Chapter 5: Descriptive Research

1
Chapter 5 Descriptive Research

• Describe patterns of behavior, thoughts, and
  emotions among a group of individuals.
• Provide information about characteristics about
  the sample rather than to test hypotheses.
• 1) Survey most common type of descriptive
  research.
• Select a sample of the population using
  predetermined questions
• this allows you to describe attitudes of
  population from which sample was drawn
• can compare attitudes of different populations
  and look for changes in attitudes over time.

2

• Surveys are usually questionnaires or interviews
• Cross-sectional survey survey one more more
  samples of the population in a single group
• allows researchers to describe characteristics of
  a population or the differences between two or
  more populations
• Successive independent samples a series of
  cross-section series is done. Survey two or more
  samples at different times with same questions.
• Allows researcher to study changes in a
  population over time, but not how individuals
  change over time.
• study peoples level of trust in the government
  at 1972 and 1980.

3

• Problems with Successive independent samples
• People surveyed at one time are not the same
  people surveyed at the next time.
• Noncomparable successive samples successive
  samples may not representative of the same
  population (e.g. could have more poor people in
  the successive sample).
• Can only describe changes in population over time
  if successive samples represent the same
  population.
• ACT scores from 1998 through 2002 decreased
  because more students wrote the ACT in 2002 who
  had no intention of going to college.

4

• Longitudinal or panel survey design same
  respondents are surveyed over time
• allows researchers to examine changes in
  individuals over time.
• Because it is correlational it is difficult to
  determine the causes of changes over time
• Attrition people may drop out of study over time
  (move, death etc) and the final sample may not
  longer be comparable to the original sample.
• Internet Surveys can be advantageous
• inexpensive, can contact people that are hard to
  reach otherwise, people can respond when they
  want
• But, researcher has little control over who
  responds and people without internet cannot
  respond.

5

• 2) Demographic Research
• describes patterns of life events and experiences
  like birth, marriage, employment etc.
• 3) Epidemiological Research
• study the occurrence of disease in groups of
  people
• Psychologists may study how diseases impact
  peoples behavior and lifestyle.
• Also study the prevalence of psychological
  disorders.

6
Sampling

• Selection of people (sample) to participate in
  research in order to make inferences about a
  larger group (population).
• Probability Sampling
• each member in a population has a specific
  probability of being chosen.
• representative sample is approximately the same
  as the population in every respect. Researchers
  can draw unbiased and accurate estimates of the
  larger population.

7

• Sampling error differences between the sample
  and the population. Almost inevitable
• Error of estimation (margin of error) represents
  how much the data from the sample differ from the
  population. Estimation of the sampling error.
• allows researchers to determine how confident
  they are that the sample results represent the
  population.
• 60 of sample chose Tide, accurate within 4
  points.
• 95 probability the true who chose tide is
  between 56 and 64

8

• Error of estimation is influenced by
• Sample size the larger the sample to more
  similar the sample is to the population.
• Economical sample a reasonable estimate of the
  population.
• Size of population sample of 100 is more
  representative if population is 500 rather than
  5000.
• Variability the larger the variability, the more
  difficult it is to estimate population values,
  and the larger the sample must be.

9

• Probability sample sample for which the
  researchers knows the probability that an
  individual is included in the sample.
• Epsem design (equal probability selection
  method) used so all individuals in population
  have an equal probability of being chosen.

10

• Ways to obtain a probability sample
• Simple random sampling each person in the
  population has an equal chance of being selected.
• Researcher must have a complete list (sampling
  frame) of all the people in the population.
• Randomly select from the sample frame
• If you want to select sample of 100 students from
  a school with 5000 students.
• You would get list of all 5000 students and
  number each from 1 to 5000.
• Use a table of random numbers to produce 100
  numbers that fall between 1 and 5000 and select
  those 100 students to be in your sample.

11

• Stratified random sampling because there may be
  different subgroups within your population (age,
  gender, ethnicity) you may want to choose people
  from each subgroup (stratum).
• Cases are randomly selected from each strata
• ensures you have an adequate number of
  participants from each stratum.
• Proportionate sampling cases are drawn from each
  strata in proportion to their prevalence in the
  population.
• If population is 70 men and 30 women, then
  sample proportionately to ensure 70 men in sample

12

• Cluster Sampling used if you cannot obtain a
  sampling frame list of the entire population
• Difficult to get a list of all health care
  workers in Alberta.
• Break the population into smaller groups
  (clusters) for which there are sampling frames
  and then randomly choose some of the clusters for
  inclusion the sample.
• Clusters are usually locations or institutions
• Randomly choose different counties in Alberta and
  then get a list of all workers in those counties
• Multistage sampling begin sampling large
  clusters, then smaller and smaller clusters

13

• Sampling Problems
• Nonrepsonse if not all people respond to survey,
  then those that do respond may have different
  characteristics than those that did not.
• If return rate is less that 100 the data may be
  biased in some way.
• Researchers try to increase response rates with
  follow up calls
• Try to determine whether respondents and
  nonrepsondents differ in particular ways.
• Misgeneralization occurs when a researcher
  generalizes to a population that differs from the
  sample.

14

• Nonprobability Samples
• Occurs when a researcher does not know the
  probability of their sample
• In many cases it is very difficult to obtain a
  probability sample.
• Much of psychological research in conducted on
  samples that are not representative whole
  population (e.g. undergrads, rats, kids in
  daycare).
• This is OK because behavioral research usually
  describes how variables are related to each other
  to support a theory, regardless of the nature of
  the sample.
• This is why replication (with different samples)
  is important to improve generalizability of
  findings.

15

• Convenience sampling using a sample of
  individuals that is readily available.
• E.g. Schools nearby, autism clinic, undergrads.
• Quota sampling convenience sampling, but you
  ensure that you get particular proportions of
  certain individuals.
• Half boys and half girls 20 4-, 6-, and
  8-year-olds
• Purposive sampling researcher decides which
  individuals are included to get a representative
  sample.
• Snowball sampling one or more individuals for a
  population are contacted, and these individuals
  are used to lead the researchers to other members.

16
Describing and Presenting Data

• Frequency Distributions
• Table that indicates how many, and is some cases
  the , of individuals in the sample that fall
  into different categories.
• Simple frequency distribution If using numbers
  (scores) possible scores are arranged from lowest
  to highest and then the frequency of each score
  is shown.

17
Frequency Table Number of Movies Seen in the
Last 6 Months Reported by Psychology Students
18

• Grouped frequency distribution combine values
  into a set of equal class intervals.
• Relative frequency proportion of the total
  number of scores that fall into each interval
• class interval must be mutually exclusive
• class intervals must encompass all scores
• all class intervals should be the same size

19
Histogram
20
Frequency Polygon
21
Measures on Central Tendency

• Gives information about the distribution
• Mean average, most common
• mean is easily influenced by outliers
• Median score in the center of the distribution
• less influenced by outliers
• Mode most frequent score
• can have more than one mode

22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
Measures on Variability

• Descriptive stats that provide information about
  about the dispersion of scores.
• range
• variance (s2) and standard deviation (s)
• standard deviation represents variability using
  the original units, not squared units as in
  variance
• most data roughly fall into the normal
  distribution
• however, in skewed distributions most scores
  falls toward one end of the distribution

26
Normal Distribution
27

• Z score represents how one score is relative to
  the rest of the data.
• Indicates how far the score is from the mean in
  standard deviations.
• Z scores can help to identify outliers.