Lecture 3 - Threats to Validity Flashcards
External Validity
The extent to which research can be applied to situations outside of the experimental setting.
Internal Validity
The extent to which effects can be reliably connected with the manipulation of the independent variable.
Construct Validity
The extent to which selected measures map onto the theoretical constructs they are purported to be measuring.
Treatment Variance
Variance among scores traceable to our own methodology.
Confound Variance
Variance among scores caused by uncontrolled variables.
Error Variance
Variance among scores caused by extraneous variables and measurement errors.
Ceiling Effect
An error of measurement sensitivity in which an artificial upper limit on scores skews the distribution of results; in certain tests, if the task is too easy for participants, an artificial upper limit is created, and many achieve perfect scores.
Floor Effect
An error of measurement sensitivity in which an artificial lower limit on scores skews the distribution of results; in certain tests, if the task is too difficult for participants, an artificial lower limit is created, and many achieve scores near zero.
Order Effects
A potential confound that results from the order in which an experimental procedure is carried out.
Practice Effects
The changes in performance resulting from familiarity with the task being completed in which participants are seen to improve over time.
Fatigue Effects
The changes in performance resulting from overexposure to dull or tedious tasks in which participants become bored.
Counterbalancing
A method designed to reduce order effects; one group experiences the conditions in one particular order, while another experiences them in the opposite order.
Latin-Square Design
A technique used to reduce order effects by partially counterbalancing the conditions of an experiment; each condition is assigned a position in a column and row where each only occurs once in their respective column and row. Individual participants are assigned their own row and experience that particular sequence of conditions.
Randomization
A technique to limit the likelihood of order effects emerging; arbitrarily organizes stimuli and trials.
Demand Characteristics
Cues within a study that are intended to help participants fulfill their role in the experiment; may potentially alter the behavior of participants as a result of perceived expectations.
Pre-Test, Experimental Treatment, and Post-Test
A pre-test consists of measuring a baseline for the variable of interest before an experiment begins.
An experimental treatment refers to the condition of experiencing the independent variable.
A post-test consists of a re-assessment following the experimental treatment.
Maturation Effects
Refers to any of the natural biological or intrapsychic processes that change over time and pose a threat to the internal validity of an experiment.
History Effects
Refers to any changes in the circumstances surrounding the lives of individual participants outside of an experiment that may influence its outcome.
Testing Effects
Refers to any effect of taking a test on respondents. It could refer to the sophistication of a test; it could also refer to the influence of having taken a pre-test on re-assessment during the post-test.
Passive Control Group
A control group involved in the completion of a meaningless alternative task.
Active Control Group
A control group involved in something that could be reasonably assumed to have an effect.
Wait List
A control group that believes they will receive the treatment at some point later.
Attrition
Refers to when participants start but do not complete a study. It is wise to be particularly mindful of differential attrition in which participants withdraw from one condition or treatment more than any other.
Matched Design
Design in which participants are matched as closely as possible on some characteristics and then assigned to separate groups.
Random Group Assignment
Participants are allocated to groups by some method that involves chance.
