W2: Practical Flashcards
What does this show?
drug 1 was scoring better than the placebo group and drug 2 was scoring additional benefit of this new drug
For interaction effect a
clustered bar chart is better
What is an interaction effect?
An interaction effect occurs when the effect of one variable depends on the value of another variable.
- Line graphs can be quite useful when you got
time series data (measurements over many time points)
Example of interaction effect - (3)
- So for drug 1, this seemed to be as effective as drug 2 for early onset Alzheimer
- Drug 1 not very effective for late onset Alzhiemer and not much difference between drug 1 and placebo for that group
- Whereas, drug 2 seems to be more effective for both types of early and late onset Alzheimer’s
What are z-scores?
A measure of variability: The number of standard deviations from the population mean or a particular data point is
Z scores are a standardised measure and ignore
measurement units
Why should I care about Z scores? - (2)
Z-scores allow researchers to calculate the probability of a score occurring within a standard normal distribution
Enables us to compare two scores that are from different samples (which may have different means and standard deviations)
How to read positive z score table to get percentile? - (2)
first colum contains first part of z score (whole number and decimal point)
top row contains remaining deicmal point
How to read positive z score table to get percentile? example- (2)
if z score is 1.25 then.. look left column for 1.2 and top row for 0.05
If trish takes a test and gets a score of 25 and shows her z score is 1.25 and percentle is 0.8944 it shows that
89.4% of students performed worse than Trish
Who performed better Trish or Josh?
89.4% students performed worse than Trish
84.1% students performed worse than Josh
Trish
68% of scores are within
1 SD of mean
95% of scores are within
2 SDs of mean
99.7% of scores are within
3 SDs of the mean
narrow CIs indicate higher
power
wider CIs indicate
low statistical power (bad).
If CIs overlap shows
two means not significantly different
If CIs do not overlap it shows
two means are significantly different
Null hypothesis is typically a hypothesis of
no difference (0)
We assume the null hypothesis is
true
We collect evidence to REJECT the
null hypothesis
We can never say that the null hypothesis is
FALSE
TheP valueor calculated probability is the estimated probability of us
finding an effect when the null hypothesis (H0) is true.
p value equals to
probability of observing a test statistic at least as a big as the one we have if the null hypothesis were true.
Statistical significance does not equal
importance
The reason why statistical significance does not equal importance due to 2 reasons - (2)
- p = 0.049, p = 0.050 are essentially same and former is statistically sig
- Importance is dependent upon exp design/aims
Statistical Sig does not equal importance as importance dependent upon experimental design/aims - example
A statistically significant weight increase of 0.1Kg between two adults experimental groups may be less important than the same increase between two groups of babies.
