pith. sign in

arxiv: 2508.00462 · v2 · submitted 2025-08-01 · 💻 cs.SE

Managing Power Gaps as an Element of Pair Programming Skill: A Grounded Theory

Linus Ververs , Janina Berger , Lutz Prechelt This is my paper

Pith reviewed 2026-05-19 01:44 UTC · model grok-4.3

classification 💻 cs.SE
keywords pair programmingpower gapgrounded theorysoftware engineeringteam dynamicsknowledge gapequalizing behavior
0
0 comments X

The pith

Power gaps can make pair programming dysfunctional even when partners understand each other well.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper uses grounded theory to show how power gaps arise in pair programming sessions. It explains that these gaps form from differences in knowledge combined with hierarchical behaviors. Even with strong togetherness, the gaps lead to defensive or disengaging actions that harm the session. Equalizing behaviors can reduce or prevent these issues. Practitioners can build better pair programming skills by recognizing and managing power gaps.

Core claim

We explain how a Power Gap can make a session dysfunctional despite the presence of high Togetherness, how it comes into existence due to a Knowledge Gap and Hierarchical Behavior, why its consequences (Defensive Behavior and Disengaging Behavior) are problematic, and how it can be reduced or prevented by Equalizing Behavior. Pair programming practitioners can improve their pair programming skill by unlearning problematic behaviors related to Power Gaps and by learning to recognize Power Gaps and apply Equalizing Behavior.

What carries the argument

The Power Gap, which arises from a Knowledge Gap combined with Hierarchical Behavior and produces Defensive Behavior and Disengaging Behavior unless countered by Equalizing Behavior.

If this is right

  • Pair programming sessions can fail even with high togetherness if a power gap is present.
  • Power gaps form when knowledge differences meet hierarchical behaviors.
  • Defensive and disengaging behaviors harm session effectiveness as a direct result.
  • Equalizing behavior can reduce or prevent power gaps and their negative effects.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Training programs for developers could add specific practice in spotting and addressing power imbalances during pairing.
  • The same dynamics may appear in other close-collaboration coding activities such as mob programming.
  • Replicating the study with teams outside Germany would test whether the patterns hold more broadly.

Load-bearing premise

The patterns observed in the 21 sessions from five German companies and six interviews from four other German companies reflect general mechanisms of power dynamics in pair programming rather than being limited to the sampled cultural or organizational context.

What would settle it

A study of pair programming sessions in other cultural or organizational settings where knowledge gaps and hierarchical behavior do not lead to defensive or disengaging behavior would contradict the mechanism.

Figures

Figures reproduced from arXiv: 2508.00462 by Janina Berger, Linus Ververs, Lutz Prechelt.

Figure 1
Figure 1. Figure 1: How often partners feel in control of decision mak [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Six forms of Unbalanced Process Agency (𝑁 = 119) Unbalanced Process Agency. Example 3 illustrates one form this can take: Example 3: DA2 - Proposed Driver-Observer Switch It’s D4’s first week in company D. D3 is tasked with onboarding D4. They use PP to implement a new UI feature in Java. Their interaction quickly changes as newcomer D3 has more general knowledge of Java and software design than D4, and D3… view at source ↗
Figure 3
Figure 3. Figure 3: Does a Knowledge Gap lead to a Power Gap? (𝑁 = 92) Unbalanced Process Agency can have a direct negative effect on the Quality of the work results, as only one developer is influencing and responsible for the decision-making process A6. Compared to solo programming, PP “appears to have a positive effect on quality” [17, Section 2.3]3 . This is also backed up by Arisholm et al.’s large-scale experiment on PP… view at source ↗
Figure 4
Figure 4. Figure 4: Does Unbalanced Process Agency harm the advan￾tages of PP? (𝑁 = 92) 3Zieris extends the meta-analysis of Hannay et al. [6] with data from two other studies [15, 16] to reach this conclusion. The Power Gap and Unbalanced Process Agency are at the heart of our Grounded Theory. They have various causes and effects (and further relationships to other concepts). We describe these relation￾ships, clustered into … view at source ↗
Figure 5
Figure 5. Figure 5: How common is Hierarchical Behavior? (𝑁 = 119) Hierarchical Behavior can lead to Unbalanced Process Agency; see again Figure 2a. The first three statements on Unbalanced Pro￾cess Agency are related to a too-dominant partner. Hierarchical Behavior is probably at play here. The majority of respondents have experienced this, although only 10% sometimes or more. One par￾ticipant acknowledged this directly: “So… view at source ↗
Figure 6
Figure 6. Figure 6: How common is Equalizing Behavior? (𝑁 = 119) One participant put it this way: “[. . . ] I think there is a lot of value in unequal pairings with respect to expertise, but not in terms of respect or dignity. I am now in a place where I am generally brought in as an expert, and often find myself teaching others, which is technically ‘not equal’, but I always try to approach problems and other people with app… view at source ↗
Figure 8
Figure 8. Figure 8: Does a Power Gap lead to Defensive Behavior? (𝑁 = 92) A13). Worse, by blocking Equalizing Behavior, Defensive Behavior makes it hard to reduce a Power Gap. 4.6 Disengaging Behavior In Disengaging Behavior, a partner gradually withdraws from the session by not asking questions, not proposing anything, and not challenging their partner’s proposals as a result of a Power Gap (Reference to GT-Arrow A7). Diseng… view at source ↗
Figure 9
Figure 9. Figure 9: Does a Power Gap lead to Disengaging Behavior? (𝑁 = 92) Besides engaging less in the decision-making process, Disengag￾ing Behavior also includes not asking questions. This damages the Knowledge Transfer, which is one important advantage of PP [17] (Reference to GT-Arrow A9). By damaging the Knowledge Transfer, the developer hinders closing any present Knowledge Gap (Refer￾ence to GT-Arrow A10). 4.7 Modera… view at source ↗
Figure 10
Figure 10. Figure 10: How present is Fragile Ego during PP? (𝑁 = 119) This is in line with our observational data, as most developers ask questions frequently or directly state their knowledge gap, and most partners respond by directly providing the missing knowledge and not commenting in any other way on the Knowledge Gap. Summing up, Fragile Ego appears to be a relevant factor for the Power Gap dynamic of PP sessions (Refere… view at source ↗
Figure 11
Figure 11. Figure 11: Does a strong Bond lead to Meta-Communication? (𝑁 = 92) Two survey participants put it this way: “[PP] was uncomfortable at first, since I was new and didn’t know anyone. But once I got to know everyone better, it became my preferred way of working [...].” (Q9:P204) and “Pairing always worked particularly well for us when there was already psychological safety within the team. When there were underlying t… view at source ↗
read the original abstract

Background: In pair programming, Togetherness (the partners understand each other's mental state well) is a main success factor. Maintaining high Togetherness is an element of pair programming skill. Some sessions appear to go badly although Togetherness appears good. Objective: Understand under what circumstances this is possible. Method: Grounded Theory Methodology based on 21 recorded pair programming sessions with 22 developers from 5 German software companies and 6 interviews with different developers from 4 other German companies. Results: We explain how a Power Gap can make a session dysfunctional despite the presence of high Togetherness, how it comes into existence due to a Knowledge Gap and Hierarchical Behavior, why its consequences (Defensive Behavior and Disengaging Behavior) are problematic, and how it can be reduced or prevented by Equalizing Behavior. Conclusions: Pair programming practitioners can improve their pair programming skill by unlearning problematic behaviors related to Power Gaps and by learning to recognize Power Gaps and apply Equalizing Behavior.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript presents a grounded theory study on power dynamics in pair programming. Drawing from 21 recorded pair programming sessions involving 22 developers from 5 German software companies and 6 interviews with developers from 4 other German companies, the authors develop a theory explaining how 'Power Gaps' can disrupt sessions despite high 'Togetherness'. The theory traces Power Gaps to 'Knowledge Gaps' and 'Hierarchical Behavior', leading to 'Defensive Behavior' and 'Disengaging Behavior', which can be mitigated by 'Equalizing Behavior'.

Significance. If the derived categories and relationships are valid, this study provides valuable insights into interpersonal dynamics in pair programming, extending beyond technical skills to include management of power imbalances. It has potential to inform training programs and best practices in software development teams, particularly in agile methodologies. The empirical basis from real sessions is a positive aspect.

major comments (2)
  1. Methods section: The description of the grounded theory process lacks detail on specific coding procedures (open, axial, selective coding), criteria for theoretical saturation, and whether member checking or other validation techniques were employed. This makes it difficult to assess the rigor of how raw data from sessions and interviews led to the reported categories like Power Gap and Equalizing Behavior.
  2. Results and Discussion sections: The study is limited to German companies. The paper should address whether the identified mechanisms (Knowledge Gap leading to Hierarchical Behavior etc.) are influenced by local factors such as labor laws or company cultures, and discuss implications for generalizability to other national or organizational contexts, as this is central to claiming these as general elements of pair programming skill.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback, which helps strengthen the clarity and contextual framing of our grounded theory study on power dynamics in pair programming. We address each major comment below and outline the revisions planned for the next manuscript version.

read point-by-point responses

  1. Referee: Methods section: The description of the grounded theory process lacks detail on specific coding procedures (open, axial, selective coding), criteria for theoretical saturation, and whether member checking or other validation techniques were employed. This makes it difficult to assess the rigor of how raw data from sessions and interviews led to the reported categories like Power Gap and Equalizing Behavior.

    Authors: We agree that the Methods section would benefit from greater specificity. The original manuscript outlined the overall grounded theory approach and data sources but did not explicitly detail the progression through open coding (initial concept identification from transcripts), axial coding (linking categories such as Knowledge Gap to Hierarchical Behavior), and selective coding (integrating around the core category of Power Gap). Theoretical saturation was reached when additional sessions and interviews yielded no new categories or relationships, but this criterion was not stated. Validation relied on team peer debriefing rather than formal member checking. We will revise the Methods section to provide a step-by-step account of the coding procedures, include brief examples of category emergence, specify saturation criteria, and describe the validation techniques employed. This will improve transparency without altering the underlying analysis. revision: yes

  2. Referee: Results and Discussion sections: The study is limited to German companies. The paper should address whether the identified mechanisms (Knowledge Gap leading to Hierarchical Behavior etc.) are influenced by local factors such as labor laws or company cultures, and discuss implications for generalizability to other national or organizational contexts, as this is central to claiming these as general elements of pair programming skill.

    Authors: The referee rightly notes the sample's restriction to German software companies, which may shape hierarchical behaviors through factors such as strong labor protections and relatively low power-distance norms in many firms. We will expand the Discussion section to explicitly consider how these contextual elements could influence the emergence and mitigation of Power Gaps. We will also add a limitations subsection clarifying that the theory offers transferable insights into pair programming dynamics but that empirical confirmation in other national or organizational settings (e.g., higher power-distance cultures) remains necessary. This discussion will be added without overstating the current evidence base. revision: yes

Circularity Check

0 steps flagged

No significant circularity: concepts emerge from data analysis

full rationale

The paper applies Grounded Theory Methodology to 21 recorded sessions and 6 interviews, deriving categories such as Power Gap arising from Knowledge Gap plus Hierarchical Behavior, leading to Defensive/Disengaging Behavior and mitigated by Equalizing Behavior. These are presented as patterns extracted from the observations rather than presupposed by definitions, equations, or prior self-citations. No load-bearing steps reduce to fitted inputs, self-referential definitions, or ansatzes imported via citation; the central claims remain independent of the sampled data by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

The study rests on the standard assumptions of grounded theory methodology and introduces several new analytic concepts derived from the data.

axioms (1)
  • domain assumption Grounded Theory Methodology is an appropriate and rigorous approach for developing explanatory theory from qualitative observations of pair programming sessions.
    The method section invokes GTM to justify the emergence of categories such as Power Gap and Equalizing Behavior from the recorded sessions and interviews.
invented entities (2)
  • Power Gap no independent evidence
    purpose: Explains why a pair programming session can become dysfunctional even when Togetherness appears high.
    Introduced as a central category emerging from the analysis of sessions where one partner exerts or experiences unequal influence.
  • Equalizing Behavior no independent evidence
    purpose: Describes actions that reduce or prevent Power Gaps.
    Derived from observed or reported developer actions that restore balance in the pair.

pith-pipeline@v0.9.0 · 5702 in / 1442 out tokens · 26996 ms · 2026-05-19T01:44:13.541986+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

  1. [1]

    Erik Arisholm, Hans Gallis, Tore Dybå, and Dag I. K. Sjøberg. 2007. Evaluating Pair Programming with Respect to System Complexity and Programmer Expertise. IEEE Trans. Software Eng. 33, 2 (2007), 65–86. doi:10.1109/TSE.2007.17

    work page doi:10.1109/tse.2007.17 2007

  2. [2]

    Kent Beck. 1999. Extreme programming explained: embrace change . Addison- Wesley Longman Publishing Co., Inc., USA

    work page 1999

  3. [3]

    Andrew Begel and Nachiappan Nagappan. 2008. Pair programming: what’s in it for me?. In Proceedings of the Second International Symposium on Empirical Soft- ware Engineering and Measurement, ESEM 2008, October 9-10, 2008, Kaiserslautern, Germany, H. Dieter Rombach, Sebastian G. Elbaum, and Jürgen Münch (Eds.). ACM, 120–128. doi:10.1145/1414004.1414026

    work page doi:10.1145/1414004.1414026 2008

  4. [4]

    Culture Factor Group. 2023. Country Comparison tool. https://www. theculturefactor.com/country-comparison-tool

    work page 2023

  5. [5]

    Hall and M.R

    E.T. Hall and M.R. Hall. 1990. Understanding Cultural Differences: Germans, French and Americans. Intercultural Press. https://books.google.de/books?id= Hr3adyadHC4C

    work page 1990

  6. [6]

    Jo Erskine Hannay, Tore Dybå, Erik Arisholm, and Dag I. K. Sjøberg. 2009. The effectiveness of pair programming: A meta-analysis. Inf. Softw. Technol. 51, 7 (2009), 1110–1122. doi:10.1016/J.INFSOF.2009.02.001

    work page doi:10.1016/j.infsof.2009.02.001 2009

  7. [7]

    Laura Plonka, Helen Sharp, and Janet van der Linden. 2012. Disengagement in pair programming: Does it matter?. In 34th International Conference on Software Engineering, ICSE 2012, June 2-9, 2012, Zurich, Switzerland , Martin Glinz, Gail C. Murphy, and Mauro Pezzè (Eds.). IEEE Computer Society, 496–506. doi:10.1109/ ICSE.2012.6227166

    work page arXiv 2012

  8. [8]

    Stephan Salinger and Lutz Prechelt. 2013. Understanding Pair Programming: The Base Layer. BoD–Books on Demand

    work page 2013

  9. [9]

    Stephan Salinger, Franz Zieris, and Lutz Prechelt. 2013. Liberating pair program- ming research from the oppressive Driver/Observer regime. In 35th International Conference on Software Engineering, ICSE ’13, San Francisco, CA, USA, May 18-26, 4This relationship is not directly observable and so our data has not produced such an arrow in the GT figure, but...

    work page doi:10.1109/icse.2013.6606678 2013

  10. [10]

    Julia Schenk, Lutz Prechelt, and Stephan Salinger. 2014. Distributed-pair pro- gramming can work well and is not just distributed pair-programming. In 36th International Conference on Software Engineering, ICSE ’14, Companion Proceed- ings, Hyderabad, India, May 31 - June 07, 2014 , Pankaj Jalote, Lionel C. Briand, and André van der Hoek (Eds.). ACM, 74–8...

    work page doi:10.1145/2591062.2591188 2014

  11. [11]

    Strauss and Juliet M

    Anselm L. Strauss and Juliet M. Corbin. 1990. Basics of Qualitative Research: Grounded Theory Procedures and Techniques . SAGE Publications, Inc

    work page 1990

  12. [12]

    big-tent

    Sarah J Tracy. 2010. Qualitative quality: Eight “big-tent” criteria for excellent qualitative research. Qualitative inquiry 16, 10 (2010), 837–851

    work page 2010

  13. [13]

    Linus Ververs, Trang Linh Lam, Janina Berger, and Lutz Prechelt. 2025. A Grounded Theory on the Teacher and Student Roles in Pair Programming. arXiv:2507.10305 [cs.LG] Preprint, https://doi.org/10.48550/arXiv.2507.10305

    work page doi:10.48550/arxiv.2507.10305 2025

  14. [14]

    Williams, Robert R

    Laurie A. Williams, Robert R. Kessler, Ward Cunningham, and Ron Jeffries. 2000. Strengthening the Case for Pair Programming. IEEE Softw. 17, 4 (2000), 19–25. doi:10.1109/52.854064

    work page doi:10.1109/52.854064 2000

  15. [15]

    Wilson, John T

    Judith D. Wilson, John T. Nosek, Nathan Hoskin, and Lin L. Liou. 1992. The Effect of Collaboration on Problem-Solving Performance Among Programmers. In Algorithms, Software, Architecture - Information Processing ’92, Volume 1, Pro- ceedings of the IFIP 12th World Computer Congress, Madrid, Spain, 7-11 September 1992 (IFIP Transactions, Vol. A-12), Jan van...

    work page 1992

  16. [16]

    Zacharis

    Nick Z. Zacharis. 2011. Measuring the Effects of Virtual Pair Programming in an Introductory Programming Java Course. IEEE Trans. Educ. 54, 1 (2011), 168–170. doi:10.1109/TE.2010.2048328

    work page doi:10.1109/te.2010.2048328 2011

  17. [17]

    2020.Qualitative Analysis of Knowledge Transfer in Pair Programming

    Franz Zieris. 2020.Qualitative Analysis of Knowledge Transfer in Pair Programming. Ph. D. Dissertation. Free University of Berlin, Germany. https://refubium.fu- berlin.de/handle/fub188/28968

    work page 2020

  18. [18]

    Franz Zieris and Lutz Prechelt. 2014. On knowledge transfer skill in pair pro- gramming. In 2014 ACM-IEEE International Symposium on Empirical Software Engineering and Measurement, ESEM ’14, Torino, Italy, September 18-19, 2014 , Maurizio Morisio, Tore Dybå, and Marco Torchiano (Eds.). ACM, 11:1–11:10. doi:10.1145/2652524.2652529

    work page doi:10.1145/2652524.2652529 2014

  19. [19]

    Franz Zieris and Lutz Prechelt. 2016. Observations on knowledge transfer of professional software developers during pair programming. In Proceedings of the 38th International Conference on Software Engineering, ICSE 2016, Austin, TX, USA, May 14-22, 2016 - Companion Volume , Laura K. Dillon, Willem Visser, and Laurie A. Williams (Eds.). ACM, 242–250. doi:...

    work page doi:10.1145/2889160.2889249 2016

  20. [20]

    Franz Zieris and Lutz Prechelt. 2020. PP-ind: A Repository of Industrial Pair Programming Session Recordings. CoRR abs/2002.03121 (2020). arXiv:2002.03121 https://arxiv.org/abs/2002.03121

    work page arXiv 2020

  21. [21]

    Franz Zieris and Lutz Prechelt. 2021. Two Elements of Pair Programming Skill. In 43rd IEEE/ACM International Conference on Software Engineering: New Ideas and Emerging Results, ICSE (NIER) 2021, Madrid, Spain, May 25-28, 2021 . IEEE, 51–55. doi:10.1109/ICSE-NIER52604.2021.00019

    work page doi:10.1109/icse-nier52604.2021.00019 2021