Survey Methodology Data interpretation and presentation

1
Survey MethodologyData interpretation and
presentation

• EPID 626
• Lecture 11

2
References

• Many elements of this lecture were taken from
• Fink, Arlene. How to Report on Surveys. Sage
  Publications. 1995.
• Babbie, Earl. Survey Research Methods. Wadsworth
  Publishing Company. 1990.

3
Analytic modes

• Univariate analysis
• Examination of the distribution of cases on only
  one variable at a time
• Aim is description
• Example SBP distribution among subjects
• Bivariate analysis
• Examination of the distribution of cases on one
  dependent and one independent variable
• Aim is explanation
• Example SBP distribution by sex

4
Analytic modes (2)

• Multivariate analysis
• Examination of the distribution of cases on one
  dependent and more than one independent variable
• Aim is explanation
• Example distribution of SBP by race and sex

5
Presenting univariate data

• Options (exampleage)
• List all respondents with their age
• Tabulate the number of respondents within
  categories of age
• Use measurements of central tendency and
  dispersion
• Appropriate for continuous data

6
Measures of central tendency

• Mean
• Average
• Mode
• The most frequent attribute
• Median
• The middle attribute in a ranked distribution of
  observed attributes

7
Measures of dispersion

• Range
• Distance between the lowest and highest values
• Often presented as min and max
• Example
• range 40
• or range 20-60
• Standard deviation
• Distribution of observations about the mean

8
Measures of dispersion (2)

• Quartile deviation or semi-interquartile range
• One-half of the distance between the bottom of
  the first quartile and the top of the fourth
  quartile

9
Constructing bivariate tables

• Column headings are determined by the most
  important comparison
• cases vs. controls, men vs. women etc.
• There are many options for format decision
  should be made based on ease of interpretation

10
Guidelines for formatting tables

• Headings or titles should sufficiently describe
  what is in the table
• The original content of the variable (the survey
  question) should be presented in the table or in
  accompanying text
• Values or categories of each variable should be
  indicated

11
Guidelines for formatting tables (2)

• When presenting percentages, the base upon which
  they are computed should be indicated
• Number of respondents with missing data should be
  indicated
• If appropriate, statistical values (frequencies
  or percents) should be in descending order

12
Example
13
Guidelines for formatting tables (3)

• Use a standardized set of symbols to call the
  readers attention to key aspects of the table
• definitions
• statistical significance
• different denominator for
• P-value format
• use for exact p-value
• use lt for other p-values

14
Presenting data

• First you tellem what youre gonna tellem,
  then you tellem, then you tellem what you
  toldem, and then you tellem what to do with
  it.
• -Preachers proverb

15
Presenting data

• A reader should be able to replicate your study
• Data without methods is meaningless
• A reader should be able to verify all percents
  and measures of association such as OR or RR

16
Interpretation of results

• Interpretation should follow the flow from your
  research question and study objectives
• Interpretation should be guided by statistical
  significance as appropriate