Pilot Study

In the pilot study, our goal was to test the stimuli intended for the main study, specifically to assess participants’ reactions to the research materials and determine whether the tasks were manageable. A sample of 20 adults (10 female) volunteered to participate in the online study. Participants were recruited through advertisements on local websites, including Facebook and Instagram, and they took part voluntarily without monetary compensation. Prior to participating, they signed an online consent form (in the pre-screening survey) in accordance with local ethical requirements.

Every participant received a link to the meeting after completing the form. We sampled a pilot group similar to our main group—adults with a balanced gender representation, excluding past or current students of psychology or related fields such as pedagogy, philosophy, sociology, political science, or resocialization. The pilot study was conducted on the Zoom platform as an individual meeting with the researcher, with cameras turned on. Upon joining the Zoom meeting, each participant used an individual code rather than their name, following pre-screening survey instructions. The meetings were recorded, with materials accessible only to the researchers for data analysis. Participants viewed figures on the screen (Fig. 1) and were asked to answer the question, “Which of the figures on the right is the same as the figure on the left?” by responding with the number 1, 2, or 3. Ninety percent of the participants (n = 18) answered all questions correctly, while the remaining 10% (n = 2) made a single error in one of twelve attempts. For the main study, we included only the sets for which all participants answered correctly.

Fig. 1Main Study Participants 

Our pre-registered target sample size was N = 55, which provided 80% power to detect an effect size of 0.15 for multiple linear regression. As preregistered, data collection was scheduled to end once we achieved N = 55, but if additional participants signed up on the same day, they would be included. All future appointments were canceled once the target was met. The final sample included N = 74 adults (35 females, 39 males) with a mean age of 27.00 years (SD = 11.72), who volunteered without financial compensation. Participants included both believers and atheists (n = 33 Roman Catholics, n = 1 Orthodox, n = 2 Protestants, n = 1 Old Catholic, n = 30 atheists, and n = 7 who selected “other”). Their education levels varied (n = 1 with primary education, n = 1 with basic vocational education, n = 3 with technical education without a matura exam, which is the examination upon leaving secondary school, n = 3 with high school education without a matura exam, n = 46 with secondary education with a matura exam, n = 9 with a first degree in higher education, n = 10 with a master’s degree, and n = 1 with a PhD). The sample also varied in occupational status (n = 38 students, n = 21 employed under an employment contract, n = 6 with a mandate contract, n = 3 business owners, n = 3 unemployed, n = 2 retired, and n = 1 pensioner). They lived in various types of locations (n = 5 in a village, n = 5 in a small town with up to 20,000 inhabitants, n = 12 in a medium-sized city with 20,000 to 99,000 inhabitants, n = 33 in a large city with 100,000 to 500,000 inhabitants, and n = 19 in a big city with more than 500,000 inhabitants). Their places of origin were similarly diverse (n = 10 from villages, n = 10 from small towns with up to 20,000 inhabitants, n = 16 from medium-sized cities with 20,000 to 99,000 inhabitants, n = 31 from large cities with 100,000 to 500,000 inhabitants, and n = 7 from big cities with over 500,000 inhabitants).

As preregistered, we excluded participants based on several criteria: one participant who was driving a vehicle during the study (n = 1; see Excluded_based_on_comments variable), one who reported difficulty recognizing geometric figures, later confirmed by a medical doctor (n = 1; see Excluded_based_on_comments variable), seven who gave incorrect answers on filler boards (questions to which all confederates gave the correct answers; n = 7; see Fillers variable; only data from participants with a score of 0 were analyzed), and ten who were familiar with Asch’s experiment (n = 10; see Whether_the_subject_knew_Asch variable). In total, we excluded 17 participants, with some failing multiple exclusion criteria.

Our final analyzed sample included 57 participants (30 males, 27 females), with a mean age of M = 27.16 (SD = 11.01). Of these, 27 identified as Roman Catholics, 1 as Orthodox, 2 as Protestants, 20 as atheists, and 7 selected “other.” Education levels varied, with 1 participant having primary education, 1 with basic vocational education, 2 with technical education (without a high school diploma equivalent), 2 with high school education (without a diploma), 35 with secondary education with a high school diploma, 6 with a first degree in higher education, and 10 with a master’s degree (second degree in higher education). Occupational status was mixed: 30 were students, 16 employed under a standard contract, 3 on a mandate contract, 3 business owners, 3 unemployed, and 2 retired. Participants lived in diverse locations: 4 in villages, 2 in small towns (up to 20,000 inhabitants), 9 in medium-sized cities (20,000–99,000 inhabitants), 28 in large cities (100,000–500,000 inhabitants), and 14 in big cities (over 500,000 inhabitants). Their places of origin also varied: 7 were from villages, 7 from small towns, 12 from medium-sized cities, 25 from large cities, and 6 from big cities. Sensitivity power analysis for our sample size, using multiple linear regression (one-tailed, α = 0.05), indicated 80% power to detect an effect size of 0.11.

Procedure and Materials

Recruitment for the main study mirrored the pilot study process, with participants able to select their preferred day and time in the pre-screening survey. Participants were instructed to join the Zoom meeting using an anonymous code and to participate from a quiet location free from distractions. Each participant met with the researcher and four confederates (two men and two women of similar age) posing as fellow participants. The researcher conducted a fake randomization, ensuring the real participant always answered last. Participants viewed 12 figures on the computer screen (one per page) and responded to the question, “Which of the figures on the right is the same as the figure on the left?” with options 1, 2, or 3. Following Asch’s conformity paradigm, confederates were instructed to give incorrect answers (Fig. 2), but the study included four filler trials in which confederates answered correctly, creating a sense of realism for the true participants (Paruzel-Czachura et al., 2024). Full details on the procedure, including scripted prompts, are available in the preregistration and materials archived on the Open Science Framework.

Fig. 2

Example of Stimuli and the Actors’ Response

Following the Zoom meeting, participants filled out an online survey measuring their religiosity and religious fundamentalism with the Religiosity Centrality Scale (Huber, 2003) and the Religious Fundamentalism Scale (Altemeyer & Hunsberger, 2004). Additionally, they provided demographic information, including sex, age, education level, employment status, place of residence, place of origin, and religiosity. Participants used the same anonymous code as in the Zoom task to ensure their survey responses could be accurately matched to their earlier participation. The complete set of survey questions is available on the Open Science Framework.

Measures Religiosity

The Centrality of Religiosity Scale (Huber & Huber, 2012) was used to measure religiosity. This 15-item scale is divided into five subscales: intellect, ideology, private practice, religious experience, and public practice. Participants responded to seven items about the importance of religious issues using a 5-point scale from 1 (not at all) to 5 (very much so), with items including “How interested are you in learning more about religious topics?” and “How important is personal prayer for you?” Six additional items measured the frequency of religious engagement on a scale from 1 (never) to 5 (very often), such as “How often do you think about religious issues?” and “How often do you pray spontaneously in response to daily situations?” The final two items assessed the frequency of prayer and religious service attendance, with participants selecting from predefined frequency options instead of using a numerical scale. Total scores were calculated by summing the responses, with higher scores indicating a greater centrality of religiosity.

Religious Fundamentalism

The Religious Fundamentalism Scale (Altemeyer & Hunsberger, 2004) was used to measure fundamentalism. This scale has 20 items rated on a 9-point scale ranging from 1 (very strongly disagree) to 9 (very strongly agree). Participants responded to statements such as “It is more important to be a good person than to believe in God and the right religion” and “God’s true followers must remember that He requires them to constantly fight Satan and Satan’s allies on this earth.” Total scores were calculated by summing all item responses, with higher scores indicating a greater level of religious fundamentalism.

Conformity

Conformity was measured using 12 boards displaying geometric figures: one figure on the left and three on the right. Participants were asked 12 times to identify which of the figures on the right (labeled 1, 2, or 3) matched the figure on the left. Four boards served as filler items to reinforce the perception that all individuals, including actors, were genuine participants. On these filler boards, the actors provided the correct answers. For the remaining eight boards—the main items used to assess conformity— the actors intentionally provided incorrect answers. The responses of the four actors preceded the participant’s answer. If participants provided the same incorrect answers as the actors on these main items, it indicated group influence. Higher conformity scores were attributed to participants who matched the actors’ incorrect answers more frequently, demonstrating a greater tendency toward conformity.