Research Instruments & Databases

Participants are asked to describe themselves in twenty different ways. We examine whether participants’ descriptions reference generalized/abstract traits (e.g., I am kind) or specific, contextualized attributes (e.g., I am kind with children; I am a member of a tennis club; Cousins, 1989; Rhee, Uleman, Lee, & Roman, 1995). We count the number of times participants indicate specific/contextualized terms as compared to generalized/abstract traits. An index is obtained by subtracting the number of abstract statements about the self (e.g., general dispositions) from 20, which is the total number of descriptions.

In the Sociogram task, participants are asked to draw their social network by using circles to represent the self and others. The size (i.e., diameter) of the self-circle relative to the other-circles is calculated as an index of self-inflation. The score is the size of self-circle divided by the average size of other circles while controlling for the overall area of the drawing defined by the outer horizontal and vertical margins.

This scale has two subscales that measure independent and interdependent self-construals.

Note: For cross-cultural purposes, it is important to note that the only previous established cultural differences is in the inclusion of the family in the self (Li, 2002).

The IOS scale is a pictorial measure of closeness. In the IOS scale, a series of two circles is provided where the degree of overlap between them progresses linearly, creating a seven-point scale of relational closeness. Participants selected one pair of circles that best represents their relationships with family members.

In the ISOQ, participants remember ten episodes (e.g. “when I had a positive interaction with friends”; “when I was reading a novel”) and indicate how much they subsequently experience each of 7 positive (e.g. feeling of closeness, elated, happy) and 5 negative (e.g. ashamed, angry, unhappy) emotions on a scale from 1 = ”not at all” to 6 = ”Very strongly.” Among those emotions, three different types are embedded: general emotions – both positive (e.g., happiness) and negative (e.g., unhappiness), socially engaged emotions – both positive (e.g., friendly feeling) and negative (e.g., guilt), and socially disengaged emotions – both positive (e.g., pride) and negative (e.g., anger).

To yield intensity of engaged vs. disengaged emotions, we first compute the average intensity of general positive emotions (happiness and joy) for each of the 10 episodes. Depending on whether the average intensity of the general positive emotions is greater or less than the intensity reported for unhappiness, the episode is classified as either positive or negative. For the positive episodes, the average intensity for the positive disengaged emotions is subtracted from the average intensity for the positive engaged emotions to indicate the relative intensity of engaged (vs. disengaged) emotions. For the negative episodes, the same procedure is repeated using negative emotions. The relative intensity scores are averaged across all the 10 episodes. 

In this measure (which was also derived from the ISOQ), we examine the relative effects of socially engaged vs. disengaged positive emotions on happiness. For each individual we perform a linear regression with socially engaged vs. socially disengaged emotions predicting general positive emotions. The unstandardized coefficients are used as scores of general tendency to predict happiness as a function of engaged vs. disengaged emotions.  

This task measures whether people draw a sharp distinction between how they treat friends versus strangers. Participants are asked to imagine going into a business deal with (i) an honest friend, (ii) a dishonest friend, (iii) an honest stranger, and (iv) a dishonest stranger. In the stories, the friend or stranger's lies cause the participant to lose money in a business deal, and the honesty causes the participant to make more money. In each case, the participants have a chance to use their own money to reward or punish the other person.

Note from T. Talhelm: Some evidence for the order effect has been found. If participants read the dishonest scenario first, they're less likely to reward on the subsequent honest scenario. However, if they read the honest scenario first, it does not affect their responses to the dishonest scenario. For that reason, it's recommended to begin with the dishonest scenario first, if running a within-subjects design.

The FLT examines the extent to which participants ignore vs. take into account contextual information in perceptual judgment (Kitayama, Duffy, Kawamura, & Larson, 2003). Participants see a set of 12 squares with a line drawn inside each and are asked to reproduce the line inside a new square of a different size either by duplicating its absolute length (ignoring the context of the square) or its length relative to the square (by drawing a line with the same proportion as in the original square). Squares vary in size. The score is the magnitude of error (adjusted for the size of the square) for the absolute judgments  minus the magnitude of error for the relative judgments. 

The Change Blindness task examines how easily participants detect changes in focal vs. background objects in order to measure whether they pay attention to focal vs. background objects (Masuda & Nisbett, 2006). Specifically, participants watch several pairs of animated scenes such as a construction site and an airport. Each scene pair consists of two similar but slightly different vignettes and participants are asked to detect the difference between them. The number of changes noticed in focal objects and contexts are counted. The score is the frequency of contextual changes noticed minus the frequency of focal object changes noticed.

This task tests holistic attention (Masuda & Nisbett, 2001). Participants see animated vignettes of fish and are asked to recall what they saw after seeing each vignette. We count the number of statements about focal objects such as fish in the foreground and the number of statements about background. The score for the task is the frequency that context is mentioned minus the frequency that focal fish are mentioned. 

For more information see Segmentation Rules (PDF)

We investigate the amount of information participants consider before making a judgment (Choi, Dalal, Kim-Prieto, & Park, 2003). Participants imagine that they are a detective investigating a murder case. They are provided with 97 clues that might or might not be relevant to the case (e.g., the number of pets the victim owned and the victim’s history of sexual abuse by his/her parents) and ask them to exclude clues which they think are causally irrelevant. The score is the number of items they think are causally relevant to the event.

Note: As argued by Cohen, Hoshino-Browne, and Leung (2007),  differences between independent and interdependent cultures are mainly pronounced when focusing on social situations.

In this task, we measure the degree to which participants take a third-person (holistic) vs. first-person (analytic) perspective when they think about their past (Cohen & Gunz, 2002). Participants are asked to recall specific instances of social (e.g., being embarrassed) and non-social or ambiguous (e.g., watching a horror movie) situations and indicate whether the memory is a first-person memory (in which only the context is seen) or a third-person memory (in which the person in relation to the context is the focus) on an 11-point scale (1= “entirely a first-person memory” to 11= "entirely a third-person memory”). The resulting ratings are collapsed into a single index. This variable often shows a skewed distribution, hence a log-transformation is recommended for subsequent analyses.

We test the extent to which participants attribute causality to an actor vs. the context (Kitayama, Ishii, Imada, Takemura, & Ramaswamy, 2006). Specifically, in the updated version, participants read four vignettes describing either positive or negative behavior of a target (2 positive and 2 negative vignettes). For each vignette, they indicate their agreement with two items reflecting dispositional attribution and two items reflecting situational attribution (1 = ”Strongly Disagree” to 7 = ”Strongly Agree”). The score is their ratings for situational attributions minus those for dispositional attributions. 

This task uses materials from Ji, Nisbett, & Su (Study 2; 2001), but is different from the original measure. In the adapted measure, participants are presented with eight graphs, each showing a trend (e.g., economic growth), and indicate the next two points on each graph. We measure the vertical distance (number of cells on the grid) between the baseline in "2004" and the prediction in "2008." These scores are multiplied by "-1" and averaged to form a single index of nonlinear reasoning.

The task examines whether participants categorize objects based on a thematic context and functional relations vs. abstract taxonomies (Chiu, 1972). In the task, there are 14 items and, for each item, participants are given a target object (e.g., picture of a cow) and asked to choose which of two alternatives (e.g. chicken vs. grass) is best associated with the target object. One alternative is thematically related to the target (e.g., picture of grass) and the other belonged to the same taxonomic category as the target (e.g., picture of a chicken). Responses on 14 items (1 = thematic vs. 0 = taxonomic) are summed into a single score representing the number of thematic vs. taxonomic categorizations. The variable often shows a skewed distribution, hence a log-transformation is recommended for subsequent analyses.