09 Bayer Achtziger Etal Subliminal Implementation Intent

8/22/2019 09 Bayer Achtziger Etal Subliminal Implementation Intent

1/19

Social Cognition, Vol. 27, No. 2, 2009, pp. 183201

183

BAYER ETAL.UnconsciousAction Preparation andInitiation

responDing to sUbLiminaL cUes: Do iF-then

pLans FaciLitate action preparation anD

initiation WithoUt conscioUs intent?

Ute C. Bayer a Aja AchtzigerUniversitt Konstanz

Peter M. GwitzerUniversitt Konstanz; New York University

Gr B. MkwitzLehigh University

Frmig impemetati iteti (I I ecuter cue X, the I wi per-rm behavir Y!) i ptuate t trigger acti iitiati withut urthercciu itet ce the pecie cue i ecutere (Gwitzer, 1999).I tw experimet uig a ijutice paraigm r a categrizati tak,critica ituati (pecie i the i-cmpet) were ubimiay pre-ete a it wa tete whether thee ituati ifuece the prepara-ti (stuy 1) a iitiati (stuy 2) the pae ga-irecte behavir(pecie i the the-cmpet). Ater the ubimia preetati thecritica ituati, impemetati iteti participat hwe trgeracti preparati a a ater acti iitiati, a cmpare t ctrparticipat (stuy 1) wh ha t rme ay ga iteti at a, acmpare t participat (stuie 1 a 2) wh ha y rme ga i-teti. Thee ig ugget that rmig impemetati itetiea t autmatic acti iitiati withut urther cciu itet.

Current research on goals recognizes the importance o automatic processes (Frst-er, Liberman, & Friedman, 2007; Gollwitzer & Bargh, 2005; Moskowitz, 2009). Forinstance, research stimulated by auto-motive theory (Bargh, 1990) assumes thatgoals are represented mentally and that they can become automatically activated

We thank Sean McCrea or his valuable eedback on an earlier version o this manuscript. We thankthe Center or Research on Intentions and Intentionality (Universitt Konstanz) and the GermanResearch Foundation (Bonn) or unding this research. Special thanks are due to Juliane Malzacher orher invaluable help in collecting the data or Study 1.

Correspondence concerning this article should be addressed to Ute C. Bayer, FB Psychologie,Universitt Konstanz, Universittsstr. 10, D-78464 Konstanz, Germany.E-mail: [email protected].


2/19

184 bayer et aL.

by contexts in which they were oten and consistently pursued in the past. Accord-ingly, it has been observed that goal striving can be triggered by subliminally pre-sented cues, and that such goal striving may run to completion without the need

o conscious involvement (Bargh, Gollwitzer, Lee-Chai, Barndollar, & Trtschel,2001; summary by Bargh, 2006). Another example o theories on automatic goalstriving is implementation intention theory (Gollwitzer, 1993, 1999). This theorysuggests that the starting point o automatic processes is orming an i-then planby a single act o will. The execution o the i-then plan is postulated to run oautomatically without the need o conscious involvement, once the critical cuespecied in the i-component is encountered. Automaticity produced by ormingi-then plans is thereore reerred to as strategic automaticity.

impLementation intentions anD strategic aUtomaticity

Gollwitzer (1993, 1999) dierentiates between goal intentions and implementa-tion intentions. Goal intentions speciy a certain end state one wants to attain andhave the ormat I intend to attain Z! Implementation intentions speciy when,where, and how one wants to initiate goal-directed action and have the ormat Icue X occurs, then I will perorm behavior Y! They create a strong link betweena critical cue (i.e., either a situation or an inner state specied in the i-component;Achtziger, Gollwitzer, & Sheeran, 2008) and a goal-directed behavior (specied inthe then-component) by one single conscious act o will. Numerous studies haveshown that the attainment o goals strongly benets rom the ormation o imple-mentation intentions (or summaries, see Achtziger & Gollwitzer, 2008; Gollwit-zer & Sheeran, 2006). These benecial eects o implementation intentions were

explained by maintaining that implementation intentions delegate the control ogoal-directed responses to critical cues; implementation intentions turn consciousand eortul top-down control o responses by goals into bottom-up control byspecied stimuli.

The eects o implementation intentions are assumed to be based on two com-ponent processes (Gollwitzer, 1993, 1999). The rst process is concerned with thecritical cue: It is suggested that orming an implementation intention increasesthe activation o the mental representation o the critical cue thus heightening itsaccessibility. As a consequence, the critical cue is more easily detected, readily at-tended to, and successully remembered (e.g., Aarts, Dijksterhuis, & Midden, 1999;Achtziger, Bayer, & Gollwitzer, under review; Parks-Stamm, Gollwitzer, & Oettin-gen, 2007; Webb & Sheeran, 2007).

The second process is concerned with the goal-directed behavior (i.e., the then-component): It is assumed that automatic initiation o the goal-directed behavioroccurs as soon as the critical cue is encountered. Accordingly, the initiation o thegoal-directed behavior is expected to show eciency, immediacy, and the absenceo conscious intent. The eciency o action initiation as a consequence o ormingimplementation intentions was observed in research with samples characterizedby action control problems (Brandsttter, Lengelder, & Gollwitzer, 2001, Studies1 and 2; Gawrilow & Gollwitzer, 2008), dual task paradigms (Brandsttter et al.,2001, Studies 3 and 4), and tasks that burden a persons cognitive capacity (e.g., in-telligence tests; Bayer & Gollwitzer, 2007). Results o these studies suggest that ac-tion initiation produced by implementation intentions does not require much cog-


3/19

UnconscioUs action preparation anD initiation 185

nitive capacity. Moreover, Cohen, Bayer, Jaudas, and Gollwitzer (2008; Holland,Aarts, & Langendam, 2006) observed that implementation intentions eectivelyhelp to block habitual antagonistic response tendencies, thereby supporting the

assumption o immediacy (Gollwitzer & Brandsttter, 1997, Study 3). Thus, thereis sucient evidence or the assumption that action initiation by implementationintentions ullls the criteria o eciency and immediacy.

the present research: action initiationWithoUt FUrther conscioUs intent

In implementation intention research the question remains whether the automatic-ity criterion o the paucity o urther conscious intent is also met. Lengelder andGollwitzer (2001, Study 2) ound some indirect evidence or this assumption with

rontal lobe patients, known to have trouble with tasks requiring conscious con-trol o action. These patients beneted rom orming implementation intentions byovercoming the hindrance o being decient in conscious control o action. How-ever, the assumption o action initiation without urther conscious intent causedby implementation intentions still requires experimental evidence. In the presentstudies, we investigated the unconscious preparation and initiation o the goal-directed behavior in the presence o the critical cue specied in the i-componento the implementation intention. We subliminally presented visual stimuli rep-resenting the i-component o an implementation intention and assessed actionpreparation (Study 1) and action initiation (Study 2).

In Study 1, we opted or the goal to complain. We presented the critical situationas a subliminal prime (i.e., the ace o an experimenter) in a priming task that de-

manded the pronunciation o target words. These target words were instrumentalto complaining (i.e., instrumental words) or were control words (i.e., describingriendly behaviors). This procedure allowed us to investigate whether words thatcan be used to enact the goal-directed behavior o complaining were highly ac-cessible when the critical cue specied in the i-component o the implementationintention was presented subliminally. We predicted or the implementation inten-tion condition that instrumental words should show a heightened accessibility ithe ace o a rude experimenter was presented subliminally as compared to theace o a neutral experimenter; whereas or the control and the goal intention con-ditions, no such priming eects were expected.

To test whether the actual initiation o the planned behavior occurred withoutconscious intent once the critical cue was encountered, we also tested the imme-diacy o perorming the goal-directed behavior itsel, and not only the accessibilityo relevant means (e.g., instrumental words). In Study 2, we used a categorizationtask in which gures were presented as primes and targets. Participants had toclassiy the targets as round or angular by pressing a key. We investigated whetherthe classication perormance was accelerated by the subliminal presentation othe critical prime specied in the i-component o the implementation intention.For implementation intention participants, we expected that subliminally present-ed cues specied in the i-component o the implementation intention should a-cilitate the initiation o the behavior specied in the then-component. This wouldprovide evidence or action initiation without urther conscious intent. Addition-ally, we inerred that targets that required the same classication response as the


4/19

186 bayer et aL.

critical targets should also be categorized aster i the critical prime was presentedsubliminally compared to a neutral prime, as the critical prime prepares or ex-actly this classication response.

For goal intention participants, we did not predict that the critical subliminalprime would aect the initiation o the goal-directed behavior, as a goal intentiondoes not create a strong association between the critical cue and the goal-directedbehavior; bottom-up control o action by the critical prime thus should not be pos-sible.

stUDy 1: aUtomatic preparationoF interpersonaL behaVior

Goal and implementation intention participants were insulted by an experimenter

(the critical experimenter). Control participants were not insulted. The insult inthe goal and implementation intention conditions was manipulated in a mannersimilar to that used in studies on injustice and retaliation (Miller, 2001). In orderto help goal intention participants to respond to the insult, they were asked to setthe goal to complain to the critical experimenter about her behavior. In addition,implementation intention participants were asked to orm an i-then plan gearedat implementing the complaint as soon as they saw the critical experimenter.

The preparation o complaining was indicated by the accessibility o words thatcan be used to complain (instrumental words). For the implementation intentioncondition, we predicted that instrumental words should show a heightened acces-sibility i subliminally primed by the ace o the critical experimenter as comparedto the ace o a neutral experimenter. For the control and the goal intention con-

ditions, no priming-eects (critical vs. neutral ace) on the two dierent types otargets (instrumental vs. control words) were expected.

methoD

PARTICIPAnTs And dEsIGn

Sixty-eight emale students o a German university participated in the experimentin exchange or monetary compensation equivalent to approximately nine U.S.

dollars (age:M = 23.8; SD = 3.85). The study ollowed a 2 (Prime: critical vs. neu-tral ace) X 2 (Target: instrumental vs. control words) X 3 (Intention: control vs.goal vs. implementation) mixed-model design, with Prime and Target as withinparticipant variables, and Intention as a between participant variable. We assessedreading latencies (RTs) or target words as the dependent variable.

PRoCEdURE

Participants arrived at the laboratory individually. They were asked to work ontwo independent experiments (a psychophysiology and a psycholinguistics ex-


5/19


periment) run by two dierent experimenters. The psychophysiology experimentwas described as investigating the physiological eects o looking at landscapepictures (run by a ctitious Experimenter 2) whereas the psycholinguistics experi-

ment (run by Experimenter 1) pertained to latencies in pronouncing words. Beorethe experiment was started, participants looked at an album containing pictures oour dierent experimenters currently working at the institute (including Experi-menters 1 and 2).

Aterwards, the alleged psychophysiology experiment was started. First, par-ticipants had to detect an image fashed on a screen. This procedure o Liebermanand Pentland (1982) was used to determine the individual threshold o subliminalpresentation. We used EPTA 5 tachistoscopes rom ZAK to establish very shortand individually determined subliminal presentation times (i.e., shorter than itis possible by means o a computer due to the computers reresh rates whichare limited to steps o 10 ms, with 10 ms being the shortest possible presentationtime). Thereater, all participants were inormed that physiological data would berecorded while they were looking at landscape pictures. The (ctitious) Experi-menter 2 was said to be in a control room, whereas Experimenter 1 would assistby attaching wireswhich led to the control roomto the arm o the participant.Finally, Experimenter 1 let the room and the participant read instructions tellingher that the assessment o physiological responses would require her to sit com-pletely still.

Aterwards, we presented landscape pictures or 13 seconds each. Ater 19landscape pictures, the presentation was interrupted and the participant heardthe ctitious Experimenter 2 through an intercom system:Hello, can you hearme? My name is A. G. I am sitting here in the control room. Goal intention andimplementation intention participants were insulted by Experimenter 2 (A.G.): I

am trying to look at your data and I must say that it does not seem that you aresitting quietly. Until now it looks like you are not concentrating on your task, oras i you are not willing to cooperate at all. I you are only here to get some moneyand are not willing to cooperate, then you will help no one. In the end, it will onlymake you stay longer and give us bad data! Control condition participants onlyheard Experimenter 2 say: I am looking at your data and it seems that everythingis going ne. I will thereore continue collecting the data and you should continueworking on the task.

Aterwards, all participants were presented three more landscape pictures. Ex-perimenter 1 returned to remove the wires. In the control condition, she imme-diately started the psycholinguistics experiment. However, in the goal and im-plementation intention condition, Experimenter 1 said that she had heard some

noise coming rom the control room and thus was wondering what had happened.Then, she motivated participants to set the goal intention to complain about theinsult: You do not have to put up with that behavior. I I were you, I would tell hermy point o view clearly. You should take matters into your own hands and makeclear that nobody can treat participants in this way! In the goal intention con-dition, no urther comments were made. Implementation intention participantswere told: It is known that it is usually not sucient to commit onesel to a goalin order to actually attain it. Instead, goals are more oten attained i one commitsonesel to when one wants to act on them. Thereore, you should commit yourselto complain to the experimenter about her behavior as soon as you see her. All


6/19

188 bayer et aL.

participants were reminded who Experimenter 2 was by showing the photo o herto control or the amiliarity o her ace.

Then, participants started with the alleged psycholinguistics experiment, being

told to pronounce presented words as quickly as possible. Aterwards, partici-pants lled out an experimenter evaluation questionnaire and were asked abouttheir suspicions. Goal intention and implementation intention participants wereasked about their commitment to their goal intention. Finally, participants werecareully debrieed, thanked, and paid.

MATERIAl

Primes. The photo o Experimenter 2s ace (i.e., the critical experimenter) waspresented as the critical prime. As the neutral prime, a photo depicting one o theother experimenters working at the institute but not involved with the current set

o studies was used (i.e., neutral experimenter).Target Words. Eight words instrumental or enacting the goal-directed behavior

(i.e., complaining) and eight control words were used as targets. One hal o the in-strumental words (disgusting, rude, cheeky, conceited), and one hal o thecontrolwords (patient, riendly, resourceul, clever) were preceded by the critical prime.The other hal o these words was preceded by the neutral prime (instrumentalwords: impertinent, unair, dumb, nasty; control words: nice, sensitive, air, gener-ous).

To check whether these words were instrumental or neutral or complainingabout the experienced insult, we ran a pretest to rate these words (n = 24 emalestudents). Participants read a vignette describing the insult. Then, they were asked

whether the 16 words listed above were instrumental or complaining about thisbehavior on a 9-point answer scale ranging rom 1 (= not at all instrumental) to 9(= very instrumental). Computing means or the words revealed that instrumentalwords (M = 5.52, SD = 1.66) were clearly rated as more instrumental or complain-ing about an insult than control words (M = 2.92, SD = 1.77), t(23) = 4.28, p < .001.

Trials. For technical reasons (i.e., changing the slides in a tachistoscope requires acertain amount o time), long SOAs (about 1010 ms) were used. A xation cross ap-peared in the middle o a projection screen or 2000 ms, ollowed by a subliminallypresented prime. In order to guarantee the subliminal presentation o the primes,the presentation time o the primes was determined individually or each par-ticipant (M = 10 ms, SD = 3.26) at the outset o the experiment. This presentationtime was then used as the individual presentation time o the primes or this veryparticipant. The prime was ollowed immediately by a pattern mask presented or100 ms. Ater the presentation o the mask, a blank screen appeared or 900 ms,ollowed by a word or 500 ms. When participants had responded by reading theword aloud, the next trial was started. Subliminal primes and words were pairedin a xed randomized order.

Manipulation Checks and Evaluation of Experimenters Behavior. Participants wereasked or each o the two experimenters: Are you dissatised with how you weretreated by the experimenter?; Did you think the experimenter had reasonabledemands?; and How did the experimenter treat you as a participant? Thesethree questions were answered on 10-point scales (0 = not at all dissatised, unrea-


7/19


sonable, and unpleasant; 9 = very dissatised, unreasonable, and unpleasant). Thenext two questions assessed the commitment to the goal to complain in goal in-tention and implementation intention participants: Do you intend to complain?

(yes/no/undecided), and How strongly do you eel committed to complaining?(9-point answer scale reaching rom not at all to strongly).

resULts

sUsPICIons And oUTlIERs

When participants were asked about suspicions concerning the study, three othem reported that they did not believe that the critical experimenter actually ex-

isted. Two participants reported that they had been able to see aces prior to thepresentation o the words; and one participant claimed that she had seen acesand that one o these aces depicted the critical experimenter. RTs or these veparticipants were excluded rom data analysis. These ve participants were aboutequally distributed across intention conditions. RTs were scrutinized or extremeoutliers such that RTs more than 3 standard deviations above and below the meano each word were excluded. Thus 1.1% o the RTs were excluded rom data analy-ses.

PRIMInG EFFECTs

A 2 (Prime: critical vs. neutral ace) 2 (Target: instrumental vs. control words)3 (Intention: control vs. goal intention vs. implementation intention) mixed modelANOVA was conducted, with Prime and Target as within participant variables,and Intention as a between participant variable. This revealed a signicant Inten-tion Prime Target interaction eect on RTs, F(2, 60) = 6.22, p < .01. The maineect o Prime was also signicant, indicating that RTs or targets presented aterthe critical prime were longer as compared to the neutral prime, F(1, 60) = 3.95, p .15; see Figure 1).

To test the hypotheses concerning our intention manipulations, or each o the

three intention conditions 2 (Prime: critical vs. neutral ace)

2 (Target: instru-mental vs. control words) ANOVAs on RTs were computed separately. In the con-trol condition, a marginally signicant main eect o Target was discovered, F(1,18) = 3.47, p = .08. RTs or control targets were shorter than or instrumental targets(instrumental:M = 502, SD = 84.85;control:M = 492, SD = 80.59). The actor Primedid not reach signicance (F < 1), nor did the Prime Target interaction eect, F(1,18) = 1.29, p = .27 (see Figure 1, let side).

In the goal intention condition, the Prime Target ANOVA revealed no signi-cant main eects o Target, F(1, 23) = 2.48, p = .13, and o Prime, F(1, 23) = 1.50, p= .23. Also the Prime Target interaction eect was not signicant (F < 1). As ex-pected, having a goal intention to complain did not acilitate access to instrumen-


8/19

190 bayer et aL.

FIGURE1

.Mean

RTo

fthe

Thre

eInten

tion

Con

dition

by

Primean

dTarget

(study

1).


9/19


tal target words compared to control words when the critical ace was presentedas a prime (see Figure 1, middle part).

In the implementation intention condition, the Prime Target ANOVA revealed

a signicant interaction eect, F(1, 19) = 13.99, p < .001. The main eect o Targetdid not reach signicance (F < 1), nor did the main eect o Prime, F(1, 19) = 1.73,p = .20. Furthermore, planned t-tests revealed that instrumental words sublimi-nally primed by the critical prime (M = 513, SD = 82.11) were responded to asterthan instrumental words primed by the neutral prime (M = 528, SD = 81.78), t(19)= 1.92, p = .03 (one-tailed). In addition, control words subliminally primed by thecritical prime (M = 532, SD = 70.16) were responded to signicantly more slowlythan when subliminally primed by the neutral prime (M = 501, SD = 80.71), t(19)= 3.29, p < .01 (see Figure 1, right side). Finally, ater the presentation o the criticalprime, shorter RTs were observed or instrumental words (M = 513, SD = 82.11) ascompared to control words (M = 532, SD = 70.16), t(19) = 2.39, p < .05. The reversewas true ater the presentation o the neutral prime (instrumental words:M = 528,SD = 81.78; control words:M = 501, SD = 80.71), t(19) = 2.43, p < .05.

FURTHER AnAlYsEs

First, we checked or participants goal commitment. When asked whether par-ticipants intended to complain, no dierences emerged between goal and imple-mentation intention participants, (2, N= 42) = 3.36, p = .15. Thus, no signicantincrease in goal commitment as a consequence o orming an implementation in-tention was observed. The question concerning participants commitment to com-plaining also did not reveal a signicant dierence between goal (M = 4.24, SD =

2.96) and implementation intention participants (M = 4.70, SD = 2.92), F < 1.Second, we analyzed the evaluation o the experimenters behavior. The three

items evaluating the two experimenters behavior were highly correlated (Cron-bachs = .87), and thus an evaluation index was created by computing the meano the items. An Experimenter Intention ANOVA with repeated measures on therst actor revealed a signicant main eect o Experimenter, F(1, 58) = 155.16,p .30 (see Figure 3).We analyzed the Intention Prime Target interaction urther: In the goal

intention condition, the Prime Target interaction did not reach signicance, F(2,42) = 1.70, p = .19, nor did any o the main eects (all Fs < 1, see Figure 3, let side).As expected, the subliminal primes did not have any eects on goal intention par-ticipants categorization responses.

In the implementation intention condition a signicant Prime Target interac-tion was observed, F(2, 38) = 3.14, p < .05. This interaction qualied a signicantmain eect o Prime, F(1, 19) = 6.11, p < .05, and o Target, F(2, 38) = 7.13, p < .01 (seeFigure 3, right side). Supporting our hypotheses, participants were signicantlyaster in categorizing critical targets preceded by the critical prime (M = 416, SD =94.42) than in categorizing critical targets preceded by the neutral prime (M = 426,SD = 86.83), t(19) = 2.18, p < .05. Moreover, we observed that critical targets (M =416, SD = 94.42) were classied aster than congruent targets (M = 434, SD = 71.59),t(19) = 2.23, p < .05, and incongruent targets (M = 456, SD = 78.71), t(19) = 3.19, p

Date post:	08-Aug-2018
Category:	Documents
Upload:	nguyen-hong-ha
View:	216 times
Download:	0 times

09 Bayer Achtziger Etal Subliminal Implementation Intent

Documents