Skip to main content

Full text of "ERIC ED246649: Conditional Reasoning and Content Effects: A Developmental Analysis of Wason's Selection Task."

See other formats


ED 246 649 



♦DOCUMENT RESUME 



'014' 404 



AUTHOR 
TITLE* , ^ 

PUB DATE 
NOTE ' ■ ■ 



PUB TYPE 



EDRS PRICE 
DESCRIPTORS 



IDENTIFIERS 

4 

I 

ABSTRACT 



Ward, Shawn L^; Overton, Willis F, 
Conditional Reasoning and Content Effects: A 
Developmental Analysis of Wason's Selection Task. 
13 Apr 84 

"12p. ; Paper presented at the Annual 7 Meeting of the 
Eastern Psychological Association (Baltimore, MD, 
April 13, 1984). < ' j-\ 

Reports - Research/Technipal (143)*-- 
Speeches/Conf terence Papers (150) 

? ' / ■ , • V' . 

MF01 Plus Postage. t>C Not Available from EDRS. 
♦Abstract Reasor^iiig; Adolescents; Age Differences; 
.♦Cognitive development; *Deduption; ^Language / 
Processing; ^Logical Thinking; Swnantics 
*Copditionals / 




*' A study investigating developmental (differences int> 

the ability to reason with conditional propositions used five .' '/£-' 

variations of Wason's selection task to assess* conditional reasoning 
in 132 eighth, tenth, and twelfth grade /adolescents. In addition to 
examirring developmental differences, th^ study had as an objective to 
examine the role>- of semantic content as a Moderator of? logical 
competence. An experimental group and ya control p group were used; with 
the experimental group performing all/five Wason tasks, including 
$s meaningful content problems and abstract content problems, ^and the 
control group receiving only the twa abstract content problems. The - 
results demonstrate developmental differences in the conditional 
relationship of deductive reasoning. The semantic relevance of 
materials moderated performance such that the twelfth graders' J 
logical competence was assessed when the original abstract content of 
materials was replaced with familiar semantic content. The probleijrS . 
. and statistical data on answer frequencies fori all grades are 
appended. (MSE ) / 




if 



****** to***********!*****************************************************. 

* Reproduction^ supplied by EDRS are the best that can be made '*•' 

*, / / from the original document. \ ■ ,, * 

*************** V** ****************** ***** ******** ************* ********* 1 



ERIC 



V 



i Conditional Reasoning and Content" Effects! 

A Developmental Analysis of Wason's Selection Task 



Shawn L.' Ward " 
Willis F. Overton 



* »!2! il,8S,ON T0 REPRODUCE THIS 
MATERIAL IN MICROFICHE ONLY 
HAS BEEN GRANTED BY 

■ v • 

r z? ■ 

to the Educational resources 

INFORMATION CENTER (ERIC),- 



U.8. OEPARTMENT OF EDUCATION 

NATIONAL INSTITUTE OF EDUCATION 

EDUCATIONAL RESOURCES INFORMATION 
CENTER (ERIC) 
/€ This documont has been reproduced as 
receivod from tho person or organization 
originating it. 
LI Minor changos have boen made to Improvo 
reproduction quality. 

• Points of View or opinions stated in this docu ■ 
merit do not necessarily represent official^NIE 
position or policy.. 



Annual Meetihg of the Eastern' Psychological Association 



April 13, 1984 



A rnojor acontomporary area of investigation in the field of cognitive 

•/ . . .... . . 

development concerns the, development of Reductive, reasoning. In deductive 

reasoning a particularly important focus of interest has been ^thq conditional 

relationship, - i.e., if p then q. This interest is warranted for the conditional 
. ' \ • ' . * • 

is central to logic" 'itself. As Braine (1978) and others have noted, the 
v f V : . ■ . ' 

-conditional is ,t£e sole logical connective that is, parallel to the metalogical 

concept of inference! and is the "heart 'of logic/* The developmental interest 

in the conditional relationship of deductive reasoning has focused upon the' 

questions of when the skill is acquired and how the skill is influenced by* 

various task and situational factors. 

^ In the language of logic the if p then q or conditional relationship 

consists of two components. The component between the '"if" and/the "theft" is , 

called the antecedent proposition or p alternative. The component following *the 

"then" is called the consequent proposition or q alternative. For example in - 

the implication rule, "if • you\ mow the lawn, then you may go to the movies" the 

phrase "you mow the lawn" is the antecedent ;atid v the phrase "you may go to tttfe 

if ' \ * 1 

movies** %s the consequent. What distinguishes the conditional 'from other'' 

■ ' ' : ■•- ■ ■ \uT ' • , ' .V - ] 

logical ofrferators is that its antecedent implies its consequent . m It does not 

assert that its antecedent is true, but only that if its antecedent is true ■ 

,then it's consequent is true also. It also does not assert that\its consequent • 

is tru^? but only that its consequent is true if its antecedent is* true. The. 

proper strategy to employ with respect to conditional joroposir ions oV what is' ; 

also fendwn as material implication is a falsification Strategy. That is, in 

order t*^ reason conditionally, it is necessary and sufficient to understand . '/ 

i that tljesp and q instances falsify the implication. Therefore, a test of an V 

hypothesis is its falsification with a' counterexample. AIL other cases will 

• • • ■■ > • ,/ . >>' •' . : / •-• • 

prove material implication! true. " ' , . ■ f " , , <, 



* ■•; Piaget's ^developmental theory as well as recent empirical work 
such as that conducted by O'Brien and Overton suggest that the competence to ■ 

■ ' , - ' ft* i ' ' ' ' \ • 

reason with conditional propositions is first evidenced in adolescence, Piaget 
-argues that to reason successfully ^requires the dovolopraont of the formal 
operational' though): stifuc^ures of the 1NRC group. The theory maintains that 
it is* only, with the coordinated transformations bf formal operational thought 
that the afiility t^o 'reason systematically with logical; relationships within a 
complete propos it ional system should be expected. It is the formal operational 
structure that enables someone presented with } the conditional, if p, then, q, to 

recognizee b©tK r tfle inverse, p and >q, and the reciprocal, if q then p, £o the 

' . '. ' • ' ' v • p ' •' ' ^ ' * 

conditional. Indeed, the hallmark of formal ^operational thought, is a systematical 

/ i ■ y ' p , ',■ . ■ ' ' ' I ' i 

understanding of the conditions with which, the conditional Can h!e tested for 

its truth or falsity, y * , \ *■ 

..' * I ' . . ' .. ' * .' f V.' ■ ' ■'. . ' ; ■ 

, Despite the evidence that conditional reasoning competence is 



acquired, during adolescence, a number of »st&dies" have" shown that poor test 
performance often extends well Ait p adult years. Overton has i ( presented a 
general competence - moderator - performance model ■ "which looks *at the interplay 
between the underlying competence and various moderator variables thAt mediate 
intimate "performance. Thejse Moderators, refer to factors which r fijfetfer; enhance, 
utilization or ar^determinant^ of< errors^. The poor perf orm^nie^&Jhat . extends ' 



into adult year^ has moSt- frequently been found" in what has i>een called the 
Wason selection task. W^spnyand Johhson-Laird theihs elves' suggest that this is 
a task d-esi^ned' to tap not specif iiraljy the underlying competence but rather • 
moderators of perfo^mantre.' As they say . .we were interested in their . .v 
determinants, trie rf actors whi'ch governed 'pferformapcfe and made it fail to reflect 
logical competence." ' ' *!■ , ^ ." * : ' 4 - : ^ J\- - \, ^ 



The Wason selection task presented a conditional reasoning, task 



consisting 'of sytnlpolic materials! and affirming antocodonts and consequents ' 
to .test performance on an implication rule. Subjects' wore shown four cards 
... displaying" on their, facing sides,, a D; an P, a 3, ancj a 7. ^iven tho 

* ■ ' . 1 ' r i 

1 ! ■ . 1 ' { 

conditional ?rulc, "If thore is a n on ono side of a caisd, then there is a 

■ " ' * ' - J ■ : : ' / • • e . 

3 on the other/ 1 subjects woro to 'test the truth or falsity of tho rule. 

. '■- 1 .4 .' • . * * ' : 

'. • , . | : > " * • * 

Performance on this selection task resulted in a dismal 10% correct * seloctiotf 
rate among intelligent b adults. TWo errors were common. Al thougty. most sub jocts 
selected the alternative, they failed to Select tho q alternative &nd frequently " 
selected the q alternative instead. V Vl' 

• An important issue then is a question of the determinants of poor 
performanc^on this problem despite the fact that they 'demonstrate /conditional 
reasoning on other tasks. Although a number Qf • alternatUyes have offered' 
potential solutions to the problem, a current, important, arid active • area Qf 
' research focuses on the semantic content. of the conditional problem. - ' 

The specific aims of ' the present study are twofold. .First, cuVrently 



r there isno available data with* respect to developmental differences on the 
selection" task. This study «is designed to obtain such developmental data. 
Second, the study examined the rfele of senjantic content as r a moderator of 
logical. -competence. v - f. l0 ^ ' 

, - 'Griggs -and Cox, amc^ng others, have demonstrated that age relevant 4 
semantic content cues relevant experiences, from memory. Tliis familiar content 
facilitates adult "Performance on- Wason's .selection^tasjc, The meaningful ■ ' \ 
-semantic material chosen for this project Was age-appropriate for the population 
= .studied. . ' T . V • V, . , • > 

• : ' " The gubj'eQts were eighth, tenth, and twelfth graders, forty-iibur ' 

from each grade level.-- , Five variations of Wason's. selection task were us'ed to. ' 
^ assess" conditional, reasoning, (PLEASE REFSR TO'*TJiE.- FIRST HANDOUT- AS IT LISTS ' 



o 

ERIC 



., *• ' ' . • ■ /■ .. ;. . • >■ ..... t. . - .. .v. 



THESE FIVE VARIATIONS. ) Throo of those problems employed meaningful content. 
These woro called (1) tho Drinking Ago Problem, (2) the Motor Vohiolcj 
Problem, and (3) the School Problem, The remaining two problems w<^ro' drawn 
directly- from Wason and consisted of abstract content* J Theso woro tljo Vowel 
Problem and tho Abstract Problem. 

. ' Each problem was presented on' a sepeyrate page of a booklet., Using 

., 'i , 1 ' i / 

tho Drinking Age Problem as an oxampie, the format of each problem was as 

i ' ' ' ' ' * ' 1 I ' / ' 

foHows. At the top of each page wore four response alternatives, prinking 

Beer, the p alternative; 22 years of #ge f the q alternative;,. Drinking' Coke, 



; the not-p alternative; and 16 years of age, the not-q alternative;/ each 



depicted on cards. The subjects -werer instructed that each ca^d has information 



relevant to a person sitting at a table. For this problem, the ihiforiftation^^A/* 



on one side of a card is a person's age and on the' ol&er £ide of the card, is 7 - '^j} 

, what the person is drinking^ Each problem has a conditional rule; here £he \f\ llA ^ 

» • ; ■ ? •■■ . .* ■* V 

rule ^stated! -"IP A PERSON IS DRINKING B^ER, IJiEN THE PERSON MUST BE OVER 2l. H 

The subject's task is to decide which card or cards necessarily must be turned 

over to decide whether or not the riil'e is being broken. / « 

. • ' * ' , • ■ ■ ■ : - ■ - • / ' / 

The correct Solution^to eafch problem involves flection of the 

• '. ■ ' . • ■■ _ ... , . ■■ . 1 , ■'■.'•■"./. 4 • • \ : - : J: 

p alternative and the-q alternative because 4ach. of these alternatives and i:^ 

• : . « * 

only" these alternatives can potentially demdnstrate the falsifying condition,/ 
p and not-q. , Ip other words, - turning the q. card eould potentially yield a p 
-alternative and turning* the p card could potentially yield a not-q alternative 



and t>nly the|je : t^ro dards can establish deducfcdve certainty with respect ^to ^ „ 
the .question is tlfe*rul^: being broken o£-$ncybeing broken. f 3 

Subjects were assigned, to either_an ^xp^lme^tal or contir^^group. 
The experimental group was administer ed^a^l^five se^ectioji tasks and^jj^ .* ■ 
meaningful set always preceded Vtl\e abstfac^sefe^ The control g?oup received 1 




combinations acrhnn yrado 1 audi a .i;'or the Drinking Agu Probion. Tho davolopnumtaJ 
. ■* ** 

c\if fcronoV for*, tho* proportion correct this problem is between tho oig'tith 

and twolftm iiradtia. Tho correct selection combination for tho Motor Vehiblo 
% f %\ * J ' # 

Problem approached significance. Tho most* frequent combination for tho School 

Problem foil into tho* "other" category. Examination of tho individual 

combinations composing this category revealed no pattern. ^ t 

Among tho abstract content task** the most frequent combination 

selected war, tho match incj ^oinbinatio^, p and cj. ( TAHLE* 3 ) . • This high nclocbidii 
, . t. : . » 

frequency of tho matching combination on abstract content tank is consistent 
with earlier work by Wason and Johnson-Laird. Thqre were no differecos across 
grade levels for the matching selection within each condition. A comparison^ 
botweon conditions revealed that only the twelfth grade experimental group 
selecbed.the matcl(ing combination significantly fewer timely than the twelfth 
grade control group. The mediation of semantic content effects on. the reasoning 
competence of the twelfth grade experimental group may have alerted them to a 
partial insight to the. logically abstract problems. 



This research demonstrates developmental differeces in the 
conditional relationship of deductive reasoning. The selection task, as designed, 
provides an assessment of possible moderators of ultimate performance. It was* 
found that semantic relevance of materials moderated performance suph that the 
twelfth graders, who have expressed logical competence on other tasks, successfully 
reaspned on problems consisting of meaningful consent. That is, logical competence 
was assessed when the original abstract content^ materials of the s.electiopi task 
were replaced with familiar semantic content. This semantic content enhanced 
logical performance, on meaningful problems as compared to logically abstract 
problems. - \ \ 



ERIC 



ft 



SANINGFLTL CONTENT PROBLEMS 



JUL 



4 



. , — Drinking Ago Problem ! 

"IP A PERSON IS -DRINKING DEER, THEN THE TERSON MUST BE OVER 21." 

* 

Motor Vohiolo Problem . >■ 

M IF A -PERSON IS DRIVING A MOTOR VEHICLE, THEN THE PERSON MUST BE OVER 16." 

\ School Problem 

\ ■ 
"IF A STUDENT HAS MOUH VlIAN 10 ABSENCES IN A SCHOOL YEAR, THEN THE STUDENT MUST 
^ REPEAL THE SO^OOL ^EAIU M ' , • 

ABSTRACT CONTENT PRODLEMS \ ' ^ 

: : : V ' ■ * : ' ' 

Vowel Problem 



)N s OI 



IF A CARD HAS A VOWEL ON. ONE. SIDE, THEN IT HAS AN ODD NUMBER ON THE OTHER SIDE. 11 



Abstract Problem 

'IF A CARD HAS A D ON ONE SIDE, VrfEN IT HAS A 3 ON THE OTHER SIDE." 



An Example of how each Problem was Presented 



(P) 



DRINKING 
COKE^ 



(q) 




16 



YEARS OE» AGE 



(P) 



** 



DRINKING. 
BEER 



-LoL 



22 



YEARS OF AGE 



Instructions! 



"Each of the above cards has information about' a person sitting at: a table. 

On one side of a card is a person's age and on the other side of the same card is 

what a person is drinking. \ + 

Here is a^rulei IF A PERSON IS DRINKING BEER, THEN . TO^ER§ON MUST BE OVER 21. 



Pick the card' or cards that you would definitely nee 
whether or not the ruij.e is be|.ng broken." 



over to decide / 



Notei The labels above each <?ard & are for illustration and did not occur in test 
materials ' 

^ ** denotes the correct selections Q N 

ERIC 



Table 1. • 

Average* Correct Selection on each 
P:ft>blem Type as a Function of Grade Level* 



r 



Drinking Ayo 
Motor Vehicle 
School 

m 

Vowel 
* Abstract 

Vowel 
Abstract 



EIG1T1I 



TEN Til 



TWEL 



ibni 



Experimental .Group 66) 



.63 
.59 
.22 
.00 
.00 



.01 
.60 
.36 
.14 
.09 



Control Group (N = 66) 
.00 .09 
.00 .09 



.95 
.06 
.10 
.14 

.14 •» 

'.09 
.09 



Notei Maximum score = 1. 



9 

ERIC 



10 



< 



Table 2, • 

Frequencies for Thematic Problems Selection Combinations 
for the Experimental Group, 





Selection ' 




i 

Grada 




Problem" 

* 


Combination 


0 

*- 


10 


12 


Drinking titjo 


P • , • 


• 

1 


i 


1 


* * 


*P»q „ 




i • 








u 








p»q 


10 


21 




other 


0 


2 




Motor Vehicle 


P 


2 


2 


2 




P»fl _ — 


1 


1 


— - 




, p#3#q 


1 


— 


— , 




p,q 


14 


15 


19 




other 


4 


4 " 


1 


School 


P 


3 


an <m 


2 




p»q _ 


2 


3 


2 




p»g»q 




1 


- 1 




p»q 


5 


8 


5 




other 


12 


10 


12 


Total 


P 


6 


3 


5 




p»q _ 


3 


5 


2 




p»q»q 


1 


1 


1- 




p»q 


32 


41 


45 




other 


24 


16 ' 


13 



< 



r 



\ 



y 




Table 3. 

Pr©<|uenai«« for Abstract Problems Selection Combination*! 



Problem 



Selection 

Combination 



Grade 
0 10 12 



experimental 



Vowel 



Abstract 



Vowel 



Abstract 



P * 

other 
P 

p#q _ 
p*g>q 
p#q 

other 




2 


2 


4 








1 




i 




> 


3 


10 


12 


9 


l 


3 


4 


12 


7 


9 * 


— * 


2 


3 


9 


10 


6 




control 




2 


3 


3* 


9 


' 9 


14 




2 


, 2 •' 


11 


' * 8 


3 


1 


4 


2 


12 


11 


14 




2 ' 


2 


9 


5 


4 ' 




12