CRIM 135 Lecture Notes - Lecture 6: Statistical Conclusion Validity, Causal Inference, Random Assignment
21 Oct 2016
School
Department
Course
Professor
Crim 220 L6
Experimental Research Designs
Introduction
- A eperiet is a proess of oseratio i riial justie researh, to e arried out
in a situation expressly brought about for that purpose
Experiments involve:
1. Taking action
2. Observing the consequences of the actions
Components of a true experiment
- Researcher is controlling IV
- Random assignment
- Causal inference (maybe)
Classical Experiment
4 central features
variables
time order
measures
groups
Major components
- IV and DV
- Pretesting and post-testing
- Experimental and control groups
Experimental group – exposed to IV
Control group – not exposed to IV
IV = The cause
E.g treatment, policy, intervention
DV = The effect
- Value depends on IV
- E.g Physical conditions, behavior
find more resources at oneclass.com
find more resources at oneclass.com
Pretesting and Post testing
- Subjects are measured on the DV prior to the IV
- Exposed to DV
- Measured after DV
Hawthorne Effect
IV:
- Would improved working conditions (better lighting)
DV
- Result in improvement in employee satisfaction and productivity
YES – BUT so did with worse lighting
Identified need for control group
Placebo
- We ofte do’t at people to ko if the are reeiig treatet or ot
- We expose our control group to a dummy independent variable just so we are treating
everyone the same
- Ensures that changes in the dependent variable actually result from the independent
variable and are not psychologically based
Double Blind Experiments
- Experimenters may be more likely to observe improvements among those who received
drug
- Experiments AND participants unaware of who is which study
o Neither group can purposely or inadvertently influence study results
Selecting Subjects
Two steps:
- Decide on target population – the group to which the results of your experiment will
apply
o GENERALIZABILITY
Pick a particular focus group
- Decide how to select particular members from that group for your experiment
o REPRESENTABILITY
Assigning participants to study groups
- The goal is to able to attribute changes in the treatment group to the intervention
- In order to do that, the groups should not be different prior to the intervention
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
A(cid:374) e(cid:454)peri(cid:373)e(cid:374)t is (cid:862)a pro(cid:272)ess of o(cid:271)ser(cid:448)atio(cid:374) i(cid:374) (cid:272)ri(cid:373)i(cid:374)al justi(cid:272)e resear(cid:272)h, to (cid:271)e (cid:272)arried out in a situation expressly brought about for that purpose. Experiments involve: taking action, observing the consequences of the actions. 4 central features variables time order measures groups. Subjects are measured on the dv prior to the iv. Result in improvement in employee satisfaction and productivity. Yes but so did with worse lighting. We ofte(cid:374) do(cid:374)"t (cid:449)a(cid:374)t people to k(cid:374)o(cid:449) if the(cid:455) are re(cid:272)ei(cid:448)i(cid:374)g treat(cid:373)e(cid:374)t or (cid:374)ot. We expose our control group to a dummy independent variable just so we are treating everyone the same. Ensures that changes in the dependent variable actually result from the independent variable and are not psychologically based. Experimenters may be more likely to observe improvements among those who received drug. Experiments and participants unaware of who is which study: neither group can purposely or inadvertently influence study results.
