Basic statistics

Question 1

What are the different levels of measurement?

Answer 1

• Nominal - act as names or labels for groupings
• Ordinal - Assign groupings/numbers so they can be ranked relative to others
• Interval - Ordinality is satisfied but the differences between numbers are meaningful
• Ratio - Same as interval scales but now zero really does mean a complete absence of the attribute

Question 2

What’s an age/grade score?

Answer 2

• Converting raw scores so that they can be equal to the mean for all persons
• Useful for educational purposes
• Not useful for research

Question 3

What’s variability?

Answer 3

• The degree to which individuals differ
• More variability = more uncertainty in your sample

Question 4

Can you have variation with a dichotomous distribution?

Answer 4

• Yes
• There’s a special formula

Question 5

Transformation of distribution and variance?

Answer 5

• When a constant is added/subtracted from all the scores in a distribution
• Will move the distribution along the x axis

Question 6

What happens to the distribution when a constant is multiplied?

Answer 6

• It will widen the shape of the distribution

Question 7

T/F: The variance of aggregated scores will always be smaller if the two items are actually correlated.

Answer 7

• FALSE
• The aggregated scores will have a higher variance if the scores are actually correlated

Question 8

Covariance?

Answer 8

• The average cross-product of two sets of deviation scores (raw scores)

Question 9

Correlation?

Answer 9

• The average cross-product of two sets of z-scores

Question 10

T/F: Variance cannot be negative

Answer 10

• TRUE

Question 11

T/F: Cannot have a negative covariance.

Answer 11

• FALSE
• Yes you can

Question 12

Can you convert covariation to correlation?

Answer 12

• No you need more information (i.e., the standard deviations)

Question 13

How can you get standardized aggregated variance?

Answer 13

• Add up the values in the correlation matrix

Question 14

What’s effect size (d)?

Answer 14

• The size of the difference between 2 group means, in SD units
• If distribution is normal, a d statistic of a given size represents different amount of overlap between distribution

Question 15

T/F: No point in having a cohen’s d of continuous variables

Answer 15

• TRUE
• Will end up having to group the variables together

Question 15

What does d = 1 signify?

Answer 15

• 1 SD difference between the two groups mean

Question 16

What are the general areas for Cohen’s d magnitudes?

Answer 16

• Small - 0.1-0.2
• Medium - 0.4-0.5
• Large - 0.7-0.8