Psyc 3000 PowerPoint 2

• Interval Scale
• Ratio Scale
• Ordinal Scale
• Nominal Scale

HINT: “IRON”

• - Always Numerical!
• - Zero does not indicate absence of variable  Example: 0 degrees does not mean there is no temp.

• -         
• Similar to Interval Scale but this can/has a true
• zero.

• -         
• Always Numerical

• -  Doubling Principle:  100 degrees is not twice as hot as 50 degrees
• but 100$ is twice as much as $50
•  
• Example: 0 seconds, 0 Dollars, height, weight,

• -         
• Mutually Exclusive you can only fall into one
• category!

• -         
• Ordered / Ranked (1st 2nd
• or 3rd place or Small/Med/Lg)

Can be Numerical or Nonnumerical

• category!
•        
• Not Ordered/Ranked (1 a = Married  2b = Divorced 3c = Never been Married)

measurable characteristic

• Antecedent
• variable/manipulated (a.k.a. predictor variable)

• Variable measured but not manipulated directly; it detects
• influence of IV (a.k.a. outcome variable)

Qualitative:

- Measured with nominal or ordinal scales

-Vary in categories/kind

 

Quantitative:

-Measured with interval or ratio scales

-Vary in Quantity, thus Quantitative

• 1. Interview participants to ensure IV
• had desired effect

• - 2. Behavioral (or physiological) indicator,
• such as blood pressure if manipulating stress (IV)

• - 3. Pre-testing, or pilot testing
• before running actual study

Main Effect

- Called “interaction effect”

- ADV of using more than one IV modifies and shows interaction/effects of other IV’s.

• - Abstract idea or consistent set of behaviors
• with a label on it. (Ex: Athletic,Friendly, Cute)

-Extent to which an abstract construct can be inferred from the operational definition of that construct.

• - Construct Validity is established by
• -Clear operational definition.
• -Face Validity: the face of it, looks valid.
• -Convergent Validity: corresponds to other measures of same construct
• Divergent Validity: does not correspond to measures of different constructs.

the extent to which the IV or DV is consistent or stable over time.

• -  Reliabilty can be established by replication.
• -  Meaning if someone else was to conduct study,experimenter should get same measures!

• -         
• The extent to which study findings can be
• generalized to “the real world”

-Results should play to be true in real world as well.

• Extent to which one can accurately state that
• the IV produced any observed effects on DV.

Cofounding Variables

• Eliminating it! (or build it into study, like
• another IV)

• Keep constant/balance across all levels of IV
• (via random assignment or matching)

History: Some influential event occurs between a pre & post test (CONTROL GROUP)

Maturation: Biological/psychological changes to participant (CONTROL GROUP)

Instrumentation: Assessment tool/person changes during study (IOA, TRAINING, “BLIND” PROCEDURES)

• Selection: Bias in selection method means characteristics of participants/groups—not under study—may affect DV (RANDOMIZATION; PRETEST TO
• ENSURE EQUIVALENT)

• Statistical regression: Groups are chosen due to high or low measurements which will regress (trend back) toward the mean over
• repeated measures—this is also a selection bias (PRETEST; RANDOMIZATION)

• Mortality/participant attrition: Loss of participants
• from study, especially when unequal across groups (CAREFUL PLANNING AND
• INCENTIVES)

Sequencing/order effect: Prior study condition (not IV) affects subsequent study condition (COUNTERBALANCING)

• Participant sophistication: Participants become familiar with subject matter or with materials/procedure (more of a problem with
• repeated measures designs) (CONTROL GROUP)

Testing effects: Participants may change their answers based on thinking further about questions and may learn from repeated testing (CONTROL GROUP)

Experimenter Attributes: Characteristics or skills of experimenter may influence results (e.g., gender, method of delivery, etc.)

• Experimenter Expectancies:
• -Recording bias
• -Bias when interpreting data
• -Can also
• actually affect participants’ responses

• Statistical regression
• Mortality problem
• Maturation problem
• History problem
• Selection problem
• Instrumentation problem

• Statistical regression
• Mortality problem
• Maturation problem
• History problem
• Selection problem
• Instrumentation problem

1. Maintain observer awareness: provide thorough and continuous training

• 2. Automation: computer/videos,etc. (take out human error and minimize bias/expectancies)
• 3. Multiple observers: measure
• inter-observer agreement (IOA)

• 4. “Blind” data recorders:
• unaware of conditions in effect (minimize bias/expectancies)

• Blind technique:
• Experimenter unaware of conditions

• Partial blind technique: Keep experimenter “blind” as long as possible
• Automation: Use
• computer, video tapes, audio recordings

Blind technique

Multiple observers

Number 1 & 2

Nada!

Demand characteristics: Cues that give participants clues as to the nature of the experiment and they change their behavior (consciously or not; positively or negatively—”screw-you effect”)

Positive self-presentation: Participants try to appear positive (and often use demand chars. in this quest!)

1) Double-Blind Placebo: Experimenter and participant unaware of the experimental conditions in effect

2) Deception: Misdirect participants as to the nature of the experiment.

3) Disguised Experiment: Participant is not aware they are participating in an experiment.

4) Independent Measures of the DV: Measure the DV away from the experimental condition.

5) Procedural control/control of participant interpretation:

Gaininsight into participants perceptions

 ^^^^ex's:

A. Retrospective verbal report (exit interview)

B. Concurrent Verbal Reports (think aloud)

C. Sacrifice groups (terminate at various times)

Random selection: is how you draw the sample of people for your study from a population

• Every member of the population as an equal chance of being selected for
• the sample

The goal is a representative sample.

Random assignment: is how you assign the sample that you draw to different groups or treatments in your study

• Every member of the sample has an equal chance of being assigned to
• each condition.

• The goal is to distribute potential extraneous variables equally to
• various groups.

By chance may get bias The larger the sample, the less likely bias will occur

• Matching: equating groups on one or more variables by measuring participants on those variables and assigning them in equal amounts
• to the various groups.

• Advantages:
•  increases sensitivity of experiment
•  morebalanced groups

• Disadvantages:                                            time-consuming;
• pre-measuring

 - induces demand characteristics