Cause and effect relationships are established by manipulating the INDEPENDENT variable(s) and observing the effect on the DEPENDENT variable.
Research design must control for the possible effects of extraneous variables that could mask, enhance, or in some way alter the effect of the independent variable on the dependent variable.
Experimental Research
Example:General study description: Recruited obese participants will spend 3 weeks in a tightly controlled laboratory setting
Dependent Variable: Weight
Loss
Independent variable: food intake
Independent variable: food intake
Independent variable: exercise
regimen
Independent variable: exercise
regimen
Internal Validity: determined by the degree to which the observed effects of the independent variable (IV) are REAL and not caused by extraneous factors
Alternative explanations Alternative explanations for the effect of the independent variable (IV) on the dependent variable (DV) threaten internal validity
KEY: controlling for the possible effects of extraneous variables
Internal & External Validity
External Validity: determined by the ability to generalize the study results beyond the study sample
Internal & External Validity
Threats to Internal Validity
alternate explanations
History Maturation(children) Testing Instrumentation Selection bias
Mortality/attrition Hawthorne Placebo
blind vs. double blind Implementation
fidelity
Randomly select participants from a well-defined study population
Randomly assign selected participants to groups
Include non-treatment control groups in the research design
Control StrategiesThreats to Internal Validity
External validity can not exist without internal validity
If the results of the study are not internally valid, there is nothing to generalize.
Researchers should be always be concerned about ensuring internal validity first.
Final Point on Int/Ext Validity
Identify and use a design that…
Controls as many extraneous variable as possible
Will still be practical and feasible to implement
Choosing a Design
X =independent variable (the treatment) X2 or Y = additional treatments
O = measurement of the dependent variable (an observation) Each observation or measurement is numbered
indicating order (O1, O2, O3 )
R = random assignment
Hawthorne effect
Experimental Designs
Survey research designs Cross –sectional Longitudinal
Trend studies –track population changes over time Youth Risk Behavior Survey (YRBS)
http://www.cdc.gov/HealthyYouth/yrbs/pdf/us_injury_trend_yrbs.pdf
Cohort study – follow a particular group or subgroup over time National Longitudinal Study of Adolescent Health (Add Health)
http://www.cpc.unc.edu/projects/addhealth/design Panel study – examine the same group of people over time
at the individual level Panel Study of American Religion and Ethnicity (PS-ARE)
http://www.ps-are.org/index.asp
Non-experimental Designs
Correlational study Identifies relationships and the degree or
closeness of those relationships
A correlation exits if, when one variable increases another variable either increases or decreases in a somewhat predictable way.
What is the relationship between participation in intramural sports and BMI among WOU students?
What is the relationship between religiosity and age of sexual initiation in seventh grade students?
Non-experimental Designs
Linear relationships Positive: both variables move in the same
direction (one variable increases as the other increases)
Negative: one variable moves in the opposite direction of the other (one variable increases while the other decreases)
Curvilinear relationships
Types of Relationships
Rough measure = scatter plot
Statistic = correlation coefficient or r (describes a sample of paired values from two different variables) Measures the closeness with which the pairs of
values fit a straight line
Range of values for r = +1.0 to -1.0 When r = 0, there is no correlation 1.0 = perfect correlation
Assessing correlation
Line of best fit
http://staff.argyll.epsb.ca/jreed/math9/strand4/scatterPlot.htm
Interpreting a Scatter Plot
Relationships cause & effect Correlation of ice cream sales and death by
drowning (r = +.86)
In the months when ice cream sales go up, so do deaths by drowning and likewise when ice cream sales go down, so do deaths by drowning
A.) Does ice cream consumption cause drowning deaths to increase? or B.) Do drowning deaths cause surviving family members and friends to eat more ice cream?