The Effect of Experimenter Bias and Demand Characteristics on Participant’s Responses Within the Bart Task

The effect of experimenter bias and demand characteristics on participant’s responses within the BART task

Emma Katherine Summerton

Abstract

When undertaking psychological studies we are given a set of instructions by the conductor of the study, however, how do we know these instructions do not influence the way we respond to the study thus rendering our results biased. The study performed was based upon Lejuez’s (2002) Balloon Analogue Risk Task (BART) in which risk based behaviour was monitored in a laboratory environment by asking participants to blow up a balloon in exchange for monetary rewards. The results collected indicate BART is a highly useful tool in terms of measuring risk taking behaviour, however in terms of participants conforming to experimenter bias, the results collected neither prove nor disprove that this has any real affect.  

Introduction

Studies have found demand characteristics are powerful in determining participants’ behaviour when the purpose of the study is conveyed effectively but not overtly obviously (Orne, 1962). Recent methods of measuring demand characteristics, such as Rubin et al. (2010) Perceived Awareness of the Research Hypothesis scale, suggest this behaviour can be scaled rather than just clarifying the existence of conformity. Many experiments have been conducted utilising Lejuez’s BART model in terms of measuring risk taking relating to another factor such as the value of rewards (Veling and Bijleveld, 2015). In this experiment, the BART task is used to measure the participant’s level of risk taking behaviour asking them within the study to measure their digit ratio and informing them whether this makes them more or less likely to take risks. In Lejuez’s BART task, the participant would be offered to cash-out after each pump or continue in order to possibly gain a greater reward but increasing the likelihood they would exceed the balloon’s threshold and earn nothing. As the balloon’s breakpoints were randomised, the participants could not predict how much to pump the balloon and thus tested the participants risk related behaviour in terms of when they believed the reward outweighed the loss. The data collected by Lejuez from the task showed BART to be a useful tool for examining risk taking for a laboratory study. The objective of this experiment is to discover whether experimenter bias, of being told they are more or less risk taking, causes participants to be more or less risk taking in the BART experiment. The predictions, based on the BART task being a highly useful diagnostic tool for measuring risk taking and studies in experimenter bias having shown experimenters are able to obtain the data they would like from participants (Fode, 1960; Rosenthal & Fode, 1961; Rosenthal, Fode, Friedman, & Vikan, 1960; Rosenthl, Fode, & Viltan, 1960; Rozenthal, Friedman, Johnson, Fode, Schill, White, & Vikan, 1960) , that experimenter bias will have an effect on participants risk taking behaviour.

Method

Design
The experiment was conducted in a private computer room, which had been booked for the purpose of the study, therefore provided a quiet focused environment. Participants before starting the trial were asked to calculate their digit ration which could be calculated by measuring their index finger on their left hand and dividing that by the length of their ring finger. There were twenty trials and within each trial a different condition was set for the balloon due to its variance in ‘breakPoint’ (the size in which the balloon would burst. This varied from .1 to  2, the full screen size, with screen coordinates ranging from (-1,-1) bottom left to (+1,+1) for top right. The size at each point in time being calculated as: 0.1 (nPumps x .015).  The aim of the experiment unbeknown to the participants being to see if due to being told they were more or less likely to take risks whether behaviour would be influenced in the experiment, thus making them less or more likely to take caution when pumping up the balloon. All participants took the test under the impression the experiment was about testosterones effect on risk taking behaviour, however Group A were told conflicting information about the study to Group B. Group B being told the higher their digit ratio the more masculine they were thus the more likely they were to take risks and Group A the converse. The experiment was a between subjects design making it easier to conduct the experiment with little contamination by extraneous factors.

Participants

Participants were placed randomly into groups with Group A consisting of 21 females and 3 males and Group B consisting of 20 females and 1 male with the age of participants in both groups ranging from 18-24 (Group A: M=18.5, SD= .659 Group B: M=18.7, SD=1.384).  All the participants were psychology undergraduates were given the same information about the study and participated due to it being a course requirement.  

Apparatus and Materials

The computers used were 27-inch Retina 5K display Apple Mackintoshes with IPS technology and 5120x2880 resolution in which the program PsychoPy (version 1.75) was run. The visual stimulus used was a red balloon (see Appendix I, Figure 2) which was programmed to pop when its maximum number of pumps was reached (Appendix I, Figure 1).  The participants were asked to wear headphones and if not to turn up the volume on their computer. An audial stimulus, “bang” (see Appendix I, Figure 3) was triggered if the number of pumps exceeded the maximum which then alerted the student they had received no money for that round.

Procedure

Upon entrance into the lab class, participants were assigned either group A or B and were then asked to open the associated powerpoint with that group. This was the independent variable. Once opened, the powerpoint informed the participants they were taking part in a social experiment which looked at the effects of testosterone levels on risk taking behaviours. Participants as a proxy measure for testosterone were asked to measure their digit ratio by dividing the length of their index finger by the length of their ring finger on their left hand from the to the crease where the hand and finger meet. Group A participants were told lower values indicate more testosterone which correlated with a more masculine character meanwhile Group B participants were told the opposite. However, both were told the average digit ratio was around 0.85 before being briefed on the different ways risk taking behaviour can be measured and the consequences and advantages of these. The participants were then told they were going to take part in the Balloon Analogue Risk Task (BART) which would measure their risk taking behaviour by asking the individual to press space to blow up a balloon with the incentive of the more they blew up the balloon the higher the financial reward. Participants were told men and people with higher testosterone tended to be more risk-taking before being asked to run the PsychoPy file.