Question

The paper "Deception and Design: The Impact of Communication Technology on Lying Behavior" describes an investigation into whether lying is less common in face-to-face communication than in other forms of communication such as phone conversations or e-mail. Participants in this study were 30 students in an upper-division communications course at Cornell University who received course credit for participation. Participants were asked to record all of their social interactions for a week, making note of any lies told.

Based on data from these records, the authors of the paper concluded that students lie more often in phone conversations than in face-to-face conversations and more often in face-to-face conversations than in e-mail. Discuss the limitations of this study, commenting on the way the sample was selected and potential sources of bias. (Select all that apply.)

-Since the students were all from an upper-division communications course at Cornell, it is not reasonable to consider these students to be representative of all students.

-Students may not have admitted all the lies they told, which would distort the results of the study.

-The study only asked 30 students, instead of the whole class.

-Participants were asked to record all of their social interactions for a week, instead of a month.

-Students knew that they were surveying themselves as to the truthfulness of their interactions, which could have changed students behavior and distorted the results of the study.

Answer 1

Ans.

Certain limitations of this study conducted can be noted down as the following:

• Looking at the objective of the study one can reasonably say that this was done to investigate the lying phenomenon face-to-face vs via some communicating medium (mail, mobile phones,etc) in everyone since everyone nowadays uses a phone and/ or a computer. So it should have taken into account the total population of people who use these medium to communicate. But the sampling scheme adopted in this survey is convenient sampling since it only considered the students of Cornell University, which itself is a non-probability sampling scheme and is prone to bias.
• Even the students chosen were all of the same age group and none from the other classes were selected (when this experiment was equally applicable to them as well) . So we can only check the intra-class variability but not the inter-class variability . Hence the 30 students belonging to a single class do not represent the entire student body since there are students belonging to different age group as well
• The sample size was only 30, which is quite small when the theoretical target population of study is quite big.
• The participants were asked to collect the data without any supervision. This decreases the credibility of the data.
• The students could have distorted the data since some of them might not be willing to disclose the number of lies by them since the experiment is most likely conducted by their own school (thus very likely that it'll be looked on by their teachers) and they might think that their answer might influence the credits they are to receive for participation or maybe concerned about their image in front of others.
• The data was collected for a time span of 1 week which is quite a small interval of time to be used for experimental purposes of these types as these type of experiments are neither expensive nor time consuming.
• Many students can forget to jot down their lie-data since people tend to lose track of time and topics while talking on the phone. This may lead to human errors like guessing, bias for certain numbers, wrong rounding off, etc.