Term
|
Definition
| People enlisted by a researcher to act as other participants or "accidental" passersby, creating a particular social situation to which "real" participants then respond. |
|
|
Term
|
Definition
| Influenced by independent variable, which in turn influences the dependent variable |
|
|
Term
|
Definition
| Temporary, changeable attribute that is influenced by situational factors |
|
|
Term
|
Definition
| Stable over time and not easily influenced by situational factors |
|
|
Term
|
Definition
when predicted relationship exists in nature, is more likely to produce a clear and convincing sample relationship
try to decrease variability among scores within each condition and to increase the differences among scores between conditions (minimize error variance and maximize differences in scores between conditions) |
|
|
Term
|
Definition
| Variability within scores that decreased the strength of a relationship. Larger error variance --> weaker relationship |
|
|
Term
|
Definition
involves conditions that are likely to produce large differences in scores between conditions
Created by: 1.) select amounts/ categories of IV that are substantially diff. from one another 2.) have participants experience condition sufficiently for it to dramatically influence their behavior |
|
|
Term
|
Definition
participants in one condition are aware of treatment given in other conditions
prevented by strong manipulation |
|
|
Term
|
Definition
| measurement, in addition to DV, that determines whether conditions had intended effect |
|
|
Term
|
Definition
participants must select from possible choices we provide
ex.) yes-no task, sorting task, self-reports, likert type: statement presented and rate response |
|
|
Term
|
Definition
| different approaches that "converge" on same behavior |
|
|
Term
|
Definition
define system for assigning different scores to different responses determines when a response is correct or not, what constitutes beginning & end of responce, how to distinguish one from another
minimize inconsistency and bias |
|
|
Term
| sensitive dependent measure |
|
Definition
| produces different scores for small differences in behavior, let us detect even a small influence that a manipulation produces |
|
|
Term
|
Definition
| occurs when range of possible scores on a variable is limited |
|
|
Term
|
Definition
lowest potential scores from worst-scoring participants are very high
scores don't differ much because can't get much higher |
|
|
Term
|
Definition
highest potential scores from best participants are very low
scores don't differ much because can't get much lower |
|
|
Term
|
Definition
| test participants as in real study, but ignore trails when analyzing the results (good for physical reactions) |
|
|
Term
|
Definition
observe several times in condition so differences in motivation or attention balance out
--> more powerful design and more confidence in reliability |
|
|
Term
|
Definition
| any influence on performance of a particular trial that arises from its position in sequence of trials |
|
|
Term
|
Definition
| influence on performance that arises from practicing task |
|
|
Term
|
Definition
| influence that particular trial has on performance of subsequent trials |
|
|
Term
|
Definition
| bias toward responding in a particular way because of previous response made |
|
|
Term
|
Definition
| systematically changing the order of trials for diff. Ss in a balanced way to counter the biasing influence of any one order |
|
|
Term
|
Definition
| people kept in dark "blind" to our hypothesis and the specific conditions they are viewing and who are trained to use our scoring criteria |
|
|
Term
|
Definition
| extent to which raters agree on the scores they assign |
|
|
Term
|
Definition
| changes in measurement materials that occur because of use, making measurements less reliable |
|
|
Term
|
Definition
mini version of study that tests procedure prior to actual study
helps by examining ODs, create clear instructions, consider automation and multiple trials, watch out for order effects, or use multiple raters with high interval reliability |
|
|
Term
|
Definition
| extraneous cue that guides or biases a participant's behavior |
|
|
Term
| Strack, Martin and Stepper (1988) |
|
Definition
facial muscles used for smiling provide neurological feedback
example of researcher watching pen in mouth producing demand characteristics |
|
|
Term
| Participants (Re: demand characteristic) |
|
Definition
| bring w/ them certain attitudes |
|
|
Term
|
Definition
| sensitive to being studied |
|
|
Term
|
Definition
| people provide socially acceptable response |
|
|
Term
| enviro. demand characteristic |
|
Definition
| random noises, changes in lighting |
|
|
Term
| measurement demand characteristic |
|
Definition
| think opinion survery as personality test--> answer ideally |
|
|
Term
| experimenter demand characteristic |
|
Definition
| formality might inhibit ppl, experimenter expectancies! |
|
|
Term
| Controls for demand characteristics |
|
Definition
1.) provide few cues as possible (single blind procedure), instructions simple, hide threatening equipment/comments, double blind
2.) make cues that must be present as neutral as possible, encourage participants to respond naturally and honestly |
|
|
Term
|
Definition
| familiarize participants w/ procedure before beginning actual date colection |
|
|
Term
|
Definition
| extent to which the measurement task engages participants |
|
|
Term
|
Definition
measure Ss' behavior w/o making them aware that the measurement is being made
placebo! |
|
|
Term
|
Definition
| define participants in terms of characteristics we require for them to participate in study |
|
|
Term
|
Definition
| arises b/c sample contains only those indivs. who are willing to participate |
|
|
Term
|
Definition
| bias occurs when participants are knowledgable about research |
|
|
Term
|
Definition
| different group of participants is randomly selected for each condition of IV--> compare groups |
|
|
Term
| problems with random assignment |
|
Definition
works less well w/ small groups
NOT guaranteed to balance Ss variable within each condition
Balance out variable effectively--> variable fluctuates within each condition |
|
|
Term
|
Definition
| across variable combine scores from diff amounts of categories of that variable |
|
|
Term
|
Definition
| each Ss in 1 condition "matches a participant in the other conditions on one or more Ss variable |
|
|
Term
| pros and cons of balancing |
|
Definition
Pro: be sure NO confounding
Con: might NOT find predicted relationship, increase error variance and pretesting may communicate demand characteristics |
|
|
Term
| pros and cons of Matched group design |
|
Definition
Pro: greater internal validity
Con: Pretesting create diffusion, increase vaiability of score which leads to less power, small N |
|
|
Term
| pros and cons of limit pop. based on that variable so variable is constant |
|
Definition
pro: increase internal validity y eliminating potential confounds, increase power by reducing error variance
con: too selective can overly restrict range of scores, decrease external vaildity |
|
|
Term
| within subjects/ repeated-measures design |
|
Definition
| Repeatedly measure the same participants under all conditions of an independent variable |
|
|
Term
|
Definition
| participants are measured on the depenedent variable before they experience the condition of the IV and then again after they receive the treatment |
|
|
Term
| pros and cons of repeated-measures design |
|
Definition
Pro: Should eliminate potential confounding from virtually all participant variables
Cons: indiv. change even from moment to moment--> become aware of all conditions--> diffusion of treatment--> demand characteristics |
|
|
Term
|
Definition
| loss of subjects b/c participation dies out before study is completed |
|
|
Term
| complete counterbalancing |
|
Definition
| balancing order effects by testing different participants using all possible orders |
|
|
Term
| pros and cons of complete counterbalancing |
|
Definition
pros: all possible orders present --> NO bias
cons: very complex design |
|
|
Term
|
Definition
| balancing of order effects by using only some of possible orders, systematically changes position of each condition in sequence but NOT before/after condition |
|
|
Term
| pros and cons of partial counterbalancing |
|
Definition
pro: simpler!
con: NOT balance carry-over effects or response sets |
|
|
Term
| nonsymmetrical carry-over effects |
|
Definition
| carry-over effects from one order of conditions do not balance out those of another order (btwn subjects design) |
|
|
Term
| Other design considerations that prevent repeated measures |
|
Definition
Consider if for one condition might produce permanent changes in behavior so one order of conditions has unique carry-over effects
Don't underestimate influence of subject history, maturation and mortality
Repeated- measures design often requires more stimuli and other testing material than a between-subjects design |
|
|
Term
| Methods for Controlling Participant Variables |
|
Definition
| Random assignment, balancing variable, matching groups, limiting population, repeated measures |
|
|
Term
|
Definition
| when characteristics of sample data are different from population they represent |
|
|
Term
|
Definition
because of sampling error, sample data poorly represent the absence of the predicted relationship in the population
Reject if its too unlikely that we'd obtain data if a real relationship does not exist in nature |
|
|
Term
|
Definition
| sample data reflect presence of the predicted relationship in the population |
|
|
Term
|
Definition
| manipulating one independent variable |
|
|
Term
| parametric inferential procedures |
|
Definition
| used w/ interval/ratio scores that are normally distributed |
|
|
Term
| nonparametric inferential procedures |
|
Definition
| used with interval/ratio scores that are not normally distributed or with nominal/ordinal scores |
|
|
Term
|
Definition
| parametric one-way design, performed when we examine only two conditions of one independent variable |
|
|
Term
|
Definition
ranked scores testing two conditions without matching/repeated measures
nonparametric ordinal scores with two conditions with between subjects analysis |
|
|
Term
|
Definition
ranked with more than two conditions w/o matching repeated measures
nonparametric ordinal scores with 3 or more conditions with between subject analysis |
|
|
Term
|
Definition
ranked with 2 conditions using matching/repeated measures
Nonparametric ordinal scores within subject analysis with two conditions |
|
|
Term
|
Definition
ranked with more than 2 conditions using matching/repeated measures
nonparametric ordinal scores within subject design with 3 or more conditions |
|
|
Term
|
Definition
| Not matching/repeated measures design, nominal scores with two or more conditions |
|
|
Term
|
Definition
| too unlikely to occur by chance--> assume represents relationship in population |
|
|
Term
|
Definition
| Way to factor in sampling error, describes a range of populations means |
|
|
Term
|
Definition
| Reject null hypothesis but it is corect |
|
|
Term
|
Definition
| Retain Null hypothesis, but false |
|
|
Term
|
Definition
| probability of making Type I error |
|
|
Term
|
Definition
| reject null hypothesis, results not too likely to occur through sampling error |
|
|
Term
|
Definition
probability that we will reject null hypothesis on those occasions when null is actually false (Type II) concluding that sample data reflect a real relationship
Don't add power after study is completed |
|
|
Term
|
Definition
1.) Seek large differences in scores and means between conditions by using a strong manipulations and obtaining sensitive measurements
2.) Minimize error variance within conditions by building in controls that eliminate influence of extraneous variables from participants, experimenter, environment, or measurement task that create inconsistency in scores
3.) Greater # of Ss, N=30 needed for min. but greater than 121 no sig. diff
4.) Use powerful inferential stats |
|
|
Term
|
Definition
| Predict mean larger, predict relationship is positive or negative |
|
|
Term
|
Definition
| doesn't predict mean, satisfied with ether type of relationship |
|
|
