Experimental Design

Section 1.3
 Designing a Statistical Study

1. Identify the variable(s) of interest (the focus) and the

population of the study.
2. Develop a detailed plan for collecting data. Make
sure sample is part of the population.
3. Collect the Data
4. Describe the data, using descriptive statistic
techniques.
5. Interpret the data and make decisions about the
population using inferential statistics.
6. Identify any possible errors.
 Ways to Collect Data

1. Observational Study
› Researcher observes and measures but does not change the
environment at all.
 Example: Researches observed and recorded what children up
to three years old did with nonfood objects (saw if they put it in
their mouths)
2. Experiment
› Treatment applied to part of a population and responses are
observed. You can also use a control group and a placebo.
 Example: Diabetics take a pill to see if helps reduce their risk of
heart disease while a control group took a water pill.
 Ways to Collect Data

3. Simulation
› Mathematical or physical model used to reproduce the
conditions of a situation
› Done when experiment is too dangerous or costly.
 Example: Automobiles use dummies when they are studying the
effects of crashes on humans.
4. Survey
› Investigation of one or more characteristics of a population
(interview, mail, telephone)
 Example: A survey conducted on females physicians to determine
whether the primary reason for their career choice is financial
stability.
 Examples:

1. A study of the effect of changing flight patterns on the

number of airplane accidents.
› Simulation
2. A study of the effect of eating oatmeal on lowering blood
pressure.
› Experiment
3. A study of how fourth grade students solve a puzzle
› Observation
4. A study of U.S. residents’ approval rating of U.S.
president
› Survey
 Three Key Elements of a well
designed experiment are:
 1.) Control influential factors
› A confounding variable occurs when an experimental
cannot tell the difference between the effects of
different factors on a variable.

 Example: A coffee shop owner wants to attract more

customers into her shop so she decorates it in bright
colors. At the same time a new shopping mall opens up.
If the business at the shopping mall increases you can
not determine if it is the new colors or the shopping
mall.
Three Key Elements of a well
designed experiment are:
› Placebo Effect  occurs when a subject acting
favorable to a placebo even when they received no
medication.

 Example: Someone who has depression is given

medicine which in fact is a water pill. The person then
starts to feel better because they believe the medicine is
working.
 Three Key Elements of a well
designed experiment are:
 2.) Randomization – Randomly assign subjects
to different treatment groups
› Could have groups being completely random.
› Could have groups be in blocks
 Blocks are groups of subjects have the same
characteristics
› Could have groups be in a randomized block design
 Example: An experiment of a weight loss drink. You may
create blocks of 20-29 year olds, 30-39, and 40-49. Then
in those blocks randomly pick people to be in the
treatment group or control group.
 Three Key Elements of a well
designed experiment are:
 3.) Replacement
› The repetition of an experiment using a large group of
subjects.
› HAVE LARGE SAMPLE SIZES
 Placebo Effect
 Placebo – a faux treatment that looks like the
real treatment (i.e. sugar pill). It acts as a control.

 Placebo Effect – occurs when an untreated

subject incorrectly believes that he/she is
receiving a treatment and reports an improvement
in symptoms.
 Example:
 The company identifies ten adults who are heavy
smokers. Five of the subjects are given the new
gum and the other five subjects are given a
placebo. After two months, the subjects are
evaluated and it is found that the five subjects
using the new gum have quit smoking.
› Sample size too small, should be replicated.
› Results of the 5 adults who were given the placebo are
not given.
 Example:
 The company identifies 1,000 adults who are
heavy smokers. The subjects are divided into
blocks according to their gender. Females are
given the new gum and males are given the
placebo. After 2 months, the female group has a
significant number of subjects who have quit
smoking.
› Groups not similar. Divide into blocks and then split
the blocks into treatment group and control group.
› Don’t know the results of the men's group
Sampling Techniques
 Sampling Techniques
 Census
› A count or measure of an entire population (costly
and difficult)
 Sampling
› A count or measure of part of a population
 Sampling error
› The difference between the results of a sample and
those of a population
 5 Sampling Techniques
1. Simple Random Sample
› Every possible sample of the same size has the
chance of being selected
› Appendix B
› Assign a different number to every member of the
population and use of a random number generator to
choose group
 5 Sampling Techniques
2. Stratified Sample
› Used when it is important to have members from each
segment of the population in our sample
› Members of a population are divided into two or more
subsets that are called strata that share a similar
characteristic such as age, gender, ethnicity, etc.
› A sample is randomly selected from each strata
› Example: Divide homes into socioeconomic levels
 5 Sampling Techniques
3. Cluster Sample
› Use when population falls into naturally occurring
subgroups, each having similar characteristics

› Divide population into groups called clusters

› Select all members in one or more clusters (not all)

› Example: Divide into zip codes, Class courses

 5 Sampling Techniques
4. Systematic Sample
› Each member of the population is assigned a number

› Members are ordered in some way

› Starting number is selected, and then sample members are

selected at regular intervals (every 3 rd, every 5th, etc.)

› Example: Assign numbers to each house in Cranberry

Township and then select every 100 th household.
 5 Sampling Technique
5. Convenience Sample
› Only use the available members of the population
› Not recommended!!
 You get biased results
 Example
 You select a class at random and question each student in the
class.
› Cluster
 You divide the student population with respect to majors and
randomly select and question some students in each major.
› Stratified
 You question every 20th student you see in the hall.
› Systematic
 You assign each student a number and generate random
numbers. You then question each student whose number is
randomly selected.
› Simple Random Sample
 Homework
 Page 25: 2, 3, 17-27