Post on 19-Dec-2015
Validity, Sampling & Experimental Control
Psych 231: Research Methods in Psychology
Announcements
The required articles for the class experiment paper are already on-line at the Milner course reserves page
Class Experiment
Collect the forms (consent forms and data summary sheets) - pass to the front
Brief discussion:– So how did it go? – What happened? – Any thing unusual/unexpected?– Any problems?
External Validity
Are experiments “real life” behavioral situations, or does the process of control put too much limitation on the “way things really work?”
Will the same basic conclusions be supported with different operational definitions, different participants, different research settings?
External Validity Variable representativeness
– relevant variables for the behavior studied along which the sample may vary
Subject representativeness – characteristics of sample and target population
along these relevant variables Setting representativeness (ecological
validity)– how do the characteristics of the research setting
compare with the “real world”
Internal Validity
The precision of the results Did the change result from the changes in the
DV or does it come from something else?
Threats to internal validity History – an event happens during the experiment Maturation – participants get older (and other changes) Selection – nonrandom selection may lead to biases Mortality – participants drop out or can’t continue Testing – being in the study actually influences how the
participants respond Statistical regression – regression towards the mean, if you
select participants based on high (or low) scores (e.g., IQ, SAT, etc.) their scores later tend to move towards the mean.
“Debugging your study”
Pilot studies– A trial run through– Don’t plan to publish these results, just try out the
methods
Manipulation checks– An attempt to directly measure whether the IV
variable really affects the DV.– Look for correlations with other measures of the
desired effects.
Sampling
Why do we do we use sampling methods?– Typically don’t have the resources to test
everybody
Population - everybody that the research results are targeted
Sample - the subset of the population that actually participates in the research
Sampling
Goals:– Maximize:
• Representativeness - to what extent do the characteristics of those in the sample reflect those in the population
– Reduce:• Bias - a systematic difference between those in the
sample and those in the population
Sampling Methods Probability sampling
– Simple random sampling– Systematic sampling– Stratified sampling
Non-probability sampling– Convenience sampling– Quota sampling
Simple random sampling Every individual has a equal and
independent chance of being selected from the population
Systematic sampling Selecting every nth person
Stratified sampling Step 1: Identify groups (strata) Step 2: randomly select from each group
Convenience sampling Use the participants who are easy to get
Quota sampling Step 1: identify the specific subgroups Step 2: take from each group until desired number of individuals
Experimental Control
Our goal: – to test the possibility of a relationship between the
variability in our IV and how that affects our DV.
– Control is used to minimize excessive variability.– To reduce the potential of confoundings.
• if there are other variables that influence our DV, how do we know that the observed differences are due to our IV and not some other variable
Sources of variability (noise)
Sources of Total (T) Variability:
T = NonRandomexp + NonRandomother +Random
Sources of variability (noise)
I. Nonrandom (NR) Variability – systematic variation
A. (NRexp)manipulated independent variables (IV) i. our hypothesis is that changes in the IV will result in
changes in the DV
B. (NRother)extraneous variables (EV) which covary with IV
i. other variables that also vary along with the changes in the IV, which may in turn influence changes in the DV
Sources of variability (noise)
II. Random (R) Variability A. imprecision in manipulation (IV) and/or
measurement (DV)
B. randomly varying extraneous variables (EV)
Sources of variability (noise)
Sources of Total (T) Variability:
T = NRexp + NRother +R
Our goal is to reduce R and NRother so that we can detect NRexp.
That is, so we can see the changes in the DV that are due to the changes in the independent variable(s).
Weight analogy
Imagine the different sources of variabilility as weights
RNRexp
NRother
RNRother
Treatment group control group
Weight analogy If NRother and R are large relative to NRexp
then detecting a difference may be difficult
RNRexp
NRother
RNRother
Weight analogy But if we reduce the size of NRother and R
relative to NRexp then detecting gets easier
RNRother
RNRexpNR
other
Methods of Controlling Variability
Comparison Production Constancy/Randomization
Methods of Controlling Variability
Comparison – An experiment always makes a comparison, so it
must have at least two groups• Sometimes there are baseline, or control groups• This is typically the absence of the treatment
– Without control groups if is harder to see what is really happening in the experiment
– it is easier to be swayed by plausibility or inappropriate comparisons
• Sometimes there are just a range of values of the IV
Methods of Controlling Variability Production
– The experimenter selects the specific values of the Independent Variables
• (as opposed to allowing the levels to freely vary as in observational studies)
– Need to do this carefully• Suppose that you don’t find a difference in the DV across
your different groups– Is this because the IV and DV aren’t related?
– Or is it because your levels of IV weren’t different enough
Methods of Controlling Variability
Constancy/Randomization– If there is a variable that may be related to the
DV that you can’t (or don’t want to) manipulate– Then you should either hold it constant, or let it
vary randomly across all of the experimental conditions
Potential Problems of Experimental Control Excessive random variability:
– If control procedures are not applied, then R component of data will be excessively large, and may make NR undetectable
Confounding: – If relevant EV covaries with IV, then NR component of data
will be "significantly" large, and may lead to misattribution of effect to IV
Dissimulation: – If EV which interacts with IV is held constant, then effect of
IV is known only for that level of EV, and may lead to overgeneralization of IV effect
Next time
Read: Chpt 8