quasi

Question 1

QUASI-EXPERIMENTAL AND NATURALISTIC DESIGNS (CH 6)

Answer 1

experimental design may not be feasible, or causal relationship not goal
health research doesn’t always rely on experimental designs
eg. human studies on obesity – 8141 articles in last year, 501 used randomized-control (experimental) design

Question 2

1. QUASI-EXPERIMENTAL DESIGNS

Answer 2

to provide reasonable evidence of cause
introduce treatment but no control group
control time at which treatment introduced or withdrawn
or control group from different sample

Question 3

Time-series designs

Answer 3

repeated observations before and after treatment
eg. causal relationship of obesity in heart disease?
select appropriate population
specify dependent variable – some measure of heart disease eg. incidence
introduce treatment to decrease risk factor (obesity)
eg. introduce package promoting exercise and healthy diet
monitor dependent variable – need readings before and after intervention eg. community health records of incidence
repeated observations before and after treatment
eg. causal relationship of obesity in heart disease?
select appropriate population
specify dependent variable – some measure of heart disease eg. incidence
introduce treatment to decrease risk factor (obesity)
eg. introduce package promoting exercise and healthy diet
monitor dependent variable – need readings before and after intervention eg. community health records of incidence

if incidence declines after intervention may indicate causal relationship
but prone to problems of internal validity
eg. history threat – was outcome result of intervention or other factor
(eg. Introduction of drugs to control of blood cholesterol)

Question 4

2. NATURALISTIC COMPARISONS AND CORRELATIONAL STUDIES

Answer 4

differences or interrelationships between health/nutrition variables in select populations

Question 5

NATURALISTIC DESIGNS

Answer 5

NATURALISTIC DESIGNS (study in natural setting without interventions)
Reasons:
variables not amenable to manipulation
eg. gender/age difference in response to treatment –
Can’t manipulate gender/age of subjects
ethically inappropriate to intervene
eg. study on smoking – can’t introduce smoking to one group
experimental design needs non-treatment group eg. cancer patient – ethical to withhold treatment?
multiple causal/risk factors
eg. heart disease – smoking, stress, diet, genetics