stats tests Flashcards
what is the null hypothesis
a hypothesis against a research question
claims no difference in result and difference is an error
what is the p- value
the probability that the null hypothesis is correct
what is an alpha-level
threshold level for a p-value usually 0.05 so is p-value is <0.05 reject null hypothesis if >0.05 accept null hypothesis
what is a type 1 error
false positive
reject null hypothesis when true
what is a type 2 error
false negative
accept null hypothesis when false
what kind of data is used in a binomial test
dichotomous, yes/ no
What are the three types of binomial test
one-tailed where observed is < expected
(cumulative probability 0 to observed)
one-tailed where observed is > expected
(1- cumulative probability from observed max)
two-tailed cumulative probability distance from mean
what is confidence level
a range of plausible values associated with confidence level (95%)
when do you use a chi-square goodness of fit test
proportions with more than two levels
when should chi-square test of association be used
comparing proportion across two or more groups
when should McNemar’s test be used
when there are two dichotomous variables e.g. 2by2 contingency table and paired samples measuring proportions
When should a one-sample t-test be used
comparing a measure to a fixed value
when should a two sample t-test be used
when comparing a measure across two groups independently
when should a paired t-test be used
when data can be paired between to groups and are comparing a measure
what is the difference between binomial test and chi squared goodness-of-fit
both compare proportions however chi-square can have more than two categories (not dichotomous)
What is chi-squared test of association also known as
test of independence
what are t-tests used for
measure difference in groups of measure
compare means of populations
what kind of data is used for t-tests
interval or ratio
what nominal test does one-sample t-test correspond to
binomial
chi-square goodness of fit
What nominal test does independent (unpaired) sample t-test correspond to
chi-square test of association
what nominal test does paired samples t-test correspond to
McNemar’s test
what does one sample t-test compare
compares mean of one sample group to a fixed value
What do independent samples t-test compare
observed difference between the means of two independent samples or categories
What doe paired sample t-test compare
mean difference of one group measure on two occasions
What assumptions are made by t-tests
normality
equality of variance
What is central limit theorem
one of the most important theorems in data science
if you take n sample from distribution and calculate means will still be normally distributed
shows sample is large enough
what are parametric tests
statistical tests based on an assumption of normality
how is normality tested
Shapiro-Wilk
Indicated by a low p-value
what test should be used if variance isn’t equal
Welch’s test
what is the t-score dependent on
degree of freedom and sample size
what does a greater t score indicate
greater difference
how is a one sample t-test result reported
mean, standard deviation then, t(x) = y, p= </> .z
how is independent samples t-test reported
mean, SD, t(x) = y, p= z
how is paired samples t-test reported
mean and SD, t(x) = y, p= z
