Where Do Data Come From

1
Where Do Data Come From ?

• Chapter 1

2
(No Transcript)
3
You Cant See Just by Watching!

• Statistical data are everywhere in our daily
  lives
• The unemployment rate is 4.5
• 70 of people 18-24 think downloading music off
  the net is no different from buying a used CD
• Where do these data come from?
• Can we trust them?
• Data good and bad are the result of human
  effort
• Bad data result from laziness, lack of
  understanding and even the desire to deceive
• The first question when we see a number is where
  do the data come from?

4
Talking about Data Individuals Variables

• Statistics is the art and science of data
• A great deal of judgment is necessary to generate
  data thats where the art comes in
• First we need some basic vocabulary
• Individuals
• Variables

5
(No Transcript)
6
Example Individuals and Variables
7
Example 1 Who recycles?

• Who recycles?
• Two neighborhoods, high income and low income
• Individuals are the households in each
  neighborhood
• Variable is the weight of the recycled material
  per week
• Result High income households contributed more
  pounds of recycled material per week
• Problem Does this really tell us anything?
  Turns out the high income households recycled
  more heavy glass bottles than the low income
  households is this really more recycling? Or
  just recycling different things?

8
Observational Studies

• Sometimes all we can do is watch
• Chimpanzees behaving naturally
• Children and teachers in their natural
  environment
• Observation that is organized and measures
  clearly defined variables is more convincing than
  just watching
• Many studies in the social sciences are
  observational
• An example you are all familiar with is the Census

9
(No Transcript)
10
Example 3 Do power lines cause leukemia?

• Electric currents generate electric magnetic
  fields
• Strong fields disturb cells grown in laboratories
• Does living near power lines disturb a persons
  cells and lead to leukemia in children?
• We cant do experiments that potentially harm
  children !!
• Leukemia is rare and locations vary in a lot of
  other ways other than proximity to power lines ?
  difficult to compare children near and far from
  power lines
• It is more feasible to look at children who have
  leukemia and compare them with children who dont

11
Example 3 Do power lines cause leukemia?

• Then look at possible causes diet, pesticides
  proximity to power lines and see for which of
  these children with leukemia differ from those
  without
• Result Careful 5 million study compared 638
  children with leukemia with 620 healthy children.
• Measured all kinds of things pertaining to the
  children and their environment
• Found no systematic difference in the exposure to
  electric and magnetic fields between the children
  with and without leukemia
• This is a well conducted observational study

12
(No Transcript)
13
Sample Surveys
You dont have to eat the whole ox to know that
the meat is tough

• Sampling To gain information about the whole by
  examining only part
• Sample Surveys Gain information on a group of
  individuals by studying only some of its members
• Selected not because they are special, but
  because they represent the group
• Vocabulary
• Population
• Sample

14
(No Transcript)
15
Population Samples

• The population is the group we want to study
• The sample consists of the members of the
  population for whom we actually have information
• To make sense of a sampling results you must know
  both the population and the sample

16
Example 5 The Current Population Survey (CPS)

• Government economic and social data come from a
  number of large sample surveys
• The monthly CPS is the most important
• Many variables in the CPS describe employment of
  people 16 and older in a household
• Monthly unemployment rates come from the CPS
• CPS Population
• The more than 11 million U.S. households
• CPS Sample
• About 60,000 households interviewed each month

17
(No Transcript)
18
Example 7 The General Social Survey (GSS)

• Social science and especially Sociology make
  heavy use of sampling
• GSS variables cover the subjects personal and
  family background, experiences and habits, and
  attitudes and opinions on subjects from abortion
  to war
• Every second year
• GSS Population
• Adults age 18 and older living in households in
  the U.S.
• GSS Sample
• About 3,000 adults interviewed in person in their
  homes

19
Sampling

• Sample is often used in a generic sense to
  describe the study group
• Sample Survey usually describes a study that
  uses an organized plan to choose a sample that
  represents a specific population
• A sample survey does not rely on judgment to
  choose the sample
• Starts with the entire population and uses
  quantifiable, reproducible methods to choose a
  sample that represents the population

20
Census

• Why not look at the whole population instead of
  just a part of it?
• Never completely true, always a sample survey
  with a very large sample
• Produces lots of local information on every small
  part of the country

21
Experiments

• The aim of all observational studies, including
  sample surveys, is to produce a picture of the
  population disturbed as little as possible by the
  act of observing

22
Experiments

• With an experiment, the object is to change
  behavior
• We actively impose a treatment in order to
  observe the response
• Can answer questions like Does aspirin reduce
  the chance of a heart attack?

23
Example 8 Helping welfare mothers find jobs

• Most welfare recipients are mothers of young
  children
• Many move off welfare by getting better jobs, and
  many of those use job-training programs to help
  them
• Question Should job-training/job-search programs
  be mandatory for all able-bodied welfare mothers?
• Observational studies are not enough
• Even if we have a properly chosen sample of
  welfare recipients, those who seek out job
  training/search programs may differ in many ways
  from those who do not
• It may be these differences that are responsible
  for their success, not really the programs at all
  ?

24
Example 8 continued

• E.g. the motivated mothers are observed to have
  more education
• To see if mandatory job training/search programs
  work, they have to be tried an experiment must
  be done
• Choose two similar groups of mothers when they
  apply for welfare
• One group has mandatory job training/search
  programs treatment group
• The other allowed to behave normally control
  group
• Compare the income and work records of the two
  groups after several years

25
Experiments

• The BIG advantage of experiments is that they can
  in principle give good evidence for cause and
  effect
• The results of a properly conducted experiment
  are good evidence that a given treatment causes a
  given response
• Cautions
• Experimental results describe the average effect
  on a group, not the specific effect on an
  individual
• Many types of experiment are not ethical

26
Summary

• A statistical study records information about
  individuals (people, animals, households, things
  ) by giving the value for one or more variables
  for each individual
• Some variables are numeric, some are not
• Some variables are useful, others not at all
• The most important fact about any statistical
  study is how the data were produced

27
Summary

• A census attempts to measure every individual in
  a population i.e. a complete sample survey
• An experiment actually does something to the
  individuals in order to observe a response
• The usual aim is to see if a treatment actually
  causes a specific response on average

28
Summary

• Observational studies try to gather information
  without disturbing the scene they are observing
• Sample surveys are a very important kind of
  observational study
• A sample survey chooses a sample from a specific
  population and uses the sample to get information
  about the entire population