Stats week 1

Size of an effect 

number of participants

probability you will correctly reject the null

A small sample shows only larger effects 

A big sample shows smaller effects

. Report means and SD to 1 decimal place, e.g., 35.1

 Report test statistics to 2 decimal places (F, t, R values), e.g., F=3.45 

Report p-values and effect sizes to 3 decimal places 

1) Look at histograms 

2) look at valued of skewness and kurtosis 

3) Look at normality tests

-1.96 <x<1.96 skew or kurtoses 

skewness/SE skewness = z skewness

shapiro wilk greater then 0.05

between participant / between subject 

- can only be in one group i.e smoker non smoker

within participant/ within group  

Each participant can be in multiple groups/ conditions

- weight before and after a marathon 

There was a significant difference in performance on the math and reading tests, t(76) = 3.75, p < .001, d = .427. By examining the mean scores it can be seen that performance was significantly higher on the reading (19.5 ± 3.5) (mean plus SD) compared to the math (17.9 ± 3.4) test.

There was no significant difference in the reading performance of smokers and non-smokers, t(75) = 0.21, p = .831, d = .049.

If Levene’s test is significant, report the results for Welch t test instead (not relevant in this example) 

- assosciation between variables 

There is a significant positive correlation between reading test score and house hold income, r(75) = 0.464, p < .001. As household income increases so too does reading test score.

Note DF= N-2

• There was no significant effect of transport used and household income, F (2,74) = 3.02, p =.055, np2= .075.

• Note. As the effect is not significant you do not report post-hoc tests 

P - pearson correlation  NP - Spearman's Correlation 

P - Independent T Test   NP- Mann Whitney T-test 

P-Dependent T- Test    NP- Wilcoxon Signed- Rank Test 

P- One- way ANOVA   NP- Kruskal Wallis Test 

P - One-way Repeated Measures ANOVA  NP- Friedman's ANOVA