Data viz, z-scores, & p

Question 1

Box plots: what 4 measures can they show you?

Answer 1

• Box → 25-75%ile
• Center line → median
• Star → mean
• Outliers show up outside of “whiskers”

Question 2

How does a boxplot show you a skewed distribution?

Answer 2

• when the whiskers are not the same lengths

Question 3

What do frq distributions show? (hint: what kind of data)

Answer 3

• Display raw data for ONE scale variable at a time

Question 4

Why use frq distributions? (ie, what do they help us do?)

Answer 4

• Look for patterns
• Check for accurate data entry
• See outliers easily
• Make sure we meet assumptions of tests

Question 5

How are frq distributions / histograms different than bar graphs?

Answer 5

• Frq dist (histogram): ONE scale variable
  → Ex. x-axis: travel time, y-axis frequency
• Bar graph: TWO variables! One scale, one category
  → Ex. x-axis studying yes/no, y-axis exam score

Question 6

Empirical vs theoretical distributions?

Answer 6

• Empirical: distribution of the raw data actually collected from the sample, approx. normal
• Theoretical: distribution based on math and logic, assumes its a normal distribution and okay to use z-scores

Question 7

Normal distribution: what’s it look like? what can we do w/ it?

Answer 7

• Bell-shaped, unimodal, symmetrical
• Allows us to find stats like percentiles, z scores, t tests & p

Question 8

Positive skew: looks like what? Ex?

Answer 8

• Tail points towards the RIGHT, towards the positive end
• Ex: household income, test that was too hard

Question 9

Negative skew: looks like what? Ex?

Answer 9

• Tail points towards the LEFT, towards the negative end
• Ex. test that was too hard

Question 10

Bimodal: looks like what? What does this say about the sample? Ex?

Answer 10

• 2 modes, two bumps
• Usually means you really have two different populations within your sample that have different means
• Ex. scores on an exam could be bimodal for “did” and “didn’t” study groups

Question 11

How does mean compare to median for normal, positive, & negative skewed distributions?

Answer 11

• Normal: M = median
• Positive skew: M > median
• Negative skew: M < median

Question 12

Outliers: what are they? How do they impact M & median?

Answer 12

• An extreme score (very high or low compared to others)
• Can really affect mean, don’t really affect median at all

Question 13

Outliers: How do you identify them in JASP & what we do with them in an analysis?

Answer 13

• Scores that surpass the whiskers on a boxplot
• AKA (1.5 x IQR)
• What do we do → usually remove them, but most often you report the descriptives with and without them

Question 14

Bar graphs: What variables/ levels of measurements do you use them with?

Answer 14

• TWO variables, usually nom/ord (categorical) IV & scale DV

Question 15

Line graphs: What variables/ levels of measurements do you use them with? (2 options)

Answer 15

• BOTH scale (often time on x-axis)

OR

• nom/ord (categorical) IV & scale DV
• ONLY when trying to highlight change

Question 16

Scatterplots: what type of variables/levels of measurement? When do you use them?

Answer 16

• 2 variables
• BOTH scale
• Specific to correlation/regression designs

Question 17

Percentiles: What are they? What do they tell you?

Answer 17

• where you are in a distribution
• 57th percentile = 57% of the data points are BELOW you

Question 18

Z-scores: What are they? What do they tell you?

Answer 18

• Also a measure of you are in a distribution
• Sign shows whether it’s above/below mean
• Value is the distance from mean in SD

Question 19

Formula for z-scores?

Answer 19

• Z = (x-m) / SD
• SAMPLE SD

Question 20

Does converting to z-scores change the distribution of the data?

Answer 20

• NO! The distribution stays exactly the same.
• The value on the x-axis changes, but does not change the shape of the distribution

Question 21

How to calculate percent more extreme?

Answer 21

1. Calculate z-score
2. Use z-score table to find the percentile
3. If necessary, 1- %ile
   → Do this if the % is > 0.5

Question 22

What does probability mean?