1/47
Looks like no tags are added yet.
when should planning for data analysis begin
during the research design, before collecting data
what are the six considerations when choosing a statistical test
data type
measurement level
distribution
sample relationship
purpose
number of groups
what is data cleaning
identifiying and addressing missing or incorrect, inconsistent or unusual
why can data collected at one point in time be limited
it provides a snapshot and may not reflect changes over time
should outliers be removed immediately
no → investigate their cause and how keeping or removing them affects conclusions
what is data coding
assigning labels or numarical codes to data to facilitate analysis
what is illogical value
an impossible or unreasonable value such as a negative age
what is generalizability
the extent to which findings apply to a wider population
what is reproducibility
the ability to obtain consistent results using the same data methods
what is descriptive statistics
statistics that summarize and describe collected data
what is inferential statistics
statistics that use sample data to draw conclusions about a larger population
what is binary data
categorical data with two possible outcomes
such as
infection: yes or no
what is discrete numerical data
countable values such as the number of paitent falls
what’s a parametric test?
a statistical test that assumes your data follows a specific shape or distribution—most often a normal (bell-shaped) curve—and uses parameters like the mean and standard deviation to analyze it
more powerful → less chance of making a type 2 error
can draw more conclusions
poor with small sample size → below 30
typically ratio and interval level
what levels of measurement typically use parametric tests
interval and ratio
what are non-parametric tests
tests that do not require the same distributional assumptions as parametric tests
useful for ordinal data, skewed data or small samples with uncertain distribution
less power → use less info
more chance of a type 2 error
good for small sample size
typically ordinal or nominal
less effected by outliers
which tests generally have greater power when assumptions are met
parametric tests
what tests are less affected by outliers
rank based
non-parametric tests
what is homogeneity of variance
assumption that groups have similar population variances
what is statistical power
the probability of detecting a real effect
what statistical test do you use to measure a binary outcome
chi-square
logistic regression
what statistical test do you use for a likertscale
frequencies, median, rank
what statistical test do you use for continuous data
T tests
anova
what are the equivalent non parametric statistical test for a Independent t-test
Mann -whitney U test
what are the equivalent non parametric statistical test for a pair sample t test
Wilcoxon signed rank test
what are the equivalent non parametric statistical test for a one way independent ANOVA
kruskal Wallis Test
what are the equivalent non parametric statistical test for a person correlation coefficient or regression
Spearman’s rank order correlation
why do researchers match participants
to improve comparability between groups on selected characteristics
which distribution matters for a paired t-test
the distribution of the differences between paired observations
what test compares two independent groups with normally distributed numerical outcomes
independent samples t-test
what test compares two independent groups with non normal ordinal outcomes
Mann- whitney U test
which test compares two related measurements with normally distributed differences
paired samples t-test
which test compares two related measurements with suitable non normal or ordinal data
Wilcoxon signed rank test
which test compares three or more independent groups with normal numerical outcomes
one way independent ANOVA
what test compares three or more independent groups with non normal or ordinal outcomes
kurskal-wallis test
which test compares three or more related measurements with normal outcomes
repeated measures ANOVA
which test compares three or more related measures with non normal or ordinal outcomes
Friedmans test
what is the null hypothesis for an independent t-test
the two population means are equal
what is the null hypothesis for ANOVA
all population group means are equal
what is the alternative hypothesis for ANOVA
at least one population group mean differs
what does the ANOVA F-ratio compare
between-group variability relative to within-group variability
does a significant ANOVA identify exactly what group differs
NO
post-hoc comparisons are needed
what test examines an association between categorical variables
chi-square test of independence
what test measures a linear relationship between two suitable numerical values
pearson correlation
what test measures a monotonic(moving in a single direction) relationship using ranked data
Spearman rank order correlation
what is the difference between correlation and regression
correlation: measures association
regression: models or predicts an outcome
what type of regression predicts a continuous outcome?
linear regression
what type of regression predicts a binary outcome
binary logistic regression