(1) Overview


This project [1] was based on the idea that interpersonal touch (e.g., a soft touch on the shoulders from an experimenter) or simulated interpersonal touch (e.g., touching a soft teddy bear) could alleviate existential concerns among people with low self-esteem. The research was guided by Terror Management Theory, which proposes that human’s unique awareness of their mortality creates the potential for overwhelming anxiety, or terror. Cultural beliefs and self-esteem together form our core protection against such existential concerns [2]. From a TMT perspective people with low self-esteem lack a crucial buffer against death anxiety. Indeed prior research has shown that these individuals find it difficult to suppress death thoughts, and respond to death reminders with increased anxiety [3, 4]. We conducted four studies to show that the existential concerns among people with low self-esteem may be alleviated by seemingly trivial experiences of (simulated) interpersonal touch.

Study 1 showed that a touch on the shoulder by a female experimenter could reduce death anxiety among people with low self-esteem. Study 2 showed that the same touch on the shoulder made people with low self-esteem feel more socially connected when reminded of their mortality. In Study 3, people with low self-esteem who were reminded of their mortality showed a greater desire for touch, as indicated by higher value estimates of a soft teddy bear (Study 3). Finally, in Study 4, people with low self-esteem were less ethnocentric when reminded of their mortality when they were allowed to hold a teddy bear on their lap.

(2) Methods


Students of the VU University were approached on campus and asked to fill out paper-and-pencil questionnaires in Studies 1 and 2. Note that the participants from Study 1 and Study 2 partly overlap. The participants of Study 1 form the high mortality salience condition in Study 2 (this is indicated by the “study” variable in the Study 2 dataset). In Studies 3 and 4, students participated in our research in the social psychology labs at the VU University Amsterdam for payment or course credit.

Several demographic variables are included in the datasets of each of the studies. For approximately half of the participants in Studies 1 and 2, we only gathered information on age and sex (their values on the remaining demographic variables are indicated as “missing”). We started measuring the additional demographic questions (e.g., the birth country of the father/mother, relationship status) halfway through running these studies. The questions were not related to our hypotheses in these studies and were added at the very end of the survey.

Study 1: Sixty-one paid volunteers (26 women; average age 23) were randomly allocated to touch versus no-touch conditions.

Study 2: One hundred and twenty paid volunteers (53 women; average age 23) were randomly allocated to high versus low mortality salience and touch versus no-touch conditions.

Study 3: Fifty paid volunteers (30 women, 20 men; average age 21) were randomly allocated to high versus low mortality salience conditions.

Study 4: Eighty paid volunteers (52 women; average age 20) were randomly allocated to high versus low mortality salience and touch versus no-touch conditions.


In all four studies, self-esteem was measured with the Rosenberg [5] self-esteem scale (RSES). The RSES consisted of 10 items (e.g., “On the whole, I am satisfied with myself”) that were answered on 5 (Studies 1–2) or 9 point scales (Studies 3–4).

In Study 1, participants were either touched by our female experimenter (a woman in her 20s) or not [6]. The main outcome measure was death anxiety, which was measured using 7 items (e.g., “I am afraid of death, because it is so final”) using 5-point scales (1 = not at all; 5 = very much).

Study 2 manipulated both touch (see Study 1) and mortality salience (MS). The MS manipulation consisted of asking participants to fill out either a questionnaire about fear of death (see Study 1) or a parallel questionnaire about fear of dentists. The main outcome measure was social connectedness, which we measured by asking participants to list the names of 7 to 10 persons they knew and rate how much they felt connected with them (1 = not all; 9 = very much).

Study 3 manipulated MS as in Study 2. As the outcome variable, participants in Study 3 were asked to estimate the retail value in Euros of a teddy bear in a cardboard box. We interpreted the estimated value as an indication for desire for touch, and this was our main outcome measure. We measured mood with the brief Profile of Mood States [7] on 9 point scales (1 = not all; 9 = very much).

Study 4 manipulated MS as in Studies 2 and 3. In addition, we manipulated touch by letting participants either view a teddy bear in a cardboard box or by asking them to hold the teddy bear on their lap [8]. The outcome variable of interest was ethnocentrism, which we measured by asking participants to rate how often a typical Dutch person and a typical Muslim person (1 = not all; 9 = very much) would experience 6 negative emotions (e.g., pain, contempt) and 6 positive emotions (e.g., pleasure, hope). The ratings were first summed separately by valence for the Dutch and Muslim targets. Ethnocentrism was indicated when participants attributed a) more positive emotions and less negative emotions to Dutch persons, and/or b) less positive emotions and more negative emotions to Muslims. We measured mood with the 23-item version of the Multidimensional Mood Questionnaire [9] on 4 point Likert scales.

The test materials and instructions are described in more detail in the paper and are available together with the datasets.

In our studies, we also measured death thought accessibility (DTA) by means of word fragment completions, a standard paradigm for assessing DTA (e.g., Routledge et al., 2010). Unfortunately, we found the internal reliabilities of the measure to be very low. Information on this measure is available online in the Supplemental Material section of our published article.


In Studies 1 and 2, participants were approached on the campus site of the VU University by a female experimenter in her 20s. Participants responded by filling out questionnaires using paper and pencil.

Studies 3 and 4, were conducted at the psychology lab of the VU University Amsterdam, which was located at the basement of the psychology building. Because experiments are run there throughout the year, participants (most students) sign themselves up on their own initiative, either in return for monetary payment or course credits. The experiments were run in cubicles and the data were automatically registered via personal computers.

Quality Control

We a priori excluded non-native Dutch participants because prior research has shown that intergroup touch has a special symbolic meaning over and above interpersonal touch [10]. Twenty-one non-native Dutch participants were discarded in Study 1, of which 10 participants also participated in Study 2. In Study 2, 4 additional participants were discarded due to missing values. Finally, 2 non-Western participants were excluded from the published dataset in Study 4.

Ethical issues

The study followed the ethical standards by the American Psychological Association. Data were anonymized by using participant numbers. We did not obtain personal identifiers, such as names, or e-mail addresses. Participants were free to discontinue participation at any time without having to state a reason.

(3) Dataset description

Object name

Data from Paper ‘Embodied terror management: interpersonal touch alleviates existential concerns among individuals with low self-esteem’

Data type

Primary data (items are already reverse scored)

Format names and versions

The data are available as a .sav file (SPSS) or a .csv file. Syntax files (.sps) for the main analyses are also available. The .sav files and .sps files were created with SPSS 21 for Windows.

ETM Data – Study 1.sav/.csv

ETM Data – Study 2.sav/.csv

ETM Data – Study 3.sav/.csv

ETM Data – Study 4.sav/.csv

Data Collectors

The data of Studies 1 and 2 were collected during Fall 2011 and Spring 2012 by Tanya Davidesko as part of her Master’s thesis research. She digitized the paper-and-pencil questionnaires. Mandy Tjew A Sin collected the data of Studies 3 and 4 during Fall 2011 and Spring 2012 as part of her Master’s thesis research. Sander Koole and Mandy Tjew A Sin were responsible for data collection and storage. Sander Koole conducted the data cleaning, after which Mandy Tjew A Sin prepared the data files for publishing.







Publication date


(4) Reuse potential

The data can be used for aggregation, further analysis, reference, validation studies, and may be particularly interesting for researchers interested in mortality salience and touch effects among Dutch student samples.