Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one...

15
Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group—family head about family Aim design to meet the unit of analysis

Transcript of Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one...

Page 1: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Study Design

Unit of analysis—the persons, individuals or groups that you are studying

May ask one person about a group—family head about family

Aim design to meet the unit of analysis

Page 2: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Design

Focus on purposes of research Describe Explain Exploration

Page 3: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Levels of measurement

Nominal Ordinal Scale Index

Page 4: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Basic Survey Designs Cross-sectional—one point in time,

describe populations Synthetic cohorts may “imply”

longitudinal data Snapshot Longitudinal—over time, describe and

explain subpopulations and populations

repetitive questions of same people

Page 5: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Longitudinal studies Trend studies—Requestioned repetitively,

problems with attrition Cohort studies—focuses on the same

subpopulation each round, subjects may vary

Panel studies—specific group of respondents, series of questionnaires over time, same sample—attrition issue; turnover table

Repeated cross-sections

Page 6: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Longitudinal

Costly Must have systems for finding and

maintaining relationships measure issues of change

Page 7: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Alternatives parallel studies—using two or more

subpopulations in parallel with same questions

Contextual studies—questions asked about the milieu in which the questions are asked

Sociometrics—physical and mental function, not attitudes

Page 8: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Choosing your design Description of a population at a

specific time—cross-section Dynamic process or change over

time –longitudinal Trend studies work if the duration

of time is short—less attrition repeat cross-sections—gross

principles over time--cheaper

Page 9: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Design issues

Combination of questions, research purpose and money available to decide on design

Paying respondents and interviewers

telephone surveys Survey antipathy Questionnaire length

Page 10: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

What do you want to measure?

change Attitudes behaviors physical or mental capabilities actions

Page 11: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Effect of survey

affects the operationalization and number of concepts covered

examples

Page 12: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Get the most from the data

Synthetic cohort divide the sample into age groups

and test the differences in one cross-section

less cost must be careful about ecological

fallacy

Page 13: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Error Sources

Measurement Misconceptualization Randomness error

Page 14: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Validity Issues

Face validity Content Criterion-based Construction Convergent Discriminant

Page 15: Study Design Unit of analysis—the persons, individuals or groups that you are studying May ask one person about a group— family head about family Aim design.

Standardized variables

IPCSR Design texts Other surveys