JACOBSON.MORTENSEN.AND CIALDINI le whether a string of letters re ented a meaningful of word prime appro hed significance in this key on t 166 60,p1 m2= nd tare vords chosen at tandom without re lacen ded norms word prime that ms.Prev s research (e.g.. 997 uracy/efficiency goa d les the ately fo owing the masking stimulu n.ared unti tion effect for both ti ccuracv efed foll ving completion of the ta Results e in co nparison to nding on the he two classes of goal words. The mean for Discussion sequent analyses (76%of resp The results of Experiment 1 support our that indi es).Trials in onwords were not included in Nomainorinteraction of particp the a began by submitting the data to a 3(word prime:descriptive norm. predicted that the descriptive norm primes would increase re and Standard Deviations for Prime-Targe Trial Categories i Experiment2 Neutral prime Target word (goals) SD SD SD y 728594 6m3;1809%68947769 Procedure. Participants received instructions to complete a lexical decision task. Their objective was to indicate, as quickly as possible, whether a string of letters represented a meaningful English word (by pressing the 5 key on their keyboard) or repre￾sented a nonword (by pressing the A key). All participants com￾pleted 72 trials, made up of 36 words (18 words related to the goal of accuracy/efficiency and 18 words related to the goal of social approval) and 36 filler nonwords. Within each of the three target word categories, individual trials were paired in equal numbers with the three types of word primes. For each trial, word primes and target words/nonwords were chosen at random without re￾placement from within each of the categories. Thus, nine distinct types of prime–target pairs were presented to each participant. All task stimuli were presented and response times recorded using DirectRT v2006 software (Jarvis, 2006a). Each trial started with a fixation point (“
”) presented on the center of the computer screen for 1,000 ms. This was immediately followed by a masking stimulus (“X”) that appeared for 70 ms, a word prime that ap￾peared for 35 ms, and a masking stimulus that again appeared for 70 ms. Previous research (e.g., Wittenbrink, Judd, & Park, 1997) has indicated that prime words are detectable but not identifiable at similar stimulus intervals. Target letter sequences were presented immediately following the masking stimulus and appeared until participants made a word/nonword decision. A 1,500-ms pause separated each decision from the beginning of the next trial. All participants were fully debriefed following completion of the task. Results Initial considerations. We excluded three participants with accuracy below 60% in identifying words/nonwords (three stan￾dard deviations below the mean) due to random responding on the task. The mean percentage of correct responses for the 84 remain￾ing participants was 95%. We excluded all incorrect lexical deci￾sions from subsequent analyses (7.6% of responses). Additionally, based on considerations for response time data discussed by Bargh and Chartrand (2000), we excluded decisions with response times less than (i.e., faster than) 300 ms and those greater than (i.e., slower than) three standard deviations above the mean (2.3% of responses). Trials involving nonwords were not included in sub￾sequent analyses. In this and all subsequent experiments, we examined the data for possible gender effects before testing our primary hypotheses. No main or interaction effects of participant gender were found in Experiment 1 or in subsequent experiments. Thus, this variable was not included in tests of our primary hypotheses. Hypothesis tests. Mean response times and standard devia￾tions for each prime-target category are presented in Table 1. We began by submitting the data to a 3 (word prime: descriptive norm, injunctive norm, neutral) 2 (goal target: accuracy/efficiency, social approval) within-subjects analysis of variance (ANOVA). The main effect of word prime approached significance in this analysis, F(2, 166) 2.60, p .078, p 2 .030, and the main effect of goal target was significant, F(2, 166) 75.82, p  .001, p 2 .477. However, as predicted, these effects were qualified by a significant interaction between word prime and goal target, F(2, 166) 17.20, p  .001, p 2 .172. We then proceeded to a set of planned comparisons designed to compare response times for goal targets in the neutral prime condition against response times for those targets in the descriptive and injunctive norm conditions (i.e., the comparisons were focused on the facilitative/suppressive effects of the normative primes on response times). On the basis of theory, priming with words related to descriptive norms should facilitate response times for words related to an accuracy/efficiency goal but should not facilitate response times for words related to a social approval goal. Sup￾porting these predictions, when preceded by descriptive norm primes, participants responded more quickly (in comparison to neutral primes) to words related to the accuracy/efficiency goal, F(1, 83) 12.77, p .001, p 2 .133. In contrast, participants actually responded less quickly to words related to the social approval goal when these were preceded by descriptive norm primes, F(1, 83) 6.78, p .011, p 2 .076. For injunctive norms, we hypothesized a facilitation effect for both the accuracy/ efficiency and social approval classes of goal words. Supporting this prediction, participants responded more quickly to goal targets when these had been preceded by an injunctive in comparison to a neutral prime, F(1, 83) 8.28, p .005, p 2 .091. Figure 1 depicts the facilitative/suppressive effects of the norm primes for the two classes of goal words. Discussion The results of Experiment 1 support our contention that indi￾viduals associate descriptive information primarily with the goal of accuracy/efficiency but associate injunctive information with the dual goals of accuracy/efficiency and social approval. Following subliminal presentation of norm-related words, word/nonword re￾sponses for the goal of accuracy/efficiency were faster than control in both the descriptive and injunctive norm conditions. For the goal of social approval, responses were slower than control in the descriptive norm condition and faster than control in the injunctive norm condition. Thus, thoughts about making accurate and effi￾cient decisions appear particularly likely following the salience of a descriptive norm, whereas thoughts about both accuracy/ efficiency and social approval are likely to follow from the sa￾lience of an injunctive norm. Although we had not explicitly predicted that the descriptive norm primes would increase re￾Table 1 Means and Standard Deviations for Prime-Target Trial Categories in Experiment 2 Target word (goals) Descriptive norm prime Injunctive norm prime Neutral prime M SD M SD M SD Accuracy/efficiency 607.28 159.14 663.53 141.02 696.47 216.56 Social approval 721.97 224.84 618.35 100.76 659.94 170.12 438 JACOBSON, MORTENSEN, AND CIALDINI This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.