The Most Popular Research Methods and Analysis Techniques in Management and Organisation Research

About this report

This technical report maps the research methods and analysis techniques that currently dominate management and organisation research. It analyses the abstracts of 759 articles published in 2024 to 2026 in six top journals of the field, indexed in the Web of Science Core Collection: Strategic Management Journal (SMJ), Journal of Management (JOM), Journal of Organizational Behavior (JOB), Academy of Management Journal (AMJ), Academy of Management Review (AMR) and Academy of Management Annals (AMA). The method is dictionary-based text mining: abstracts were screened with an extended dictionary of 11 data collection categories (Dimension 1) and 30 analysis technique categories (Dimension 2). Each article could receive more than one label. Journal differences were tested with chi-square tests, and annual trends are reported as percentage-point (pp) changes.

1. Data Collection Methods (Dimension 1): Overall Frequencies

At least one data collection method was detected in 43.6% of the 759 articles (n = 331). The remaining 56.4% consists mainly of conceptual and theoretical papers in AMR, review and integrative studies in AMA, and articles whose abstracts do not state the method explicitly.

Table 1. Frequency distribution of data collection methods and journal profile (N = 759)

Finding: Surveys (12.8%) and archival or secondary data (11.7%) are the two dominant data sources in the field. Together they account for 56% of all detected data collection methods.

Finding: ESM and diary studies (4.5%) are almost exclusive to JOB (17.1%), reflecting the journal’s micro-level identity focused on daily experience.

1.1. Journal differences: chi-square tests

Differences in data collection preferences across the four empirical journals (SMJ, JOM, JOB, AMJ) were examined with chi-square tests.

Table 2. Chi-square tests of Dimension 1 categories across journals

Finding: The strongest differentiation is in ESM and diary methods (χ² = 61.24, p < .001): JOB practically holds a monopoly on this method. The use of archival data and surveys also differs markedly across journals.

1.2. Annual trends (2024 to 2026)

Table 3. Annual trends of Dimension 1 methods (%, percentage-point change)

Finding: The interview method jumped from 4.4% in 2024 to 10.2% in 2026 (+5.8 pp), indicating growing interest in qualitative research. Text and NLP data are also rising steadily.

2. Analysis Techniques (Dimension 2): Overall Frequencies

Screening with an extended dictionary of 30 analysis technique categories detected at least one specific technique in 36.4% of the 759 articles (n = 276). It is expected that this rate is lower than for Dimension 1: abstracts more often contain general expressions such as “regression analysis” than names of specific techniques.

Table 4. Frequency ranking of analysis techniques, top 15 (N = 759)

Finding: SEM and path analysis (12.5%) and the related mediation and moderation testing (12.4%) form the most dominant analytical framework in the field. The two categories overlap by 95% (90 of the 95 SEM articles also test mediation or moderation).

Finding: DiD (4.0%) is the most common causal inference method and is firmly positioned in strategy research (SMJ 6.6%). It is followed by IV / 2SLS (1.6%), PSM (0.5%) and RDD (0.4%).

Finding: Qualitative analysis techniques (process theory, thematic analysis, case study, grounded theory) together account for 8.8% and form a strong second block after quantitative methods.

2.1. Journal differences: chi-square tests

Table 5. Chi-square tests of Dimension 2 categories across journals (df = 3)

Finding: The most dramatic differentiation is in SEM and mediation: one third of JOB articles use this approach, compared with 8% in SMJ. This reflects the difference in methodological culture between organisational behaviour and strategy.

2.2. Annual trends: rising and declining techniques

Table 6. Annual trends of Dimension 2 techniques (%, percentage-point change)

Finding: The use of SEM and mediation almost doubled between 2024 and 2026 (8.8% to 16.9%), showing that demand for mediation and moderation frameworks continues to grow.

Finding: NLP and text analytics rose from 0.9% in 2024 to 4.2% in 2026 (+3.3 pp). Despite the small n, the upward trend is consistent and signals that computational methods are gaining ground.

Finding: Traditional qualitative techniques such as thematic analysis and grounded theory are declining. This may mean not that qualitative research is disappearing, but that it is evolving towards more sophisticated qualitative approaches such as process theory and narrative analysis.

3. Cross-Analyses: Data Method × Analysis Technique

This section shows which data collection methods are used together with which analysis techniques. These co-occurrence patterns define the “methodological bundles” of the field.

3.1. Most common Dimension 1 × Dimension 2 combinations

Table 7. Most common data collection × analysis technique combinations

3.2. Co-occurrence of analysis techniques

How often two different analysis techniques are used in the same article:

Table 8. Most frequent Dimension 2 × Dimension 2 technique pairs

Finding: SEM and mediation or moderation are an almost inseparable pair (90 shared articles). Management researchers use SEM mainly to test mediation and moderation hypotheses.

Finding: DiD combined with mediation or moderation (n = 6 to 7) is a rising sophisticated combination that integrates causal effect identification with explanation of mechanisms.

4. Advanced Findings

4.1. Adoption of causal inference methods

Adoption rates of the “causal inference toolkit” (DiD, IV / 2SLS, RDD, PSM and event study) by journal:

Finding: The overall adoption rate of causal inference methods is 7.3%, with SMJ leading (9.9%). The transfer of these methods from economics to management is still at an early stage.

4.2. The spread of multi-study designs

Multi-study designs consisting of two or more independent studies have an overall rate of 4.9% (n = 37). By journal: JOB (12.5%) far ahead of JOM (6.9%), AMJ (6.5%) and SMJ (0.0%). By year: 2024 (3.5%), 2025 (3.9%), 2026 (11.0%). Multi-study designs tripled in 2026, indicating stronger editorial expectations and a strengthening replication culture.

Finding: Multi-study designs rose to 11% in 2026 (2024: 3.5%). They have become a strong norm especially in JOB (12.5%), where editors encourage replication across multiple samples and contexts.

4.3. The rise of technology-supported research methods

The status of AI / ML and computational methods both as research tools and as research topics:

Finding: Computational and big data methods rose to 3.4% in 2026 (2024: 1.3%), and NLP text methods jumped to 2.5%. This signals the beginning of a “computational turn” in management research.

5. Methodological Profiles of the Journals

SMJ (Strategic Management Journal). Archival and secondary data (20.8%) is the dominant data source; it leads in causal inference with DiD (6.6%) and has the highest ML / AI share (3.3%); SEM is low (8.0%); case study (3.8%) is its preferred qualitative design. Profile: an econometric and causal identity, the journal most open to computational methods.

JOM (Journal of Management). A balanced profile: archival data (15.6%), surveys (11.9%), interviews (8.8%); strong SEM (11.2%) and mediation (10.0%); balanced DiD (5.0%) and meta-analysis (2.5%). Profile: methodological pluralism.

JOB (Journal of Organizational Behavior). Surveys (27.0%), experiments (13.2%), ESM / diary (17.1%) and observation (12.5%) dominate; SEM (32.9%) and mediation (33.6%) are by far the highest; HLM (9.2%) is strong; multi-study designs (12.5%) are the norm. Profile: a micro-level, individual-focused identity centred on SEM and mediation.

AMJ (Academy of Management Journal). A varied mix of interviews (11.1%), archival data (16.7%) and observation (8.3%); the highest shares of SNA (7.4%) and process theory (8.3%); low SEM (4.6%). Profile: methodological diversity and a qualitative-quantitative balance, with a tradition of process research.

AMR (Academy of Management Review). As expected, predominantly conceptual and theoretical; the share of empirical methods is minimal. Process theory (7.4%) is the most discussed approach.

AMA (Academy of Management Annals). Review and integrative studies. Text / NLP (6.5%) has the highest share as a data source, reflecting bibliometric and systematic review methods.

6. Synthesis: The Methodological Landscape of the Field

6.1. Five core methodological bundles

The results reveal five dominant methodological bundles in management and organisation research:

1. SEM-mediation bundle: survey or ESM data, then SEM / CFA, then mediation or moderation testing. The most common bundle in the field (12 to 13%), dominant in JOB at 33% and increasingly reinforced by multi-study designs.
2. Econometric-causal bundle: archival data, then panel FE, DiD or IV, then causal effect estimation. Strongest in SMJ (9.9%); the gold standard of strategy research.
3. Qualitative-process bundle: interviews or observation, then process theory, grounded theory or case study. Strong in AMJ (8.3% process theory). Declining, but still a core research tradition.
4. Multilevel micro bundle: ESM or diary data, then HLM / multilevel models, then within-person mediation. Specific to JOB (17.1% ESM, 9.2% HLM); research on daily experience and fluctuation.
5. Computational / NLP bundle: text or platform data, then NLP / ML, then computational analysis. Still at 2 to 3%, but the fastest-growing bundle, with a marked jump in 2026.

6.2. Signals of methodological transformation

Rising trends: (a) accelerating use of SEM and mediation, (b) the entry of NLP and computational methods, (c) multi-study designs becoming normative (11% in 2026), (d) a revival of the interview method (+5.8 pp).

Declining trends: (a) traditional thematic analysis and grounded theory are in decline, (b) an unexpected stall in causal inference methods (a fall to 3.4% in 2026), (c) panel data models are under-represented.

Structural gaps: (a) Bayesian methods are almost never used (0.3%), (b) QCA / fsQCA is still niche (1.1%) but rising, (c) PLS-SEM has zero representation in these top journals, where CB-SEM is positioned as the only accepted form of SEM.

6.3. Implications for health management research

These findings carry several strategic implications for health management researchers:

• Adopting causal inference methods, especially DiD and IV / 2SLS, can provide a strong publication advantage in health policy research. Natural experiments in health systems (policy changes, reform implementation) offer a rich laboratory for these methods.
• The rise of NLP and text mining methods holds great potential for analysing patient feedback data, clinical notes and social media in health management.
• Multi-study designs are becoming an increasingly strong normative expectation for Q1 journals. Replication with samples from multiple hospitals or countries increases the likelihood of publication for health management studies.

6.4. Methodological limitations

The main limitations of this analysis are: (1) only abstract text was screened, and full-text analysis could yield higher coverage rates; (2) the regex dictionary is context-insensitive, so polysemous terms such as “model” may produce false positives; (3) the 2026 data cover only part of the year (n = 118, records indexed up to the search date), so 2026 percentages are provisional and trends ending in 2026 should be read with caution; (4) the SEM category overlaps heavily with mediation and moderation, so it is more accurate to evaluate the two categories together.

Appendix A: Abbreviations

Journals: AMA, Academy of Management Annals; AMJ, Academy of Management Journal; AMR, Academy of Management Review; JOB, Journal of Organizational Behavior; JOM, Journal of Management; SMJ, Strategic Management Journal.

Statistical and methodological: 2SLS, two-stage least squares; ANOVA, analysis of variance; BERT, Bidirectional Encoder Representations from Transformers; CAR, cumulative abnormal return; CB-SEM, covariance-based structural equation modelling; CEM, coarsened exact matching; CFA, confirmatory factor analysis; CI, confidence interval; DiD, difference-in-differences; EFA, exploratory factor analysis; EMA, ecological momentary assessment; ESM, experience sampling method; FE, fixed effects; fsQCA, fuzzy-set qualitative comparative analysis; GMM, generalised method of moments; GPT, generative pre-trained transformer; HLM, hierarchical linear modelling; IV, instrumental variable; LDA, latent Dirichlet allocation; LLM, large language model; MANOVA, multivariate analysis of variance; MASEM, meta-analytic structural equation modelling; MCMC, Markov chain Monte Carlo; ML, machine learning; NLP, natural language processing; OB, organisational behaviour; OLS, ordinary least squares; PLS-SEM, partial least squares SEM; PSM, propensity score matching; QCA, qualitative comparative analysis; RCT, randomised controlled trial; RDD, regression discontinuity design; SEM, structural equation modelling; SNA, social network analysis; SVM, support vector machine; XGBoost, extreme gradient boosting.

Data sources and other: AI, artificial intelligence; API, application programming interface; IPO, initial public offering; MTurk, Amazon Mechanical Turk; n, sample size or number of observations; pp, percentage point; SEC, Securities and Exchange Commission; WoS, Web of Science; Δpp, change in percentage points; χ², chi-square test statistic.

Appendix B: Short Definitions of Methods and Techniques

B.1. Data collection methods

B.2. Analysis techniques

How to cite

This article is the English version of the technical report: Kurutkan, M. N. (2026). Yönetim ve organizasyon alanında en revaçta araştırma yöntemleri ve analiz teknikleri (Technical Report No. 8). Düzce University, Department of Health Management. https://www.researchgate.net/publication/401970333