Research into statistical validity for the self-reporting test
See all posts in this thread below.
-
-
Christian
Updated: 06-23-2015 01:07 am
-
Questions to find answers to:
- What kind of result would a null hypothesis give?
- What is the alternate hypothesis?
- What kind of result would support the alternate hypothesis?
- What is the probability that a null hypothesis would give a result which would support the alternate hypothesis?
If the probability of the null hypothesis for a result is below 5% then the result is statistically significant and the null hypothesis can be rejected. That doesn’t mean that the alternate hypothesis is true though.
What to do:
- Calculate probability that a null hypothesis would give a trait score.
- Calculate probability that a null hypothesis would give a type score.
How to do it:
1. Calculate probability that a null hypothesis would give a trait score.
- Determine distribution of random answers and trait score. No assumptions.
2. Calculate probability that a null hypothesis would give a type score.
- Determine distribution of random answers and type score. No assumptions.
- Determine distribution of random random trait scores and type score. Assumes: 1. People make 75% correct answers. 2. Questions for a trait is correlated.
- Determine distribution of random trait + inverted trait and type score. Assumes: 1. People make 75% correct answers. 2. Questions for a trait is correlated. 3. Questions for a trait and it’s inverted trait is anti-correlated.
-
-
