1/21
Looks like no tags are added yet.
how do you calculate the degrees of freedom in a chi-square test
based on number of rows and columns
DF=(# of rows -1) x (number of columns -1)
ex: 2 rows and 2 coloms = (2-1)(2-1) =1
ANOVA (F-ratio)
ANOVA stands for Analysis of Variance. It is used to determine whether there is a statistically significant difference between the means of three or more groups.
F ratio = between group variance/ within group variance
what are the multiple levels of ANOVA
one way: one independent variable and one dependent variable
-ex: compare mean anxiety scores for three different treatments
two -way: impacts of two IV (2 or more groups) on one DV(continuous)
-ex: effect of age and gender on anxiety score
repeated measures: one IV and one DV but have the same measure at more than one time point
-ex: same patients anxiety scores at baseline, one month and three months
what are the assumptions for ANOVA
the dependent variable is continuous
the measurements in each population are approximately normally distributed
the measurements in each population have equal variance across the groups
the samples are independent of each other
no extreme outliers
ANOVA components
ANOVA determines whether the differences between groups are large enough compared to the differences within groups to be statistically significant.
what is the sum od squares SS
2. Sum of squares (SS)
Sum of squares measures the total amount of variation by squaring differences from a mean and adding them together.
how do you calculate the degrees of freedom
K= number of groups
N equals total number of participants
Between groups: K-1 (DF1) numerator
within groups: N-K (df2) denominator
DF total =(N-1) df1+ Df2
what are the properties of the F distribution
the F statistic is positively skewed beginning at 0
the larger the F statistic the stronger the evidence against equal means
there are many F distributions and each one has a different shape. we donate these quantities by DF1 and DF2
POST-HOC test
anova tells us that at least one group mean differs but not which group
need to conduct a post hoc test to determine
what are the different post hoc tests
scheffe
bonferroni
others
-least significant difference test
-tutees honestly significant difference
-student Newman-keuls
-tukey’s wholly significant differnce
what happeneds if anova assumptions are not met
Use welch anova if unequal variances
for ordinal datât or non-normal outcomes use kruskal-wallis H-test → non parametric aquivalant of ANOVA
what is the kruskal-wallis H-test
non parametric equivalent of ANOVA
compares ranked scores of 3 or more independent samples that are NOT normally distributed
what are the two types of correlation statistics
pearson correlation coefficient - parametric
spearman rank correlation coeficient - non parametric
what are the assumptions of pearson correlation
level of mesurment: the two variables should be measured at the interval or ratio level
liner relationship: there should exist a linear relationship between the two variables
normality: both variables should be roughly normally distributed
random: data is randomly selected
what is covariance
indicated if two variables tend to increase or decrease together
it is the variance shared between variables
describes how the variables change together (up down or not at all )
provides a measure of strength of correlation
pearson correlation standardize covariance to produce a coefficient between -1 and +1
what does the correlation formula generate
R value
what is the strength of linear association with R value
r value =1 -→ perfect positive liner relationship
r value =0 → no linear relationship
r value = -1 → perfect negative linear relationship
Coefficient of determination
R² → tells you how much of the variation in the dependent variable can be explained by the independent variable.
what does correlation tell you about the relationship
discribes STRENGTH AND DIRECTION
R statistic ranges from -1-+1
it cannot explain causation
correlation does NOT = causation
causation
Causation means that one variable directly influences or causes a change in another variable.
Finding a correlation is the first step, but you need more evidence before concluding that one variable causes the other.
Five things to consider when determining causation
Question | What it means |
|---|---|
Does one occur before the other? | The cause must happen before the effect. |
Is there a strong association? | Are the variables strongly related? |
Is it a plausible relationship? | Does the relationship make logical or biological sense? |
Is it consistent with existing theory? | Does previous knowledge or research support it? |
Are there confounding variables? | Could another variable explain the relationship? |
what is hills 9 criteria of causation
temporal relationship : the cause must happen before the effect
strength of association: The stronger the relationship between two variables, the more evidence there may be for causation.
dose-response relationship : As exposure increases, the risk of disease increases.
consistency: Different studies find the same relationship.
plausibility: The relationship makes biological or scientific sense
Consideration of alternate explanations: Could something else explain the relationship?
Experiment: the condition can be altered by an intervention
specificity: One specific cause produces one specific effect.
coherence: The relationship agrees with what we already know from existing research, scientific knowledge and theories.