:::info
Authors:
- Sira Vegas
- Patricia Riofr´ıo
- Esperanza Marcos
- Natalia Juristo
:::
Table Of Links
Abstract
1 Introduction
2 Original Study: Research Questions and Methodology
3 Original Study: Validity Threats
4 Original Study: Results
5 Replicated Study: Research Questions and Methodology
6 Replicated Study: Validity Threats
7 Replicated Study: Results
8 Discussion
9 Related Work
10 Conclusions And References
Abstract
A recurring problem in software development is incorrect decision making on the techniques, methods and tools to be used. Mostly, these decisions are based on developers’ perceptions about them. A factor influencing people’s perceptions is past experience, but it is not the only one. In this research, we aim to discover how well the perceptions of the defect detection effectiveness of different techniques match their real effectiveness in the absence of prior experience.
To do this, we conduct an empirical study plus a replication. During the original study, we conduct a controlled experiment with students applying two testing techniques and a code review technique. At the end of the experiment, they take a survey to find out which technique they perceive to be most effective. The results show that participants’ perceptions are wrong and that this mismatch is costly in terms of quality.
In order to gain further insight into the results, we replicate the controlled experiment and extend the survey to include questions about participants’ opinions on the techniques and programs. The results of the replicated study confirm the findings of the original study and suggest that participants’ perceptions might be based not on their opinions about complexity or preferences for techniques but on how well they think that they have applied the techniques.
1. Introduction
An increasingly more popular practice nowadays is for software development companies to let developers choose their own technological environment. This means that different developers may use different productivity tools (programming language, IDE, etc.). However, software engineering (SE) is a humanintensive discipline where wrong decisions can potentially compromise the quality of the resulting software. In SE, decisions on which methods, techniques and tools to use in software development are typically based on developers’ perceptions and/or opinions rather than evidence, as suggested by Dyb˚a et al. [19] and Zelkowitz et al. [55].
However, empirical evidence might not be available, as certain methods, techniques or tools may not have been studied within a particular setting or even at all. Alternatively, developers may simply not be acquainted with such studies, according to Vegas & Basili [49]. On this ground, it is important to discover how well developers perceptions (beliefs) match reality and, if they do not, find out what is behind this mismatch, as noted by Devanbu et al.[14]. According to Psychology, experience plays a role in people’s perceptions. This has also been observed by Devanbu et al. [14] in SE.
However, this research sets out to discover how well matched perceptions are with reality in the absence of previous experience in the technology being used. This makes sense for several reasons: 1) experience is not the only factor affecting developers’ perceptions; 2) development teams are usually composed of a mix of people with and without experience; and 3) it is not clear what type of experience influences perceptions. For example, Dieste et al. [17] conclude that academic rather than professional experience could be affecting the external quality of the code generated by developers when applying Test-Driven Development.
We aim to study whether perceptions about the effectiveness of three defect detection techniques match reality, and if not, what is behind these perceptions. To the best of our knowledge, this is the first paper to empirically assess this issue. To this end, we conducted an empirical study plus a replication with students. During the original study we measured (as part of a controlled experiment) the effectiveness of two testing techniques and one code review technique when applied by the participants. We then checked the perceived most effective technique (gathered by means of a survey) against the real one.
Additionally, we analysed the cost of the mismatch between perceptions and reality in terms of loss of effectiveness. Major findings include:
– Different people perceive different techniques to be more effective. No one technique is perceived as being more effective than the others. – The perceptions of 50% of participants (11 out of 23) are wrong.
– Wrong perception of techniques can reduce effectiveness 31pp (percentage points) on average.
These findings led us to extend the goal of the study in a replication to investigate what could be behind participants’ perceptions. To do this, we examined their opinions on the techniques they applied and the programs they tested in a replication of the controlled experiment. Major findings include:
– The results of the replication confirm the findings of the original study.
– Participants think that technique effectiveness depends exclusively on their performance and not on possible weaknesses of the technique itself. – The opinions about technique complexity and preferences for techniques do not seem to play a role in perceived effectiveness. These results are useful for developers and researchers. They suggest:
– Developers should become aware of the limitations of their judgement.
– Tools should be designed that provide feedback to developers on how effective techniques are.
– The best combination of techniques to apply should be determined that is at the same time easily applicable and effective. – Instruments should be developed to make empirical results available to developers.
The material associated to the studies presented here can be found at https://github.com/GRISE-UPM/Misperceptions. The article is organised as follows. Section 2 describes the original study. Section 3 presents its validity threats. Section 4 discusses the results. Section 5 describes the replicated study based on the modifications made to the original study. Section 6 presents its validity threats. Section 7 reports the results of this replicated study. Section 8 discusses our findings and their implications. Section 9 shows related work. Finally, Section 10 outlines the conclusions of this work.
:::info
This paper is available on arxiv under CC BY-NC-ND 4.0 license.
:::
