Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
46 views2 pages

3.2 - Selecting A Stat Test

This document provides guidance on selecting an appropriate statistical test based on the variables in a study. It outlines a 4-step process: 1) Identify variables, 2) Classify variables as nominal or interval, 3) Make additional distinctions, 4) Choose a test. Variables can be nominal (categorical) or interval (continuous). Common tests include chi-square for 2 nominal variables, t-tests for 1 interval and 1 nominal variable, and ANOVA for comparisons of interval variables across multiple nominal groups. Hints are provided on hypothesis testing, study designs, and key statistics.

Uploaded by

Flower
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
46 views2 pages

3.2 - Selecting A Stat Test

This document provides guidance on selecting an appropriate statistical test based on the variables in a study. It outlines a 4-step process: 1) Identify variables, 2) Classify variables as nominal or interval, 3) Make additional distinctions, 4) Choose a test. Variables can be nominal (categorical) or interval (continuous). Common tests include chi-square for 2 nominal variables, t-tests for 1 interval and 1 nominal variable, and ANOVA for comparisons of interval variables across multiple nominal groups. Hints are provided on hypothesis testing, study designs, and key statistics.

Uploaded by

Flower
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

Selecting a Statistical test:

Select a test sequence:


1) Identify the variables in the study
2) Classify variables as nominal or interval
3) Make additional distinctions
4) Finalize choice of statistical test

Select a test guidance:


1) Read narrative and list the variables
2) Most common: 2 variables
a. One nominal + one interval
b. Two nominals
c. Two intervals
3) Less common: repeated measures on a single group

Nominal: categorical usually 2-5 possible values


- Variable: gender (male or female), blood type, disease state (well or diseased), Dx test result (negative or positive), exposure status
(exposed, unexposed)
- 2 nominal: test for difference in proportion, test for association
- First choice: Chi-square
- N<25: Fisher’s Exact Test
- Gender by SES (2x3)
- Gender by vaccination status (2x2)
- Foodborne outbreak investigation: illness (ill/not ill) by food consumption (ate/did not eat)

Interval: continuous, many possible values across a continuum


- Glucose, weight, calories, IQ, BP, height, etc.
- 2 intervals: test for relationship, test for association or correlation
- Regression: prediction involved
o Treatment to decrease body weight
▪ Predict corresponding decrease in SBP, DBP
▪ Predict corresponding decrease in cholesterol
o Medical college admissions test score
▪ Predict two-year basic sciences G.P.A

One Interval and One Nominal: test for differences in the means of the interval measure for each of the groups defined for the nominal measure
- T-test: one nominal with two groups
o Compare mean height by gender
▪ Height: interval
▪ Gender (male, female): nominal with 2 groups
o Compare mean blood pressure by race
▪ Blood pressure: interval
▪ Race (Asian, White): nominal with 2 groups
- Matched pairs t-test: one nominal with two groups or one identical group measured twice
o Compare blood pressure by treatment
▪ Blood pressure: interval
▪ Treatment status (treated, untreated); nominal with 2 demographically-matched groups
o Compare weight in 2016 and 2017
▪ Weight: interval
▪ Identical group: 2 times (2016, 2017)
- 1-way ANOVA: one nominal with 3 or more groups
o Compare mean height by nation
▪ Height: interval
▪ Nation (Canada, US, Mexico): nominal 3 groups
o Compare mean blood pressure by race
▪ BP: interval
▪ Race (Asian, White, African American): nominal with 3 groups
- 2-way ANOVA: one interval and two nominal measures
o Compare mean height by gender and ethnicity
▪ Height: interval
▪ Gender (male, female): nominal with 2 groups
▪ Ethnicity (Hispanic, non-Hispanic): nominal with 2 groups
- Repeated measures ANOVA: one nominal one interval several different times
o Compare BP in 2016, 2017, 2018
▪ BP: interval
▪ Identical group: 3 times (2016, 2017, 2018)

Hints:
▪ Hypothesis testing:
o Type I error: erroneously claiming a treatment works when it is NOT effective (based on strange sample, flaws in the research)
o Type II error: erroneously failing to claim that a treatment works when the treatment IS effective
▪ What type of observational study design is this:
o Case report
o Case series
o Cross-sectional study
o Case-control study
o Cohort-study
▪ Key statistics:
o Cross sectional study: chi square
o Case-control study: odd ratio
o Cohort study: relative risk

You might also like