Statistics for the Social Sciences Quiz #5

when is it appropriate to use chi square test of independence?

to determine whether there is an association/relationship between two categories

what scale of measurement must the variables be?

both must be categorical

what does it mean to say all data must be independent of one another?

each participant/observation can only appear in one cell of the contingency table; no repeated measures or paired date

null hypothesis (H₀)

no relationship; variables are independent

alternative hypothesis (H₁)

there is a relationship; variables are not independent

observed frequencies

actual counts observed in each category in your sample

expected frequencies

the counts you anticipate if there were no relationship between the variables; if H₀ were true

how can you use percentages to describe the trends observed in the data?

convert frequencies into percentages within categories to make trends easier to see

effect size?

how strong the relationship between the variables is

chi square test of independence?

used for 2x2 tables

cramer's v?

used for larger tables; when more than two rows or columns

p < .05

reject H₀; there is a significant relationship

p > .05

retain H₀; there is not a significant relationship

when is it appropriate to consider effect size?

if the relationship is significant (p < .05), report effect size

when is it not appropriate to consider effect size?

if not significant, effect size is not usually emphasized

a description of the context of the study

what is being studied

a description of the trends in the data supported by appropriate percentages

describe with percentages

a statement of the test conducted

chi-squared test of independence

a description of the findings as indicated by the outcome of the test

X2 (df, N) = value, p = value, effect size

supporting statistical notation including, when appropriate effect size

what that means in plain language