2x2

Practthe stwo datten ProcThereeffec Initia

2-wa

Main

Initia Get unian / m / pr / pl / de

2-Factor B

The studytice Difficulty

same difficultydifficulty levelsndance was re

cess: e are a lot of

cts will require

al Analysis Get descr Determine Consider

ay Interaction Get 2-way

n Effects Get estim Why are t

al Analysis

descriptive

nova TestPe

method = sstyprint descriptivot profile(Praesign = Prac

Between G

y examined thwas a 2-cond

y as the exams. The schedecorded (1= n

steps to a coe different sub

riptive meanse what effectswhat main ef

ns y cell means &

ated marginathe “Descriptiv

e means, plo

rf by AtndRe

pe(3) es

actDif * AtndRtDif AtndRe

roups ANO

he relationshipdition variable

m problems (=dule showed tnot attend, 2=

mplete analysbsets of these

, plots & F-tess are significaffects are likel

& follow-up a

al means & fove” and “Estim

ots & F-test

ev PractDif

Rev) ev PractDif*A

OVA

ps of exam Re - practice p2). Different the class mee

= attend). The

sis of a 2-waye. Here’s a pre

sts ant ly to be intere

nalyses to de

llow-up analymated” margi

ts

AtndRev.

eview Attendroblems weresections of th

eting during we dependent v

y design. Diffeview…

esting – based

escribe the 2-w

yses to descrinal means dif

lists DorderDesc

corre get de get pl speci

autom

ance and Prae either easiehe course wewhich the exavariable was

ferent pattern

d on the aggr

way interactio

be each mainfferent ?

DV “by” IVr determines lcriptive Statistcts each effeescriptive cellot of cell meaify the designmatically calcu

The “Descr“uncorrecte The margindifferential aggregated For exampPractDif is ( (60.833 *

actice Difficulter than the exare randomly am review wouperformance

ns of significa

regations invo

on

n effect

Vs left-to-right ortics table ct for all othel and marginaans (x-axis * “ including theulates from th

riptive Statistied” means.

nal means aresizes of the c

d.

ple, the margi

12) + (57.500

ty with exam am problems assigned to reuld occur & ston an exami

nt and non-si

olved

rdering of IVs r effects al means “separate linee interaction the IVs specifi

ics” are the ra

e weighted bycell means be

inal mean for

0 * 8) ) / 20 =

performance(=1) or abou

eceive the tudent’s nation.

gnificant

in the

es” ) hat is ed above)

aw or

y the eing

r the Easier

59.500

. t

Dete

ermine what effects aree significan

nt

From thelooks likeafter prathan witheffect is attended Another pattern wusing eaperhaps attendingwhen uspracticesadvantag

We haPractiReviesignifi

e means and e performanc

actice with samh easier probllarger for tho

d the review.

way to descrwould be that asier practice

a small advag the review,

sing the similas, there is a sge to attendin

ave a significace Difficulty, w Attendancecant interacti

the plots, it ce is better me difficulty lems, and thisse who

ribe the data for when there is

antage to not whereas

ar difficulty substantial ng the review

ant effect for no effect for e and a on.

s

.

Consider what lower-order effects we will need to check for descriptive/misleading patterns Because of the significant 2-way, the means patterns of each main effect will have to be carefully checked against the corresponding simple effects to determine if they are descriptive or misleading. Remember, this will have to be done whether the main effect is significant or not – main effect nulls can be misleading! Consider what lower-order effects are likely to be interesting – based on the aggregations involved PractDif

These conditions are really pretty arbitrary. More importantly, it is unclear what population is represented by an average of those who attended and not attend

the review session! So, this main effect is only likely to be interesting if that main effect is descriptive, and so, it describes the

behavior of both those who did and did not attend the review. Attend the Review

This is a straightforward operationalization of a simple variable However, the marginal means are of dubious value, because the PractDif conditions are arbitrary, and so it is not

clear what population would be represented by the aggregate of the easier and similar difficulty performances So, this main effect is only likely to be interesting if that main effect is descriptive, and so, it describes the

behavior of both those who practiced with similarly difficult and easier materials. Remember – non-significant lower-order effects that are involved in a significant higher order effect must be compared to the corresponding simple effects, to determine whether they are descriptive or misleading!!! 2-way Interaction Pairwise Comparisons You will usually want both sets of simple effects. One of those sets will be used to describe the pattern of the significant interaction. Each set will be used to determine if the corresponding main effect pattern is descriptive or misleading. Select the set of simple effects that most directly addresses the research question or research hypothesis The statement that, “We wanted to know if the relative difficulty of the practice material was related to test performance, and if this effect was different for those who did and did not attend the review session.” makes the selection of the simple effects to use to describe the interaction straightforward. From this, we’ll want to focus on the simple effect of practice difficulty (easier vs. similar) and then examine how this simple effect is different those who did and did not attend the review session.

ObtaAtte / emm

aining and dndance

means tables

describing

s ( AtndRev by

the pairwis

y PractDif ) c

e simple ef

compare ( Pr

ffects of Pra

ractDif )

actice Diffic

this asks fthe 2-way

the order o“table” commeans

the variabtells which These arDescriptivmatch the

The F-tessimple efcondition With onlypairwise tests. Not Atten Same t2

culty for eac

for the an anainteraction

of the variablemmand contro

ble specified inh set of simple

re the same cves table aboe tables comm

sts tell us thatffect of Practic of Review At

y 2 Practice Dcomparisons

nd t2 = (9.167

= (21.875 / 3

Thin N A ThalthD

ch level of R

alysis of the c

es in parenthols the displa

n the “compae effects to te

ell means as ove, but rearramand.

t there is a sigce Difficulty fottendance.

Difficulty condare redundan

7 / 3.488)2 = 6

3.309)2 = 43.7

he pattern of nteraction is:

ot Attend Easier < S

Attend Easier <<

his interactionllows us to an

he main effectifficulty will be

Review

cell means for

esis of the ay of the

re” commandest

in the anged to

gnificant or each

itions, the nt with the F-

6.908 = F

711 = F

the

Same

< Same

n pattern nticipate that t of Practice e descriptive

r

d

e

ObtaDiffi / emm

aining and dculty

means tables

describing

s ( PractDif by

the pairwis

y AtndRev ) c

e simple ef

compare ( At

ffects of Rev

ndRev )

view Attend

this askfor the

the ord“table” means

the varcommatest

Thein t TheRebut

Witconred Ea Sa

dance for ea

ks for the an a2-way interac

der of the variacommand co riable specifieand tells whic

e cell means the “Descripti

e F-tests tell eview Attendat not Easier P

th only 2 Revnditions, the pdundant with t

sier t2 = (3.33

me t2 = (9.37

Ti E S TatAm

ach level of

analysis of thction ables in pare

ontrols the dis

ed in the “comch set of simp

will be the save Statistics”

us that the simance is significPractice.

view Attendanpairwise compthe F-tests.

33 / 3.488)2 =

75 / 3.309)2 =

The pattern onteraction is:

Easier Practi Not Attend

Same Difficu Not Attend

This interactioallows us to athe main effecAttendance wmisleading

f Practice

he cell means

nthesis of thesplay of the

mpare” le effects to

ame as given above.

mple effect ofcant Same

nce parisons are

= .913 = F

8.029 = F

of the

ice = Attend

ulty Practice< Attend

on pattern anticipate thatct of Review

will be

e

f

t

An A This Stand Why showlargemean An aany tANOdiffervariaterm”base

Why

1. Itpc

2. Ttoc

Alternative A

is a BG modedard Errors o

care? Becauwn above) thaer or smaller sn difference c

lternative is totwo cell mean

OVA error termrent design ceability, so the a” approach is

ed on the df fro

do people wh

t is based on pairwise errorcomparisons aThis approacho get SPSS tcompare thes

Analysis of

el. Although aof the follow-u

use the followat uses the stasample sizes,can be signific

o use this “fulns. This is an m for BG modells, because aggregate of that since the

om all the pa

ho like this ap

the same estr term approaare significanh allows the co provide thise means.

f Cell Means

all the F-tests ps vary with s

w-up analysesandard error i the standard

cant for one c

l model error extension of els. That assthey each rethem all is thee best estimarticipants.

pproach like it

timate of variach. So, it prov

ntly different ucomparison ofs significance

s to Describ

and follow-upsample size.

s are based on the denomi

d error can becomparison a

term” as the the “homogensumption is thpresent a same best estima

ate is derived

t?

ability, but larvides a more sing this appf nonadjacenttest, but the

be Simple E

p analyses ar

n a t-test (thanator. So, de

e larger (smalland not signifi

basis for comneity of varian

hat it makes smple taken froate of the variafrom using th

rger sample spowerful signroach (thought cells meansComputators

Effects and

re based on a

at isn’t shown epending on wler ns) or smacant for anoth

mputing an LSnce” assumpt

sense to combom different pability of each

he full design

size, and, so, nificance test,h the reverse. Sometimes

s will give us a

the Interact

a single error

in the outputwhether the cealler (larger nsher.

SD value that tion that is mabine the withipopulations thh. The extenssample, the s

uses a smalle, and more pa can happen

s, with larger dan LSDmmd t

tion

term (MSe=5

t, but how to cells being coms), and the sa

is then used ade when wen-group varia

hat all have thsion in the “fusignificance te

er standard eairwise cell mon occasion)designs, therethat we can u

58.385), the

compute it is mpared have ame cell

to compare compute the

ability from thehe same ll model errorest should be

error than the mean ). e is no easy

use to

e e

r e

Ano

Anotrun s

sort c

tempsplit f

unino /des

ther Alterna

her approachseparate analy

cases by Pra

porary. file layered by

ova testperf bsign = AtndRe

ative Analys

h to testing simyses for each

actDif .

y PractDif .

by AtndRev ev,

sis of Cell M

mple effects th partition of t

Means to De

hat shows uphe design.

escribe Sim

p in many exa

sorts

specanalysis splits

spec

mple Effects

amples is to u

s the cases by

cify that split cs command

s the cases b

cify DV “by”

The Sfrom comp

The SEMMbaseEMMcells. The dEMMbasecompn from

The MSSer

The F

Hereof praatten

sort c

tempsplit f

unino /des

s and the In

se the “split f

y the selectio

command will

by the selectio

IV (simple eff

SS effect (Atnthe EMMEAN

pares the sam

SS Error are MEANS analysed on data froMEANS were .

df-error are dMEANs analysed on n from tpared, while Em all four cell

MSerror are drror and df-err

F-values and

e is the syntaxactice difficult

ndance condit

cases by Atn

porary. file layered by

ova testperf bsign = PractD

teraction

file” option in S

on variable

l only apply to

on variable

fect variable)

ndRev) are thNS analyses me cell means

different fromses above. Thm two cells, wbased on dat

ifferent from tses above. The two cells bEMMEANS wls.

different, becaror are differe

p-values are

x to get the simty for each retion.

ndRev.

y AtndRev.

by PractDifDif.

SPSS and

o the next

he same as above. Each s

m the hese are while ta from all fou

the These are being

were based on

ause both theent.

e different.

mple effects view

r

n

e

Desc / emm

Howedesc Easie Sam

cribing the

means tables

ever, we knowriptive for tho

er Practice

e Difficulty P

Main Effect

s ( AtndRev )

w from the paose in the Eas

N

Practice N

t of Review

comp

attern of the insier Practice c

Not Attend =

Not Attend <

Attendance

pare ( AtndRe

nteraction thacondition.

Attend

Attend

e

ev )

at this is not

You shouldnot the sam“DescriptiveAttend and Also, the F-above and comparing “Descriptive Because ththe design interaction correlated).analyses usunique concontrolling f So, in a faceffects beinmarginal mregression a regressiobivariate cocriterion! The overallis:

This main ebecause it

d notice that thme as the mare Statistics” a72.08 for Att

-test for “Atndshown belowthe data meae Statistics” a

here are uneqconditions, thare all colline. Thus, like asing Type III Stribution of eafor the other e

ctorial using Tng tested are

means, the samincluding qua

on weight thatorrelation betw

l or main effec

Attend =

effect must beis potentially

he means shorginal means

above (there 6end).

dRev” in the Aw (which matcans shown in above.

qual sample she main effectear (nonorthogall other multivSS, the modeach effect to teffects in the

Type III SS, thdifferent thanme as a multiantitative variat is not the saween a variab

ct for Review

= Not Attend

e communicamisleading.

own here arefrom the

64.5 for Not

ANOVA table ch) are not the

izes among ts and the gonal, or variate el tests the the model, model.

he main n the raw dataple ables will testme as the ble and the

Attendance

ted carefully,

a

t

Desc / emm

cribing the

means tables

Main Effect

s ( PractDif )

t of Practice

compare (

e Difficulty

PractDif )

Agaheremeaabo). Thetablecorrmaibothamothe effe ThepattDiffi

Alsointe Not Atte

ain, you shoule are not the ans from the “ve (59.5 for E

e F-test matche above, becrected or uniqn effect to theh are testing tong the estimgroups, after cts in the mo

e pairwise comtern of the maiculty to be:

Eas

o, we know froraction that th

t Attend

end

d notice the msame as the “Descriptive SEasier and 76

hes what’s in tause both are

que contributioe model. Saidthe mean diffeated marginacorrecting fodel.

mparisons shoain effect of P

sier < Harder

om the patterhis is descript

Easier < Sa

Easier << S

means shownmarginal

Statistics” 6.25 for Same

the ANOVA e for the on of this d differently, erence al means of r the other

ow the ractice

r

rn of the tive.

ame

Same

n

e