lecture 1 - chi squared Flashcards
what does the shape of the curve on a distribution represent?
scores from the data collected
what does positive kurtosis mean?
there are more scores in a range
what does negative kurtosis mean?
less scores in that range
what is kurtosis?
how many scores fall within the SD
does chi squared require a normal distribution?
no
what does categorical variables mean?
it can be grouped, we form categories based on continuous categories e.g. gender, relationship status, do you have a masters degree? - yes or no
what are numerical observations?
can be placed in order e.g. age, test score, income, height, num of errors made on a memory test
what is continuous data?
can make categories from them e.g. height - it can be anywhere along a range
what is ordinal data?
rank within them e.g. military ranks, where you finish in a race
what is nominal data?
can be assigned to a particular group but there is no ranking within them e.g. religion, gender, race, ethnicity
when do we use chi squared?
if we have categorical data and use counting frequency
when do we use chi squared goodness of fit?
categorical data and frequency counts and one categorical variable
when do we use chi squared independence?
categorical data and frequency counts and 2 or more categorical variables
what is the goodness of fit test?
Does the observed data match what we would expect to see or do the findings deviate from what we expect to see?
what is the equation for testing goodness of fit?
x2 = (observed frequencies - expected frequencies) squared divided by expected frequencies
when do we reject the null hypothesis?
if there is a lack of goodness of fit between the observed and expected frequencies
what goes in the table?
levels of the single categorical variable, observed and expected outcome, calculate diff (observed –expected), diff squared, take diff squared divided by the expected
how do we calculate the degrees of freedom?
num of levels in the variables minus 1
how do we find out if something is sig or not?
1 – determine level of sig and highlight the level youre at
2 – use df – already calculated
3 – read value – this is the chi squared critical value
what do we do if the result is sig?
reject the null – there is not goodness of fit between the expected and observed
what does a right hand tail mean?
very diff from the expected
what does left hand tail mean?
too close to expected so have goodness of fit
what are the restrictions of x2 test
df determined by sie of the table not the sample size, need to be careful with small sample sizes, don’t use x2 if there are fewer than 5 observations
