Organizing and interpreting data

Página 1

Example:

Treating patients with LAS - ther ex and Mulligan LAS tx

Track outcomes immediately after one tx

Total sample size is 8; alpha is 0.05 (essentially false positive rate, chance of type I error) ; beta is 0.75 (chance of type II error, false negative rate)

Effect size is found to be 1.2 (small is 0.2, medium is 0.5, large is 0.8)

because it's large that means our results are meaningful

Power is 0.25 (we ideally want Power at 0.8; rejecting null hypothesis when it is false)

to increase power, we can increase n and increase alpha and increase effect size (if set before hand), use more reliable measures

p value was 0.42 which means the results were not statistically significant

Accept null hypothesis Reject null hypothesis

Null true Correct decision (1-a) Type I error (a level)

Null false Type II error (B level) Correct decision (1-B)

Página 2

Distribution tables

rand order

Helpful when data is spread/varies and for small data sets

Simple frequency

can summarize larger data sets similar to rank order, but may not be easily understood as other data visualizations

grouped frequency

can summarize very large data sets like the SF

Groupings must make sense for your data

not ideal for data sets that are small or too spread out

Graphs

histograms

frequency polygon