STATS WAVIER

Independent Variable

Thing you change or group people by

Dependent Variable

Outcome you’re measuring

Types/Levels of Measurement

• Nominal

• Ordinal

• Interval

• Ratio

Nominal

names and categories (no order)

examples: nike, adidas

Ordinal

ordered but uneven spacing

examples: rankings 1st, 2nd, and 3rd

Interval

equal gaps and no true 0

examples: temperature in F

Ratio

interval and true 0

examples: time spent and income

Descriptive stats

mean, median, mode

Mean

average

Median

middle value

Mode

most frequently seen value

Standard Deviation

how far data is spread

Normal Curve

bell shaped and most scores fall near average

Chi Square Test

use when you have 2 categorical variables

Chi Square Test SPSS output

• look at p value

• if p<.05 variables are significantly related

• use cramers v to measure strength 0 to 1

Paired T-Test

same group at 2 time points

(pre/post tests)

Independent Samples T-Test

2 separate groups

T-Test SPSS output

• means for each group/time

• sig (2 tailed) → if p<.05 difference is significant

ANOVA

analysis of variance

3 or more groups

ANOVA SPSS output

• look for F value

• sig (p<.05 = groups are different)

Pearsons Correlation

tests relationships between 2 scale variables to determine strength and association

r=0

no correlation

r>0

positive correlation

r<0

negative correlation

Regression

predict outcome using 1 or more variables

Linear regression

1 predictor

Multiple regression

2+ predictors

Regression SPSS output

R2 = % of variation explained

B= slope (direction and strength)

sig = if predictor matters

Ethical Questions

• Are authors being transparent

• Who benefits from data

• Who is harmed by the data

• Could date be manipulated to support one group and erase another?

Good Use for stats

• support fair policy (access and justice)

• identify health disparities

Harmful use for stats

• cherrypicking data

• ignore underrepresented groups

• misrepresent who benefits

Chi Square IV and DV

IV = nominal categories

DV= nominal categories

Independent T-Test IV and DV

IV= nominal (2 group)

DV= scale (interval or ratio)

Paired T-Test IV and DV

IV= time 1 vs time 2 (same group)

DV= scale (interval or ratio)

ANOVA IV and DV

IV= nominal (3+ groups)

DV= scale (interval or ratio)

Correlation IV and DV

BOTH= scale (interval or ratio)

Regression IV and DV

BOTH= scale (interval or ratio)

EPSEM

equal probability of selection method (same chance of being selected)

descriptive stats

summarize, organize, and describe

descriptive stats categorical

nominal and ordinal → counts, percentages, mode

descriptive stats continuous

interval and ratio → mean, median, mode, SD, min/max, range, variance

measures of association

are 2 variables related/how strong?

Cross tabulation

table shows relationship between 2 or more categorical variables

Linear Regression

predict 1 outcome 1 variable

Multiple regression

predict 1 outcome using multiple variables

How to use regression equation to predict DV

simple linear y=b0 + b1X

Cramers V/Phi

Categorical

use - both variables nominal (2 or more categories)

Phi → use both variables binary (2 options each) (ranges 0 no association to 1 perfect association)