equal
what is the probability required for significance called
alpha level
alpha level
the probability required for significance
research hypothesis is also called..
alternative hypothesis
how to specify null hypothesis formally
u1=u2
u1=0.5
u1=u2=u3=u4..(iv has more than two levels)
how to specify if researcy hypothesis is correct in formal terms
u1 =/= u2
u1=/=0.5
u1>0.5, etc
“at least one differs from the rest’
sampling distributions are based on..
the assumption that the null hypothesis is true
deals with probability charts?
p value
actual probability of result if null hypothesis was true
importance of sampling size
as sample size increases, you’re more likely to obtain an accurate estimate of the true population value
Type I error
rejecting the null hypothesis (H0) when it is actually true
Type II error
failing to reject (or, accepting)
H0 when an alternative hypothesis (H1) is
actually true.
t test
used to examine weather two groups are statistically different to each other.
reflects all possible outcomes we could expect if we compare the means of two groups and null hypothesis true
what is the t value
ratio of two aspects of data, the difference between means and the variability within groups
group difference
the variance between obtained means.
(deals with t test)
under null hypothesis, you expect this to be 0
the value of t increases as…
the difference between your obtained sample means increases
degrees of freedom
(df) its the total number of participants minus the number of groups
what determines if something is a one or two tailed test?
depends on whether you origionally designed your study to test a directional hypothesis
what do you do if you cant calculate the probability of an outcome (step 2 of inferential statistics)
convert value to a statistic (eg a score/measure/etc) then determine the probability of getting that statistic
(USE Z SCORE! Z SCORE IS THE STATISTIC)
statistic
a score/measure/etc
statistical vs practical significance
statistical significance- whether an effect exist
practical-refers to the magnitude of the effect
another name for significance level
alpha level
probability of making a type one error is determined by..
determined by choice of significance or alpha level
probability of type 2 error is determined by..
alpha level
sample size
effect size
sampling distribution
other possible statistics of the same size
The Central Limit Theorem
“When variables are added (such as w/ means), the distribution of these new values (e.g., 𝑋തs) will
approach a normal distribution even if the original variables themselves are not normally distributed”
the central limit theorem suggest that..
with a large enough sample size, almost every sampling distribution becomes normal
standard error
average distance between a sample mean and a population mean
standard deviation of the distribution of sample mean and pop mean
factorial design
an experiment with more than one independent variable (also called factor)
main effects
the effect one IV has by itself on the DV in a factorial design
how do you know when theres interaction
when effect of one IV on the DV depends on the particular level of the other IV
format for factorial designs
levels in first IV X #levels in 2nd IV Xnumber of levels in 3rd IV
The ____ the t (or z) ratio, the more likely the results are ___
Larger
Significant
R squared value
Represents the percentage of variability in u that can be accounted for in x
A criterion variable is generally analogy’s to an ___variable in the experimental method
Dependent
Precictor variable
Variable used in regression to predict another variable
Predicts outcome of critertion. Can also be thought of as dependent variable
John finds Pearson r coefficient between variable a and b to be 0. Before concluding there is no relationship, what should he do?
Construct a scatter plot of the data
T test and Pearson correlation are used to test what kind of data?
Interval
Multiple correlation is a technique for
Increasing accuracy of prediction
Ways to commit a Type two error
Alpha is too low
Sample size is too small small effect/small effect size
Experimenter error such as bad instructions or weak manipulation
Main effect
The effect that one independent variable has by itself on the dependent variable
The mean difference between levels of one factor are called the____
Main effects of that factor
What do you plot to determine if there’s a main effect between interactions
Cell means
What would a plotted data point look like if there’s a main effect
The data point won’t overlap
Overall differences in bar heights
How would you determine if there’s no main effect
If plots overlap
Interaction
When effect of one independent variable on the dependent variable depend on the particular levels of the other independent variables
How to determine if there’s NO interaction between factors
Parallel lines or no changes in bar patterns
R squared
Amount of variance explained
Percent of shared variance between two variables
What is y prime
The predicted value of y (critertion variable)
Y
The critertion variable
X
Predictor variable
A
The y value at the point where the regression line crosses the y axis
The value of when x=0
B
Slope of the line
Related to but not the same as r
Restricted range
Range of values that has been condensed or shortened
It makes the correlation misleading
A correlation coefficient is always used to
Describe the strength and relationship between two variables
Multiple correlation uses many ______ variables and one ___\\
Many predictor
One criterion
Effect size
quantitative measure of the magnitude of the experimenter effect.