Week 2: Statistical Inference & Correlation: descriptive Flashcards
What is Statistical Inference?
Me words: This is when observation is based on statistics and research that produce an accurate answer across the population.
Real definition:
Statistical inference is based on the scientific method by which data is sampled under controlled conditions. The obtained statistic is then compared to a chance (i.e., no effect) probability distribution of this statistic. If the statistic is sufficiently different to that expected from chance, then it is (statistically) significant
What is Ad Hoc Inference?
My words: This is an observation mad about a population or effect etc based on your own merits which may include your experience, what you have been exposed to.
Example: In mill park there are lots of hills. This may be true to you but may not reflect it comparison to all other suburbs and mill park may have the average amount of hills
Real definition
Ad hoc inference is intuitive, often based on inadequate sampling of data (e.g., casual observation, limited past experience)
Whats the meaning of N = 1
This equation highlights the ad hoc inference. For example, a meta-analysis would have n = 100 onwards. For example, a meta-analysis on a hypothesis that does smoking cause lung cancer. A meta-analysis would suggest this is the case where an ad hoc inference may be that someone has a friend (n=1) smokes a lot and never got cancer. This may be true for this case but among the true population a meta-analysis would be more appropriate to build an understanding from
P value/ Probability value:
- Can show the statistical difference between 2 different means.
- The P value highlights the significance.
- If the p value is less than .05 – this means there is less than a 5% chance that the null hypothesis is true.
- In other words, we are 95% sure that the different between two means are not due to chance.
- In other studies (Particularly medical studies) they put the p value at .01 to be sure
