revsion pp: stats Flashcards
Common Quan Stats
P-Value
Measures probability that observed difference is chance. The lower the score, the less likely due to chance, and so the more likely it is significant (significant = p-value <0.05)
Common Quan Stats
T-Test
- inferential statistic
- the difference between means of 2 sets of data.
- The bigger the score, the more likely the test is repeatable.
Common Quan Stats
Pearson’s correlation
- numerical x 2
- age and BP
Linear relationship - directly impact each other
An interval statistic – measures a linear relationship between 2 variables, where a change in one variable causes a proportional change in the other (e.g. age and BP)
Common Quan Stats
Spearman’s correlation
- categorical (ordinal) + numerical variable
An ordinal or monotonic statistic – the variables change together, but not necessarily at the same rate
Common Quan Stats
ANOVA
categorical (ordinal or nominal) + numerical variable
need at least 3 categories; social media use (low, medium, high)
hours of sleep (6,7,8)
Common Quan Stats
Chi2
Categorical (nominal) x 2
- eg. eye colour and hair colour linked
- A categorical statistic - looking for independence of data (one variable not related to the other)
Common Quan Stats
Cohen’s Kappa coefficient
Measures inter-rater reliability, e.g. performing GCS – assessing if agreement is by chance
1 = complete agreement
-1 = complete unagreement
0 = agreed by chance
Common Quan Stats
Confidence Interval
The confidence you have that the result will lie between a predetermined range of values
Sample and sample Error
There will always be sampling error because the sample will not 100% represent the whole population
Variation & sample size impact confidence interval
Larger variation and small sample size = large confidence interval
Common Quan Stats
Intention to Treat
Including results from participants who do not complete the trial (e.g. died or withdrew) – maintain the original randomised groups of participants
Common Quan Stats
Is a higher confidence interval better?
A large confidence interval suggests that the sample does not provide a precise representation of the population mean, whereas a narrow confidence interval demonstrates a greater degree of precision.
Common Quan Stats
Descriptive statistics -
mean, median, mode, standard deviation – terms used to describe numbers and patterns
Common Quan Stats
Inferential statistics
used to determine the statistical significance of any observed differences between groups
Common stats
What p-value makes something statistically significant
= or less than 0.05
0.05 means 5% by chance
reject null hypothesis (no effect)
accept alternative hypothesis (the independent variable caused an effect)