Significance Levels
when using a directional hypothesis is that 1 or 2 tailed test
1
Normal Level of Significance used in Psychology
5% (0.05 or 1/20) this means their is less than a 5% likelihood the results occured by chance thus you can probs reject null hypothesis
When is *Mann-Whitney U* used? - Group Design - Level of Measurement
Independant Measures Design Ordinal Data
When is *T-Test related* used? - Group Design - Level of Measurement
Independant Measures Design interval data
When is *Chi Squared (X2)* used? - Group Design - Level of Measurement
Independant Measures Design + Matched Pairs also Relationship Nominal Data
3 Criteria for Parametric statistic Tests
Interval data, intervals between data are truly equal data is drawn from a population that has a normal distribution, (population, not sample) the Variance (SD) of 2 samples isn't significant
Parametric test 3 Pros
More Powerful - they can detect significance in situations where non-parametric tests can't based on actual scores as opposed to ranked No info lost due to ranking
When is *Pearsm's* used? - Group Design - Level of Measurement
Parametric Interval
When is *Spearmans Rank* used? - Group Design - Level of Measurement
Relationship Ordinal Data
When is *Wilcoxon* used? - Group Design - Level of Measurement
Repeated Measures Deisgn, Matched Pairs Ordinal Data
When is *Sign* used? - Group Design - Level of Measurement
Repeated Measures Design, Matched Pairs Nominal Data
Homogeniety of Varience
SD(2) Very very similar calculated with - Range - F-test that tests for differences between 2 sample variances
8 Statistical Tests used
Sign Chi Squared (X2) Wilcoxon Mann Whitney U Spearmans Rank T-test related T-test unrelated Persons R (r)
Parametric Test
Use mean and SD to calculate test statistic are more powerful than non Parametric which used rank data) they can detect significance in situations where non-parametric tests cant
what is a "Significant Result"
a result where their is such a low probability it occured by chance that we decide it's safe to reject the null hypothesis
4 step Procedure for using a Test
calculate Test Statistic (learn this for relevant test) obtain Critical Value (for relevant test) compare calculated test statistic with critical value, if results are sig. at 5% level reject null hypothesis, if not then accept it (read sentence on greater or less than in exam) sentence explaining why significant
How do u test for normal distribution of population
ideally Normally Distributed on distribution curve = mean, median and mode should be roughly the same
How do you know weather calculated value should be greater or less than critical value to be signifant?
if their is a R in the name it should be greater than critical value to be significant
Why is a statistical test used in Psychology
in psychology we always try to reject the null hypothesis and find evidence to support our alternate hypothesis no matter how convincing the results we can Never be certain the null hypothesis is false, theres always a possibility it was due to chance statistical tests show wheather we have a significant result
What are Statistical tests
tests preformed to find out if the results of a study are significant they measure the Strength and Direction of any effect shown by the results they are performed by calculating a Test Statistic from the raw data and comparing it with a Critical Value which is looked up in a table
Explain Relationship between Calculated Value and Critical Value
the Calculated value is the test statistic produced from performing a statistical test on any set of data the critical value is a value the test statistic must reach in order to be deemed as significant the researcher doesn't calculate the crtical value, they do calculate test statistic
What must we do to our 2 hypothesis (null + experimental) if our significance level is >5%
we must retaian Null hypothesis and reject alternate
when may a Significance level of <1% be used?
when we must be especially confident our results weren't chance ie. testing new drug or controverial research