Implicit Assocation Experiment

Social Balance Experiment

By Michael Birnbaum (mbirnbaum@fullerton.edu)
CaliforniaState University at Fullerton
and Ken McGraw (pymcgraw@olemiss.edu)
University of Mississippi

Back to Instructor's Page

Introduction
Method
Analysis
References

Introduction

Social balance theory as proposed by Fritz Heider (1946) is the theory that people strive for cognitive balance in their network of likes and dislikes. This theory is most easily understood in triadic relationships. Consider, for example, that you dislike a person who is loved by your best friend. This is a situation that lacks balance because people you dislike should be disliked as well by those whom you love. The presence of this imbalance will motivate a change in your attitudes. To achieve greater balance, you might decide that your best friend is not really a best friend after all. Or you might decide that person you initially disliked is a really a pretty good egg after all. Making these adjustments restores balance.

The current experiment demonstrates that people have an expectancy for balanced social relations. Moreover, it can be used to test a very strong version of balance theory for triadic relationships. The strong theory is that when you and another person (Bill) have an attitude toward a 3rd person John, the degree to which you like John is the product of how much you like Bill and how much Bill likes John. This is a strong theory because it makes a precise quantitative prediction of liking. It says that if you like Bill by a value of +2 and Bill dislikes John by a value of -1, then you will dislike John by a value of -2 (because +2 X -1=-2). This experiment can be used to test this strong theory.

At a more general level, the experiment provides an interesting class exercise that demonstrates the concept of "disordinal" statistical interaction, an interaction where the direction (and not just the magnitude) of the effect of a variable depends on the values of a second variable.

Method

The dependent variable in this experiment is how much you judge you will like a person named John. Choices can range from 1 (dislike very very much) to 9 (like very very much). To make the determination, you are given information about how much you like a person named Bill and how much Bill likes John. These two pieces of information constitute separate independent variables in the experiment. Values for the variable "How Much You Like Bill" are hate, dislike, neutral, like, and love. Values for the variable "How Much Bill Likes John" are hates, dislikes, neutral, likes, and loves.

« Return to Top»

Design

The experiment employs a 5 x 5 within subjects factorial design. The first variable is How Much You Like Bill and the second is How Much Bill Likes John.

Analysis

Complete description of how to use the data from the experiment to test the strong version of balance theory are given in Birnbaum (2001, Chpt. 13). The test of the theory requires comparing predicted values to observed values for each of the 25 means obtained in the 5 x5 factorial design. An alternative analysis is to use an analysis of variance to determine the significance of main effects and interactions. The analysis should show that the interaction of the two "liking" variables (i.e., How Much You Like Bill and How Much Bill Likes John) accounts for nearly all of the non-error variance.

Data Format for Downloadable csv file

The data obtained via the Download data link on the PsychExperiments homepage are comma delimited. The first entry (up to the first comma) is the affiliation of the research participant. Typically this is the class to which the participant belongs. Non-affiliated participants use the default entry of Interested Person. Next is the randomly assigned ID code for the research participant. The third and fourth entries are the data and time (24-hour format) on the server when the data were received.

The task data consist of 25 trials on which participants make a "liking" rating that can range from 1 (dislike very very much) to 9 (like very very much). The liking rating is labeled "YouToJohn" in the table below, which gives the data format (excluding user information) The text line below the table is copied from the text file that is obtained from the DownloadData link on the PsychExperiments homepage. This is the actual data format including user information.

Trial Number	YouToBill	BillToJohn	YouToJohn	Trial Number	YouToBill	BillToJohn	YouToJohn	Trial Number	YouToBill	etc.
1	like	loves	6	2	dislike	likes	-1	3	love	etc.

ATI Workshop--CSUF,BEH683,8/2/2002,13:16,1,like ,loves ,6,2 ,dislike ,likes, -1, 3, love . ..

References

Birnbaum, M. (2001). Introduction to behavioral research on the Internet. Prentice Hall.

Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychology, 21, 107-112.

« Return to Top»

Development of this experiment was supported by the National Science Foundation under Grant BCS-0129453. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.