PSY 301 RESEARCH METHODS
researchers use natural group designs to meet the first two objectvies of the scientific method
description and prediction
data analysis uses both
descriptive stats and inferential stats.
only the repeated measures design involves an
independent variable in which participants responses are contrasted in different experimental conditions.
internal validity
is the degree to which differences in performance on a dependent variable can be attributed clearly and unambigously to an effect of an independent variable, as opposed to someother uncontrolled variable.
In block randomization procedure, it is unlikely
it is unlikely that changes in the participants attention, motivation or experience with rating the photographs would affect any one of the exposrue condtions more than another. any one condition is not more likel to appear at the beginning middle or end of the 66 trials. the practice effects can be expected to average out across the three experimental conditons ( 100, 500, 1000 ms exposure times. )
bottom row of table 7.1 PAGE 227 labled Practice Effect (nonlinear) .
labled Practice Effect (nonlinear) . the change from the first trial to the second trial is from 0 to 6, indicating a participants performance changed dramatically after one trial. several warm up trials before actual data collection begins can sometimes offset early abrupt changes in practice effects PAGE 227 TABLE 7.1
when is the matched group design best
matched group design is better than the random group design when - a good matching variable is available and - when only a small number of subjects is available for an experiment that requires separate groups for each condition
describing the results of an experiment involves the use of
means , standard deviations and measures of effect size
meta analysis
measure of effect size. its a stat technique used to summarize the effect sizes from several independent experiments investigating the same IV or DV. used to answer questions like are there gender differences in conformity? what are the effects of class size on academic achievements?
meta analysis makes sure of
measures of effect size to provide a quantitiaative summary of the results of a large number of experiments on a research question
advantage of block randomization
- Block ran- produes groups of equal size - controls for time related variables - block random will balance ( or average) any characteristics of participants (including the effects of time related factors) across all the conditions of an xperiment. BLOCK RANDOM increases Internal validity by balancing extraneous variables across conditions of the independent variable.
constructing a latin square
- Even number of IV levels- example 4 levels -Low, Low Medium, Medium High -Randomize first order -High (1) Low Medium (2) Low (3) Medium (4) -Then add 1 to each value to get next row 4 of previous row becomes 1 -Continue this until you have a 4 * 4 square
matched group designs
- may be used to create comparable groups when there are too sfew subjects available for random assignment to work effectively. - matching subjects on the DV (as a pretest) is the best approach for creating matched groups, but scores on any matching variable must correlate with the DV. - after subjects are matched on the matching variable, they should then be randomly assignmed tto the conditions of the IV.
threats to internal validity
- threats to internal validity are threats to potential alternative explanations for a studys findings. - researchers control threats to internal validity in order to make a clear cause and effect inference about the IV. -randomly assigning intact groups to different conditions of the independent variable creates a potential confounding due to preexisting differences among participants in the intact groups - selective subject loss, not mechanical sub loss, threatens the internal validity of an experiment. - placebo control groups are used to control for the problem of demand characteristics, and double blind experiments control for both demand characteristics and experimenter effects.
Latin square design
-To balance the effects of practice effects - in the latin square, the general rule for balancing practice effects is MET. -All possible orders of the levels of the IV are presented -The number of potentially different order combinations varies depending on the number of levels N! = N factorial- if there are 4 levels of the IV then the possible combinations = 4! = 4*3*2*1 = 24 possible combinations -More levels greater the possible combinations. Often limited to 4 or fewer levels -Latin square constructs the order of presentation to counterbalance the effects of practice- not all possible combinations but making sure that every order is represented for every level of the IV -If 4 levels each level has to be first, second, third, and fourth order of presentation
why researchers use repeated measures designs
1 - conduct an experiment when few participants are available 2. conduct experiment more efficiently 3- increase the sensitivity of the experiment 4- study changes in participants behavior over time
Dittmar concluded that exposure to thin body images, such as barbie, CAUSES young girls to be dissatisified with their own bodies. they are able to make this conslusion bc ..
1. manipulated an IV that varied the images girls viewed 2. eliminated other plausible explanations through holding relevant conditions constant 3. balanced individual differences among the groups through random assignment to conditions.
data analysis
3 states - getting to know the data - summarizing the data- use descriptive stats and graphical displays to summarize what was found -confirming what the data reveal. - try to find out what is going on in data set , look for errors, make sure it makes sense. seek evidence for what the data tells us about behavior, and make our conclusions abut the data using variaous stats techniques.
the specific techniques for balancing practice effects differ for the two repeated measure designs (complete and incomplete) , but the general term used to refer to thse balanceing techniques is callllledddddddddddddddddddd
COUNTERBALANCING
controlling for practice effects ordering ******
Counterbalancing: varying the order of presentation Complete design: each participant receive the varied order of presentation Incomplete design: The order of presentation is not presented fully to each participant but across participants all orders will be presented- used if there are multiple different levels of the IV- but need to balance presentation of the order of the levels
selected orders with the 4 level independent variable from the stigma experiment. what are the two types of balancing more directly
IN A LATIN SQUARE- the general rule for balancing practice effects is MET. USE PG 231 TABLE 7.2 1ST 2ND 3RD 4TH G S C N S N G C N C S G C G N S we can see in the latinsquare, condition "G" appears exactly once in the first 2nd 3rd and 4th ordinal positions. this is true for each condition in a latin square, each cond. precedes and follows each other cond exactly once. table 7.2 shows that the order GS appears once, SG appears once and , SC and CS appears once. and so on for every combos of conditions
to confirm whether the IV has produced an effect in an experiment, researchers use
INFERENTIAL STATISTICS bc of the nature of the control prove=ided through random assignment in experiments random assignment DOES NOT eliminate the individual differences among subjects , random ass simply balances or averages the individual differences among subjects across the groups of the study. INFERENTIAL STATS- allow researchers to test whether differences btw group means are due to an effect of the IV , not just due to chance (error variation) two types of inferential stats to make decisions about whether an IV has had an effect 1. null hypothesis testing 2. confidence intervals
when there are no confoundings , an experiment has
INTERNAL VALIDITY
two methods for selecting specific orders in the incomplete design are
LATIN SQUARE and Random Starting Order with Rotation
Balancing practice effects in the incomplete design
Practice effects are balanced ACROSS subjects in the incomplete, rather than for each subject as in complete. the rule for balancing practice effects in the incomplete design is that each cond. of the exper. must be presented in each ordinal position (first second third ect) equally often best method for balancing practice effects in the incomplete des with 4 or less conditions is to use alll possible orders of the conditions two methods for selecting specific orders to use in complete are LATIN SQUARE and Random Starting Order with Rotation whether using all possible orders or selected orders, participants hsould be randomly assigned to the diff sequences.
what is the ultimate test of the reliability of an experimental outcome
REPLICATION
repeated measure designs require fewer participants than an independent group desins TRUE OR FALSE
TRUE
Independent Group Design Outcome of Interest Effects of receiving a behavioral intervention on alcohol behaviors - random assignment -matched groups
Threat to internal validity: Individual differences may account for the outcome rather than manipulation of the IV. Holding effects constant, balancing and randomization balances these effects; but there may be outcomes of interest where there are multiple levels of the IV or when the DV is measured mutliple times or both
When can you say experiment has INTERNAL VALIDTIY (and what does that mean)
When the 3 conditions for CAUSUAL INFERENCES are met 1. covariation 2. time order relationship 3. elimination of plausible alternative causes then you can say INTERNAL VALIDITY IS MET because the IV caused the difference in behaivor as measured by the DV = IV caused the DV
if the three groups had differed on a factor other than the picture books, then the results of the experiment would have been interpretale. suppose the participants in the barbie condition had heard a diff story, we wouldnt know wheter the observed difference in the girls body disattispaftion was due to viewing the images of barbie or to the different story. When the independent variable of interest and a different potential IV are allowed to covary....
a confounding is present.
manipulation
a control technique. to test their hypotheses about girls body disattisfaction. the dv was a body shape dissatisfication score
when the cofounding occurs...
a plausible alternative explanation for the observed covariation exists, and therefore, the experiment lacks Internal Validity. Plausible alternative explanations are ruled out by holding conditions constant.
the sensitivity of an experiment refers to the
ability to detect the effect of the independent variable even if the effect is a small one. an experiment is more sensitive when there is less variablity in participants responses within a condition of an experiment, which is less error variation. - the increased sensitivity of repeated measures designs is attractive to researchers tho study IVs that have small and hard to see effe3cts on behaivor.
advantages of random assignment
all the individual differences are balanced, SOOO we can rule out alternative explanations due to any individual differences among participants.
disaddvantage of ABBA counterbalancing :anticipation effects
anticipation effects occur when a participant develops expectations about which condition occuts in the next sequence, as we suggest you might when seeing this sequence suppose you are presented the following sequence Mod Fast Slow Slow Fast Mod Mod Fast Slow Slow Fast Mod Mod Fast Slow Slow Fast Mod a participants respo se to a cond. may be influenced by more expectations than by the actual experience of the cond. itself. for ex, imagine a partricipant "relaxing" after a Slow condition, anticipating a second Slow cond. immediatley
Disadvantages of ABBA
appropriatly used ONLY when practice effects are linear. if they are linear , the same amount of practice effects is added too or subtracted from performance on each successive trial. one "unit" of hypothetical practice effects is added to performance on each trial. because there would be no practice effect associated with the first trial, the amount of practice added to trial 1 is zero. trial two has 1 unit of hypothetical effects added bc of participants experience with the first trial, in trial 3 there are two units added bc of participants experience with two trials , and sooo onnnn.
When deciding to use all possible orders...
atleast one participant must be tested with each of the possible orders of the conditions. meaning-> at lease one participant should recieve the GSCN order, at least one participant should revieve the GSNC order and so onnnnn the use of all possible orders requires at least as many participants as there are possible orders. so if there are 4 conditions in the ex, at least 24 participants are needed ( or 48, 27 , or some other multiple of 24)
Random assignment to conditions is used to form comparable groups by
balanceing or averaging subject characterisitics (individual differences) across the conditions of the independent variable manipulation
thesecond balance technique using selected orders requires you to
begin with a random order of the conditions and to rotate theis sequence systematically with each cond. moving one position to the left each time. USE TABLE 7.2 PAGE 231 USING random starting order with rotation effectively balances practice effeccts bc like the latin squeare, each cond appears in each ordinal position. however, the symstematic rotation of the sequences means that each cond. always follows and always precedes the same other conditions (N always comes after C and before G) which is not like the Latin Square technique.
another term for independent groups design is
between subjects design both terms are used to describe studies in which groups of participants are compared and there is no l=overlap of participants in the groups of the study, each participant is in only one condition
researchers have two choices in deciding how to arrange the order in wchich the treatments in a complete design are administered:
block randomization and ABBA counterbalancing
block randomization
block randomization balances subject characteristics and potential confoundings taht occur during the time in which the experiment is condructed and it creates groups of EQUAL SIZES.
block randomization vs simple randomization
block randomization that each condition of the experiment is represented an equal number of times at all stages of practice - ex early, middle, late. simple randomization procdure - its possible that one coni=dition may appear more frequently , by chance, at one stage of practice compared to anotherrr. - if this occurs , we cannot assume practice effects are balanced across conditions for a participaant. when many random sequenses are used, the chances are less that any one conidtion will apeear more often at the beg, middle or end of the trials. , also any differences in practice effects should be balanced when results are summarzied accross all subjects in the experimetn. unlike block rando- ******************simple rando of orders is used to present conditions in an experiment, many diff random orders must be used **********
ABBA counter balancing
can be used to balance effects in the complete design with as few as two administrations of each condition ABBA counter involves presenting the conditions in a random order followed by the opposite of that order. - even number of repititions is required. secquense with only two conditions (A and B) in the experiment, so ABBA counter experiment with 3 conditions, secquence of trials could be ABCCBA, order of the three conditions on the first three trials were reversed for trials 4 to 6 USE TABLE 7.1 ON PAGE 227
block randomization is a common procedure for a
carrying out a random assignment in block randomization - we assign subjects to contitions one block at a time.
experimental control allows researchers to make the
causal inference that the IV *caused* the observed changes in the CV..
the two types of repeated measure designs ______ and _________ , differ in the specific ways they control for practice effects
complete and incomplete
holding conditions constant is a ______ that researchers use too ___
control technique, that researchers use to avoid confoundings a factor that is held constant, such as the story the girls heard, does not change in the experiment. bc it is constant, it cannot possibly covary withh the manipulated IV, nor the DV. SO researchers can rule out factors that are held constant as potential causes for observed results.
in an independent groups design,
each group of subjects participates in a different condition of the independent variable
Block randomization can be used to order the conditions for
each participant in a complete design. ex . study on "first impressions " when participants were asked to rate photographs of ppl on a personality trait, such as likeability or trustwothiness and the photos were presented very briefly. researchers used three exposure times in milliseconds which are 100 ms, 500 ms and 1000 ms , with such short presentation times, it would not make sense to use a between group expeiment. -in this experiment all participants rated 22 faces at each oth the three exposeure times for a total of 66 facial ratings, the experiment was reapeated 5 times, once for each trial, and diff participants rated only one personality trait for the 66 photos. to use a block randomization- the 66 trials would be broken up into 22 blocks of 3 trials. each block would contain the three condtions of the experimentation in a random order. - IN GENERAL the number of blocks in a block randomiztion procedure is equal to the number of times each condition is administered, and the sizze of the block is equal to the number of conditions in the experiment PAGE 225 FIGURE.
the goal of random assignment is to establish
equivalent groups of participants by balancing, individual differences accross the conditions. when random assignment is used to balance these individual differences across the groups, we can logically rule out the alternative explanation that differences we obtain between the groups on the DV are due to characteristics of the participants. when we balance a factor , such as body weight, we make the three groups equal , interms of their average body weight.
the nonsystematic (random) variation due to differences among subjects within each group is alled
error variation - the presence of error variation poses a potential problem bc the means of the different groups in the experiment may differ bc of the error variation, not bc the IV has an effect.
the individual differences studied in the natural groups design are usually confounded- groups of individuals are likely to differ in many ways in addition to the variable used to classify them
ex indivduals who divorce and individuals who stay married may differ with respect to a number of characteristics other than marital status. for ex, their religious pracices or financial circumstance. the manipulation done by "nature" is the rarelly controlled type we have come to expect in establishing the internal validty of an experiment nature does the manipulation because nature causes something to happen for them to get divorced. this is controlled because researcher doesnt manipulate it or cause this.
effect size d
examines the difference btw 2 group means relative to the average variablility iin the experiment the mean difference between 2 groups is always relative to the aerage varability in the participant scores.
replication
repeating the procedures used in a particular experiment in order to determine wheter the same results will be obtained a second time.
for block randomization, to be affective in controlling practice effects.... ******
for block rando, to be affective in controlling practice effects- each condition of the experiment must be repeated several times before we can expect practice effects to average out. in the example- each condition appeared 22 times, we would not expect practice effects to be balanced out after 2 or 3 blocks. typically all participants folllow the same block randomization sequence. - each trial was preceded by a fixation point (+) for 500 ms at the center of the screen. after the photo briefly appeared, the participants made a trait judgement (yes or no to this person is trust worthy), and then a confidence judgement on (a 7 point scale).
for when the use of all possible orders is not practicallllll-
for ex. w we wanted to use the incomplete design to study an IV with 7 levels, we would need to test 5040 participants if we used all posssible orders. one participant for each of the possible orders of the seven conditions is (7! orders)
matched groups design example
if the DV in the experiment is blood pressure, participants should be matched on blood pressure prior to the start of the experiment. the matching is accomplished by measuring the blood pressure of all participants, and then forming pairs or triples or quadruples of participants, depending on the number of conditions in the experiment, who have identical or very similar blood pressures. at the start of the expirement, participants in the different groups will have , on average, equivalent blood pressure. so researchers can then reasonably attribute any group differences in blood pressure at the end of the study to the treatment, presuming other potential variables have been held constant or balanced.
counterbalanccing for complete and incomplete design
in complete deesign- practice effects are balanced for EACH participant by administering the conditions to each participant several times, using different orders each time -- each participant can then be considered a "complete" experiment. in incomplete design- each condition is administered to each participant ONLY ONCE. the order or administering the conditions is varied across participants rather than for each participant.
error variation
in general, participants in a repeated measure desin will vary within themselves less over the time of an experiment than participants in a random group design will vary from other participants - there is usually more variation between people than within people. - error variation will generally be less in a repeated measures desin - the less error variation- the easier it is to detect the effect of an independ variable.
practice effects
in repeated measure design, participants can undergo changes during hte experiment as they are repeatedly tested. participants may improve with practice, bc they learn more about task for bc they become more relaxed, for examples. OR they get worse with practice, for ex bc of fatigue or reduced motivation these temporary changes are called practice effects **practice efffects that participants experience due to repeated testing in the repeat measure design cannot be eliminated . when balanced across the conditions, practice effects are not confounded with the IV and the results of the experiment are interpretable. practice effects may threaten internal validtiy of a repeated measure experiment when the different conditions of the independent variable are presented in the same order to all participants.
how to know if the IV is responsible for the difference in random group designs
in the random groups design, the groups are formed using randomization procedures such that the groups are comparable at the start of the experiment. if the groups perform differently following manipulation, and all the other conditions were held constant, it is presumed that the IV is respoonsible for the difference.
Mechanical subject loss is
in which an error in procedures cause the loss of a subject, less problematic than selective subjectet loss.
experiments involving IV whose levesl are selected, like individual differences variables, are called
natural group designs. used in situations in which ethical and practical constraints prevent us from directly manipulating independent variables. ex interested in effects of a surgery, but cant perform surgery on a randomly assigned group of participants, and then compare symptoms with those who did not recieve surgery. we can compare the ppl who have had the surgery vs those who had not- in natural group design. use subjects who naturally belong to the group
a simple randomization procedure is ___ as controlled as block randomiztion
not as controlled simple is less controlled than block
Participants performance in repeated measures designs may change across conditions simply because of repeated testing (and not bc of the ______,,, ) and these changes are called
not because of the independent variable called practice effects.
the most effective independent groups design is
one that used random assignment of subjects to conditions in order to form comparable groups proior to implementing the independent variable.
the matched group desin is an alternative to the random group design when
only a small number of subjects is available, when a good matching variable is available and when the experiment requires separate groups for each treatment.
in the incomplete design , each participant is given each treatment ---
only once SO the results for any one participant cannot be interpreted bc the levels of the IV for each participant are perfectly confounded with the order in which those levels were presented. for ex, the first participant in an incomplete design exp might be tested first in the experimental cond. "E" and second in the control cond. "c" . any differences in the participants performance btween the exp and control conds. could be due to the effect of the IV or to the practice effects resulting from the EC order. to break this confounding of the order of conds. and the IV , we can administer different orders of conds to diff participants. for ex administer the conds of our incomplete design to a second participant in CE order, testing the control cond. first and the experimental cond. second. - in this way, we could balance the effects of order accross the two conds. using two participants instead of oneeeeeeee.
threats to internal validty in repeated measure designs
participants may change over time. the repeated testing of participants in the repeated measure design gives them practice with the experiental task. so as a result of practice, participamts may get better or worse. - changes in participants in the repeated testing is called practice effects. in general- practice effects should be balanced acrosss the condtions in repeated measures design so that preactice effects average out across conditions.
demand characteristics and experimetner effects can be minimized through the use of proper experimental procedures but can be best controlled by using ________and _________
placebo control and double blind procedures.
in an incomplete design,
practice effects MUST be balanced by varying the order in which the conditions are presented. Each condition of the exper must appear in each ordinal position . 1st 2nd 3rd 4th equally often if appropriate balancing is carried out then we are able to determine wheteher the IV , not practice effects, infllueenced the participants behavior , DV.
balancing practice effects in the complete design
practice effects are balanced in complete designs within each participant using block randomization or ABBA counterbalancing. in block randomization- all of the conditions of the experiment (a block) are randomly ordered each time they are presentted. In ABBA counter balancing0 a random sequence of all conditions is presented, followed by the opposite of the sequence. Block randomization is preferred over ABBA counterbalancing when practice effects are not linear or when the participants performance can be affected by anticipation effect --- ABBA needs linear bc its a sequence In complete- participants are given each treatment enough times to balance practice effects for each participant. - when the task is simiple enough and not too time consuming, it is possible to give one participant several experiences with each treatment. -- SO in some complete designs, only one or two participants are tested, and each participant experiences literally hundreds of trials.
what is the most common method of forming comparable groups what is the most commmon technique for carrying out random assignment
random assignment - its an attempt to ensure that the differences among subjects are balanced, or averaged, across conditions in the experiment. block randomization is the most common tech for carrying out random assgment.
Random assignment DOES and DOES NOT ??
random assignment DOES NOT eliminate the individual differences among subjects , random ass simply balances or averages the individual differences among subjects across the groups of the study.
what are three independent groups designs
random group design matched group design natural groups design
when random assinmment to conditions is used, the indepent group desin is called a
random group design. Groups are formed so as to be similar on all important characteristics at the start of the experiment. next, in the experiemnt itself, the groups are treated the same except fo rht elevel of the IV. THUS any difference btw the groups on the DV must be caused by the IV.
EXTERNAL VALIDITY
refers to the extent to which findings from a research study can be generalized to individuals, settings and conditions beyond the scope of the specific study.
use when it is more effective to have each subject participate in all the condtions of an experiment. this is called
repeated measure designs- or within subject designs. in a independent group desins, a separate group serves as a control for the group given the experimetnal treatment. in a repeated measures design, subjects serve as their own controls bc they participate in both the experimental and control conditions.
when wew balance a factor such as body weight, we make the three groups equal in terms of their average body weight. for holding body weight constant,
require all the girls in the study to have the same body weight
explain approaches for drawing causual inferences in the natural groups design
requires that individual differences be studied in combo with IVs that can be manipulated. this combo of more than one IV in one experiment requires the use of a complex design.
to avoid anticipation effects
researchers generally consider ABBA counter balancing only when the number of conditions and the number of repetitions of each condition are relatively smallllll. - - even number of repetitions of conditions is required in an ABBA procedure these concerns above, make block randomizaion a better counterbalance in most complete repeated measure designs
in most matched group desisn- a pretest matching variable is used to match subjects. the challenge is toooo
select a pretest variable that equates the groups on a dimension that is relevant to the outcome of the experiment. the matched group design is only useful when a good matching variable is available. the most preferred matching variable is one that will be used in the experiment itself.
selective subject loss occurs when
subjects are lost differentially across the conditions and s(c)ome characteristic of the subject that is related to the outcome of the experiment is responsible for the loss. -prevent this by restricting subjects to those likely to complete the expirment successfully or - we can compensate for it by selectively dropping comparable subjects from the group that did not experience the loss
block randomiztion guarantees
that each condition of the experiment is represented an equal number of times at all stages of practice - ex early, middle, late.
the experimental method is ideally suited toooo.....
the experimental method is ideally suited to indentifying cause and effect relationships when the control techniques of manipulation, holding conditons constant, and blancing are properly implemented.
biggest problem in matched group design
the groups are equated only on the characteristic measured by the matching variable.
when practice effects are linear,
the number of hypothetical "units" of practice is the same for all conditions see TABLE 7.1 PAGE 227 with three conditions, 5 units are gained for each condition. for ex. the moderate condition gets the lease (0) and the greatest (5) influence from practice effects , for a total of 5 units- the sum of hypothetical units is also 5 for the fast (1+4) and slow (2+3) conditions.
when a experiment calls for only a few repetitions of conditions in a block randomiztion procedure......
the researcher should usee more than one random sequence of conditions.
repeated measures designs cannot be confounded by individual differences variables becauseeeeee
the same individuals participate in each condition, level, of the IV
describing the results
the two most common descriptive stats used to summarize results are the MEAN and STANDARD DEVIATION Measures of effect size indicate the strength of the realtionship between the IV and DV , and they are not affected by the sample size. meata analysis uses measures of effect size to summarize the results of many experiments investigating the same IV or DV.
PAGE 231 TABLE 7.2 how do investigators studying stigma might counterbalance practice effects there are 4 vignettees describing individuals with a gamblind disorder, cancer, schizophrenia, or mild, non clinical stress( neutral condition)
there are 4 vignettees describing individuals with a gamblind disorder, cancer, schizophrenia, or mild, non clinical stress( neutral condition) so 4 vignettes = 24 total sequences would be required to obtain all possible orders of conditions 4x3x2x1=24 sequences= orders of conditions the first ordinal position shows this balancing most clearly: the first 6 sequences begin with gambling disorder (G) condition , and each of the next 6 sequences begins with the schizophrenia (S) and then Cancer (C) , and then neutral (N) conditions this same pattern applies at each of the four ordinal positions for ex. G cond also appears 6 times each in the second , third and fourth ordinal positions. this is the same for S C and N condtitions.
one key to the logic of the experimental method is forming comparable, similar groups. participants in each group should be similar, similar in indiviudal differences. the control technique of balancing is required bc
these facrots often cannot be held constant.
researchers use Holding conditions constant
to make sure that the IV is the only factor taht differs systematically across the groups. ex , all of the girls heard the same stroy about shopping and attending a birthday party and they al looked at their picture books for the same amount of time, all recieved the same instructions throughout the expirement and recieved the same questionnaire at the conclusion
matched group designsss
to work right, the random groups design requires samples of sufficent size to ensure that individual differences the matched group design is a good alternative when neither the random group design nor the repeated measure desin can be used. instead of trusting random ass to form comparable groups, the reascher makes the groups equivalent by matching subjects. once comparable groups have been formed based on matching, the logic of the matched design is the same for the random groups design. in most matched group desisn- a pretest matching variable is used to match subects
two types of errors that can occur when researchers use inferential statistics
type 1 error false alarm- when we claim that an outcome is statistically significant and the null hypot (no difference) is really true, we are makinga type 1 error. - saying there is a fire when there is not. when we conclude that we hve insufficeiencet evidence to reject the null hypot and it is, in fact, false, this is a type 2 error.
perferred technique for balance practice effects in the incomplete design is to
use all possible orders of the conditions. each participant is randomly assigned to one of the orders. with only two conds, there are only two possible orders . (AB and BA). with 3 conds, there are a possible of 6 orders (ABC, ACB, BAC, BCA, CAB, CBA) THERE ARE "N!" - N factorials, possible orders with N conds, where N! equals N(N-1) (N-2) ... (N-/[N-1}) !!! means factorial there are 6 posible orders with 3 conditions, which is 3! (3 x 2 x 1 =6) the number of required orders increases ALOT with increasing number of conditions. for 5 conds. there are 120 possible orders bc 5x4x3x2x1= 120 for 6 conds- 720 possible orders bc 6x5x4x3x2x1=720
advantages of block randomization***
used to randomly assign subjects to groups - Block random produces groups of equal sizes, important bc the number of observations in each group affects teh relability of the descriptive statistics for each group and it is desirible to have the relability of these measures comparable across groups - block random controlls for time related variables. ----bc experiements often take a long time to complete, some participaants can be affected by events that happen during the time of experiment. ---so in block randomiztion- every condition is tested in each block so thses time related variables are balanced accross the contidions of the experiment. - block random will balance , or average, any characteristics of participants (including the effects of time related factors) across all the conditions of an xperiment.
practice effects ccan be balanced by
using just some of all the possible orders the number of slected orders will always be equal to some multiple of the nymber of conds in the experiment. ex. to do an experiment with 1 IV with 7 levels , we need to select 7, 14, 21, 28 or some other multiple of 7 orders to balance practice effects
researchers also chose repeated measure design because some areas of psychological research require it!!!
when research quesiton involves studying changes in participants behaiovr over time, in a learning experiment for example, repeated measure design is needed. also whenever the experimental procedure requires that participants compare two or more stimuli relative to one another, a repeated meaaures design must be used repeated testing can also be used when researchers investigate the reliability (consistency) of a measure. they may obtain two or more measures of the same individuals in order to resetablish the reliability of a measure, called - test retest reliability
