Basics of Experimentation 1 Chapter 6 Variables and

Basics of Experimentation 1 Chapter 6 Variables and Control

“Onscreen Bullying Study” • Read the article you were given.

Variable Anything that can assume two or more values (i. e. , levels) • • • Height Gender Year in school Number of times married Income

Article Question • Name all the possible variables.

IV & DV • An IV or DV must each have more than one possible value or level? • IV usually limited to only 2 -3 levels. • DV can have an infinite number of levels. (Extraneous variables have all combinations (0 -infinite)

Article Question • How many IVs? – Identify the IV(s) and all its levels • Identify the DV(s)

Operational Definitions The exact way we measured or manipulated the variable and all its levels.

Example: Love - Ratings - Physiological reactivity - Money spent - Length of gazing

Article Question • How is “Relational Aggression” operationally defined?

Why do we need operational definitions? 1. Accurate replication 2. Compare studies and make generalizations 3. Evaluation of measures

Article Question • What would you need to know about the video clips and the DV measures in order to replicate the study?

How do you operationally define something? No exact way to define any variable, but some ways are better than others. 1. Use established O. D. s 2. Make it logically related to your construct (variable).

Article Question • If you were to replicate this study, would you use the same measures or method of manipulation?

Types of IV

Physiological Alteration of biological state.

Experience Manipulation of training or learning

Stimulus or Environmental Manipulating the stimulus or environment that participants are exposed to.

Participant Characteristics Ex post facto IV: Not a “true” IV because you can not randomly assign people to the levels. Age, Sex, Personality, Marital status, etc.

What makes a good IV? Strong enough to produce differences in the groups on the DV.

Article Question • Are the video clips strong enough to activate the scripts? • How could you modify the study to more strongly activate the scripts? – Would this strengthen your IV? – What would it do for “ecological validity” – What about ethics?

Ways to Measure the DV

Correctness Whether or not participants correctly answered a question or passed some type of test.

Rate How rapidly a behavior occurs during a time period. Number of bar presses per minute.

Frequency How often a behavior occurs. Number of temper tantrums.

Degree or Amount A rating or measurement of the extent to which some trait or behavior is present.

Latency How quickly a response is made. Reaction Time

Duration How long a response lasts. Length of sleep per night.

What makes a good DV? 1. One that is sensitive enough to detect differences in groups as a result of the IV. 2. Reliable and Valid

Article Questionn • What assumption did they make about their DV? (G) • How do we know if their assumption about the DV is correct? • Can you think of other ways to test this same assumption?

Women exhibit a higher willingness to cooperate during and shortly after menstruation • The researchers measured the subjects' individual willingness to cooperate using a well-established psychological scale, the "Social Value Orientation. " To do so, they asked the women to divide fictitious money between themselves and another person who was a complete stranger to them. • Anderl, C. , Hahn, T. , Notebaert, K. , Klotz, C. , Rutter, B. , & Windmann, S. Cooperative preferences fluctuate across the menstrual cycle. Judgment and Decision Making, Vol. 10, No. 5, September 2015, pp. 400 -406

Constructs vs. Operational Definitions • A construct is the “layman’s terms” or Webster (Wiki) definition of a concept. – Homesick means to feel a longing for familiar surroundings O. D. : how you measure or manipulated the construct. e. g. , Homesickness scale

Why Does it Matter? • Constructs and relationships between them are what we really care about, but we have to talk about O. D. when we talk about methods and validity of conclusions. – Homesick students are more likely to drop out of school, suffer from psychosomatic symptoms, have lower GPAs, etc. – “Homesickness” is a label based on scores.

Class Activity • • Go to the latest edition of APA Monitor. Click on the “In Brief” link. Skim the headlines and a few summaries. Select 1 construct (e. g. , “Supportive coaches”) and attempt to find the full text copy of the article. • Skim the method’s section and find out how they operationally defined that construct.

Class Activity • Operational Definitions Exercise • Divide up • Assigned Construct – Define it as both a DV and an IV. – Discuss the type of DV and IV you selected. – Discuss other ways this same construct could be measured or manipulated.

Types of Extraneous Variables • Nuisance Variables: causes scores on the DV to be more spread out for both groups. – Not as problematic and largely unavoidable. • Confounding Variables: causes differential changes in the DV. – You have an unintended, unforeseen and unwanted IV.

Ways to Control Extraneous Variables

1. Randomization (Let it vary) • Assumes that most characteristics will be evenly distributed among the groups. • Controls for lots of unknown extraneous variables. • Very easy to do. • Works well if you have enough subjects in each group (15+)

2. Elimination/Hold it Constant • Reduce the variable to only one value/level. • Guarantees that the effect is the same for all groups. • Reduces generalizability. • Takes effort.

3. Balance • Equal distribution of the values of an extraneous variable to all groups. • Requires planning and time. • Keeps the study generalizable.

Details • “Team exercise” = pairs of undergraduates playing a flight simulator video game. • “Aggressive teammates” = students who scored high on the Conditional Reasoning Test of Aggression. • “Performance” = video game scores

4. Counterbalancing • Controls order effects by presenting the levels of the IV in different sequences. • Only necessary when you are doing “repeated measures. ”

Repeated Measures • Type of design in which the same participants are used in every group and so are exposed to every level of the IV. • Participants must be measured on the DV multiple times. • Ensures that the groups are identical; so eliminates many E. V.

Within Subjects Counterbalancing • Every participant gets all possible sequences.

Within Groups Counterbalancing • Different participants get different sequences such that each sequence is given to the same number of participants. • More common.

Complete Counterbalancing • All possible sequences are presented

Incomplete Counterbalancing • Random selection of sequences chosen.

Why bother with counterbalancing? • Controls order effects • Controls carryover effects

Order Effects • When the position of a treatment (level of IV) in a series determines, in part, the participants’ response. • Example: Memory technique (phonetic, appearance, meaning) and recall.

Carryover Effect • The effects of one treatment persist or carryover and influence responding to the next treatment. • Example: Type of movie (horror, comedy, romance) and emotion

Differential Carryover Effect • Response to a treatment depends on which specific treatment came before it. • An unforeseen and unavoidable problem!

Moral of the Story! • Beware of repeated measures, particularly if you have more than 2 levels of the IV.