We’re ready to TEST our Research Questions!

In science, how do we usually test a hypothesis?

EXPERIMENTAL OBSERVATIONALA study is set up to determine the effect of changing one or more variables on an outcome, while all other variables are held constant

Risk factors (variables) are under the direct control of the investigator

A study is created in which observations of subjects and measurements of variables are made without changing any variables

Risk factors (variables) are NOT under direct control of investigator

Was our ACT passage today EXPERIMENTAL or OBSERVATIONAL?

Is EPIDEMIOLOGY an EXPERIMENTAL or OBSERVATIONAL study?

Example: Suppose we want to study the effect of smoking on lung capacity in women…

EXPERIMENTAL OBSERVATIONAL1. Find 100 women age 20 who do not currently smoke.2. Randomly assign 50 of the 100 women to the smoking treatment and the other 50 to the no smoking treatment.3. Those in the smoking group smoke a pack a day for 10 years while those in the control group remain smoke free for 10 years.4. Measure lung capacity for each of the 100 women.5. Analyze, interpret, and draw conclusions from data.

1. Find 100 women age 30 of which 50 have been smoking a pack a day for 10 years while the other 50 have been smoke free for 10 years.2. Measure lung capacity for each of the 100 women.3. Analyze, interpret, and draw conclusions from data.

1. What might be a problem here?2. Which type of study would be more reliable?

Why can’t we always do experiments if they are more reliable?

• Ethics– Ex: Trying to determine if abortion causes breast cancer– Why experimental design wouldn’t work: It’s unethical to ask

subject to have abortions • Lack of Control

– Ex: Trying to determine if smoking bans reduce lung cancer rates– Why experimental design wouldn’t work: Researchers can’t control

laws and policies • Impractical

– Ex: Trying to determine if a certain medicine causes rare symptoms– Why experimental design wouldn’t work: Population size would

have to be extremely large

Steps to Testing your Research Question:

1) Determine whether primary or secondary data is best for answering your research question

2) Choose a study design3) Collect & analyze data

1 m

Step 1: Determine whether primary or secondary data is best for answering your research question

What would you guess is the DIFFERENCE between Primary and Secondary data?

Primary and Secondary Data

• Primary data: you collect it yourself– Ex: a survey that you administer that asks about the

health behaviors of people in your community. • Secondary data: has already been collected by

someone else, and you analyze it in a new way to answer the question that you are interested in– Ex: data collected by the Centers for Disease Control

and Prevention (CDC) on disease rates in different states

Quick Quiz• If your question is: Why are rates of heart

disease different around the US?What type of data would you most likely collect?

Primary data OR

Secondary data This is probably not data you would be able to collect yourself!

Step 2: Choose a Study Design

• Descriptive: answers who, what, when, where• Analytical: answers why or how

Steps to Identifying the Problem:

1) Choose health-related outcome2) Clearly define the outcome (“case”)3) Choose a population4) Describe the problem (descriptive study)

PrevalenceIncidence

Step 2: Choose a Study Design

• Descriptive: answers who, what, when, where• Analytical: answers why or how

Case-Control

Cross-SectionalCohort

Analytical Study Design

Case-ControlCross-SectionalCohort

STEPS:Cohort

(FORWARD)(CAUSE to EFFECT)

Cross-Sectional(“Snapshot”)(cause & effect)

Case-Control(BACKWARD)

(EFFECT to CAUSE)

Break up your population into:

- Have RF- Don’t have RF

- No initial groups - Have Outcome- Don’t have Outcome

…then: - Wait- Usually a long period of time

- survey/interview/test them- identify who has RF & Outcome at the same time

-Look backwards into their past- Usually done through survey, interview, or lab tests

…to determine if: - Outcome?- No outcome?

RF & OUTCOME show up in the same people?

- Had RF?- Didn’t have RF?

Cohort Set-up

What might be some drawbacks to a cohort study?

Cohort Example

Case-Control Set-up

Population

Cases (Outcome)

Controls (No Outcome)

RF No RF RF No RF

(Outcome)

(Risk Factor)(Risk Factor)

Case-Control Example

Illinois Population

Cases of Brain Tumors

Controls (no Brain Tumors)

High usage of cell phone

Low usage of cell phone

High usage of cell phone

Low usage of cell phone

(Outcome)

(Risk Factor)(Risk Factor)

What might be some advantages of a case-control study over a cohort study?

Quick Quiz:Identify the study method below as

CASE-CONTROL or COHORT

1) Assemble a group of 300 persons with lung cancer and a group of 300 persons without lung cancer and question them about their past smoking history

2) Follow a group of smokers and a group of nonsmokers over time to see who develops lung cancer

Cross-sectional Study• Information on the exposure or risk factor is

collected at the same time as information about the health outcome

• Often performed by using a survey

Example: High school athletes given a survey including questions on: 1) Risk factors: helmet use, training, & type of sport they play2) Outcomes: types of injuries, severity of injuries

Step 3: Collect & Analyze Data

• We’ll practice this during the case!

Extension:

• Clinical trial (experimental)•Cohort•Case-control•Cross-sectional

MOST Reliable

LEAST Reliable

What calculation would I use?• Cohort: Follows population forward in time,

from suspected cause to effect– Quantified by calculating the relative risk for the

exposure• Case-control: Works backwards, from

suspected effect to cause – Quantified by calculating the odds ratio (we’ll

learn about odds ratio later on, but it’s quite similar to relative risk!)

Extension:

In general, five criteria must be met to establish a cause-and-effect relationship:

• Strength of association—the relationship must be clear.

• Consistency—observation of the association must be repeatable in different populations at different times.

• Temporality—the cause must precede the effect.• Plausibility—the explanation must make sense

biologically.• Biological gradient—there must be a dose-response

relationship.