Term
|
Definition
- distribution is known/ can be estimated
- assume there is no effect or no difference |
|
|
Term
|
Definition
| there is an effect or difference |
|
|
Term
|
Definition
| no reason to think that hypothesis is wrong, no evidence that treatment has no effect |
|
|
Term
| why fail to reject and not accept? |
|
Definition
o It is tradition
o Decision is based on just 1 data
o Claim can be disproven with one example, but even 1 million examples in favour, cannot prove a claim |
|
|
Term
|
Definition
| the null hypothesis is true, but you reject it (alpha) |
|
|
Term
|
Definition
| you fail to reject the null hypothesis, but it is false (beta) |
|
|
Term
|
Definition
0.05 (split into 2 = 0.025)
- Use when not sure if data will increase or decrease (can move either way) |
|
|
Term
|
Definition
0.05
- One sided movement
- More likely to reject |
|
|
Term
|
Definition
| measure the size of effect or difference, without influence of sample size |
|
|
Term
|
Definition
probability of reject the null hypothesis when it is false
- goal of a lot of studies
- generally want it to be 0.8 |
|
|
Term
| factors that effect power |
|
Definition
- increasing the effect size, increases power
- decreasing variability increases power
- decreases standard deviation or increasing n increases power
- increasing standard error increases power
- increase alpha (type 1 error) increases power |
|
|
Term
|
Definition
- paired samples, matched samples, dependent samples, repeated measures, within subject
- 2 samples
- Pre-test vs post-test
- Performance of same group of participants at 2 different times/conditions
o Ex. Measure on midterm, measure on final
o Related samples = ex. Husband and wife |
|
|
Term
|
Definition
- Between subjects
- 2 samples that are unrelated (example PSYB07 this year and last year)
- Variance Sum Law = variance of a sum/difference of 2 independent variables is equal to the sum of their variables
- Ho = no difference between last year and this year (ulast – uthis = 0 or ulast = uthis) |
|
|
Term
|
Definition
| - Measures how well the data fits expected values of the model/hypothesis |
|
|
Term
|
Definition
how good does the data fit expected value of model/hypothesis
- If data deviates from expected values to a large extent, we reject the model/hypothesis
- Expected values are generated with the assumption that variables are independent |
|
|
Term
| chi square test assumptions |
|
Definition
- Independent observations
- Categories of each variable are mutually exclusive and exhaustive
- Sufficient count for each cell
o > 5 for each cell |
|
|
Term
|
Definition
o Degree of relationship between 2 measurement data
o Linear relationship
o More about continuous data
o Quantify how strong and the direction of relationship
o Line of best fit/regression line |
|
|
Term
|
Definition
- How closely the data clusters around the regression line
- Represents strength and direction of linear relationship |
|
|
Term
|
Definition
measure of how much 2 variables vary together
- It is a function of the standard deviation of the variables -> r controls for that |
|
|
Term
|
Definition
o not an unbiased estimator
o tends to be a little bigger than it should
o smaller smaple = greater bias |
|
|
Term
|
Definition
- regression line overestimates sometimes
- only can make predictions only in the range of the data given |
|
|
Term
| standard error of estimate |
|
Definition
- data doesn’t fall on line; assuming an error
- standard error of estimate: how much data deviates from the regression line
- it is like standard deviation, but around regression line (predicted scores) rather than the mean |
|
|
