Solutions
Kurtosis can be an indicator of
problematic outliers
Discrete variables
Can only take specific values (usually whole numbers like
students in a class, number of cars in a family)
Continuous variables
Can take all possible values (decimals, fractions, like income
and height)
The only limitation is how precise the researcher wants to be
Bar graphs
Used for categorical/nominal variables
Can easily distinguish categories and make comparisons
Histograms
Used for quantitative variables
Presents variable frequency
Line graphs
Often used to show how one variable changes over time
Effective for demonstrating variable relationships
Pie charts
,Best for a small number of categorical (or ordinal) variables
Visually impactful
Each piece represents the proportion of the sample with a
specific characteristic
Chi-square test for independent variables
Used when you have two categorical and/or ordinal variables
Tests to see if the two variables are independent (null
hypothesis) or if we can conclude there is an association (reject
the null hypothesis if independent)
Independent two-sample t-test
Used when you have a quantitative (ratio/interval) variable and a
binary categorical variable
EX: age in years and history of angina
When the p-value is less than 0.001 we can
Reject the null hypothesis
One-way ANOVA
Use when you have one quantitative variable and one
categorical or ordinal variable with three or more groups
Tukey HSD
Shows us which pairs of groups had significant differences, if
any
Pearson's correlation
Use when you have two quantitative variables
Used when the data is linear, and normally distributed
, Linear regression
Use to model the relationship between two quantitative variables
The outcome variable is continuous or categorical
Research question
Does alcohol consumption increase risky sexual behavior
associated with STI transmission among college students?
Hypothesis
We hypothesize that alcohol consumption will increase the
likelihood of risky sexual behaviors, increasing STI
transmission.
Methods
The survey was created from AUDIT-C and SSBQ, with 27
questions. It was a cross-sectional study, qualtrics and SPSS
were used. Convenience sampling and snowball sampling was
used.
Results/Conclusion
Audit-C and SSBQ scores yielded a negative and weak
correlation, yet statistically significant. We rejected the null
hypothesis. People who scored higher on the AUDIT-C tended
to score lower on the SSBQ.
What does each row in a properly formatted dataset typically
represent?
A. A variable
B. A participant or case