Transcript 1.3 Types of Statistical Study

Statistical Reasoning
Intro to Probability and
Statistics
Mr. Spering – Room 113
1.3 Types of Statistical Study
Subjects – people or objects chosen for
the sample
 Participants – when the subjects are
people usually referred to as participants

“You can observe a lot by just watching”
Yogi Berra
1.3 Types of Statistical Study

Variables:
Variable of interest- Items or quantities that the
study seeks to measure
 Explanatory variable- Cause and effect
 Response variable- Variable which is a response
to the cause and effect relationship

1.3 Types of Statistical Study

3 Basic Types
study – observe and measure
statistics, do not influence or modify
characteristics
 Experiment – apply treatment then observe
effects
 Meta-analysis – study a topic that considers
previous studies to find trends not evident in
original study
 Observational
1.3 Types of Statistical Study

Observational Studies:

Retrospective study (Case-control) - Uses data
from the past.

When do you think this type of study is most prevalent?
Ethics???

Prospective study (Longitudinal) - Set up to
collect data in the future from a predetermined
sample.
1.3 Types of Statistical Study

Case-control study (Retrospective) –
observational study that resembles an
experiment because the subjects naturally divide
into groups.
Cases – engage in behavior
 Control – do not engage in behavior
 Used many times when an experiment may be
deemed unethical, allow subjects to voluntarily
exhibit behavior (i.e. Do seat belts save lives? Make
sure you wear your seatbelt!)

1.3 Types of Statistical Study

Study Groups:
group – receive treatment
 Control group- do not receive treatment
 Treatment

Note: Study could involve more than two groups,
and the groups should be chosen randomly from
larger sample
1.3 Types of Statistical Study
Confounding – can not determine the
effects of specific factors
 Confounding factors – factors that lead to
confusion

1.3 Types of Statistical Study

Strategies For Selecting Groups
a randomized experiment – participants
assigned randomly
 Use a large number of participants – unlikely
to differ in significant way
 Use
1.3 Types of Statistical Study



Placebo – Identical in appearance to the
treatment, but lacks active ingredient,
participants can not distinguish
Placebo effect – patients improve simply
because they believe receiving treatment (i.e.
power of suggestion)
Experimenter effect – when
experimenter somehow influences
subjects through factors such as facial
expression, tone of voice, or attitude
(i.e. education: self-fulfilling prophecy)
1.3 Types of Statistical Study

Single-blind – participant does not know
which group they are part of
Double-blind – participant nor experimenters
know who belongs to which group
1.3 Types of Statistical Study
1.3 Types of Statistical Study

a.
Which type of Study?
What is the mean income of stock brokers?
Observational
b.
Can lifting weights improve runner’s times?
Experiment, no blinding
c.
Does skin contact with particular glue cause rash?
Experiment, single-blind, not double-blind because experimenter can not
cause a rash
d.
Can a new herbal tea remedy reduce severity of cold?
Double-blind experiment, severity of cold may be influenced by
experimenter
e.
Dozens of individual studies have given contradictory
results, can exercise increase life span?
Meta-analysis
1.3 Types of Statistical Study

Summary
 Why
do we need a control group?
 What two strategies do we keep in mind when
choosing groups?
 What is a placebo?
 Actual example of experimenter effect?
 When are case-control studies usually used?
Class work—pg 32 #9-28 all