Data

Question 1

what are the three characteristics of normality?

Answer 1

• Unimodal
  
  • Symmetrical
  • No sharp cut-offs

Question 2

what is a skew?

Answer 2

• Measure of lateral deviation from morality
  Skewed distributions are not symmetrical
  #

Question 3

what is a negative skew?

Answer 3

• Long tail is on negative side of the peak;
  
  • Also described as skewed to the left
  • Mean is on the right side of peak.

Question 4

what is a positive skew?

Answer 4

• Long tail is on the positive side of the peak
  
  • Mean mean is also on the right of the peak
  • Also described as skewed to the right.

Question 5

what is kurtosis?

Answer 5

• Measure of vertical deviation from normality
  
  • Leptokurtic - curve is more peaked than normal distribution
  • Platykurtic - curve is flatter than normal distribution curve.

Question 6

what do parametric stats tests assume?

Answer 6

• Dv is interval or ratio
  
  • Observations are independent from each other
  • Data follows normal distribution
  • Samples obtained are from populations with same variance (levene’s test)

Question 7

what are non-parametric tests?

Answer 7

Do not make assumptions about distribution
- Are less powerful and may fail to detect differences
- Ideal for categorical and ranked scales or where data is normally distributed.

Question 8

what is hypothesis testing?

Answer 8

• Null hypothesis
  States there will be no relationship

Question 9

what is the purpose of statistical testing?

Answer 9

• The aim of statistical testing is to determine whether we can accept or reject our null hypothesis
  
  • HOWEVER, our null hypothesis is never proven.

Question 10

what is the difference between type 1 and type 2 errors?

Answer 10

Type 1 error = reject H0 = null hypothesis - when it is actually true (false positive)
Type 2 error = fail to reject H0 = null hypothesis when it is negative (false negative).

Question 11

what happens as significance level increases?

Answer 11

the less confident we are in our findings.

Question 12

what is effect size?

Answer 12

• Tell us how ‘large’ our effect is
  
  • This Is important as there is a greater likelihood of achieving statistical significance with larger sample size.
  • BUT statistically significant change could only cause ‘small’ effect.
    Partial eta squared = commonly used.