The DOI matches its Crossref record.
The Cost of Dichotomization
Open accessAbstract
Assuming bivariate normality with correlation r, dichotomizing one variable at the mean results in the reduction in variance accounted for to .647r²; and dichotomizing both at the mean, to .405r². These losses, in turn, result in reduction in statistical power equivalent to discarding 38% and 60% of the cases under representative conditions. As dichotomization departs from the mean, the costs in variance accounted for and in power are even larger. Consequences of this practice in measurement applications are considered. These losses may not be quite so large in real data, but since methods are available for making use of all the original scaling information, there is no reason to sustain them.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“However, grouping loses information and may reduce statistical power [ 87 ] especially when dichotomization is used [ 82 , 85 , 88 ].”
From Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration (2007) · cites this paper as [87]
Citations
Cites
No references could be resolved yet.
Cited by
Sorted by citations
Items 801–900 of 945
Designing historical control studies with survival endpoints using exact statistical inference
Han, G. (2020) · Pharmaceutical statistics · doi:10.1002/pst.2050 · Cited by 2
“As the one-sided counterpart of the two-sided LRT in Reference 14, the test (5) is the uniformly most powerful (UMP) test.”
MethodDealing With Artificially Dichotomized Variables in Meta-Analytic Structural Equation Modeling
de Jonge, H., Jak, S., & Kan, K. (2020) · Zeitschrift für Psychologie · doi:10.1027/2151-2604/a000395 · Cited by 5
“However, for the relationship between an artificially dichotomized and continuous variable, the point-biserial correlation does not provide an accurate estimate of the true underlying population correlation (e.g., Cohen, 1983; MacCallum et al., 2002).”
BackgroundResponse interference by central foils is modulated by dimensions of depression and anxiety
Weldon, A. L., Liu, Q., Heller, W., & Buetti, S. (2020) · Attention Perception & Psychophysics · doi:10.3758/s13414-019-01933-y · Cited by 3
Frequency Sensitivity of Neural Responses to English Verb Argument Structure Violations
Sassenhagen, J., Blything, R., Lieven, E., & Ambridge, B. (2018) · doi:10.1525/collabra.87 · Cited by 2
“(Cohen, 1983), it was decided not to include frequency as”
BackgroundDo sales people trust new customers because of who they are?
Gaczek, P., Leszczyński, G., & Zieliński, M. (2018) · The IMP Journal · doi:10.1108/imp-01-2018-0008 · Cited by 2
Sharpley, C. F., Bitsika, V., McMillan, M., Jesulola, E., & Agnew, L. L. (2018) · Neuropsychiatric Disease and Treatment · doi:10.2147/ndt.s168291 · Cited by 3
“24 This limitation might have contributed to the disagreement in the previous literature regarding the existence of the 5HTTLPR–stress–depression association.”
Flora, P., McMahon, C., Locke, S., & Brawley, L. (2018) · Applied Psychology: Health and Well-Being · doi:10.1111/aphw.12106 · Cited by 2
“Evidence”-based medicine in eating disorders research: The problem of “confetti p values”
Kraemer, H. (2017) · International Journal of Eating Disorders · doi:10.1002/eat.22667 · Cited by 2
Détecter la centralité des éléments d’une représentation sociale : un indice alternatif de mesure
Baggio, S., & Renaud, O. (2015) · doi:10.3917/cips.105.0033 · Cited by 2
First, M. B. (2015) · Wiley eBooks · doi:10.1002/9781118753378.ch41
Carretta, T. (2014) · doi:10.21236/ada613694 · Cited by 2
“For the three courses with pass/fail training outcomes, the observed correlations were corrected for range restriction and dichotomization of the criterion (Cohen, 1983).”
BackgroundMethodInfluentialVoisin, D. R., Walsh, T. S., Flatt, N., Eavou, R., Bertozzi-Villa, C., Eder, M., & Schneider, J. A. (2014) · Journal of Nursing Education and Practice · doi:10.5430/jnep.v4n12p27 · Cited by 3
Miwa, K. (2013) · ERA: Education and Research Archive (University of Alberta) · doi:10.7939/r31c1tq15 · Cited by 2
Wand, H., & Ramjee, G. (2012) · Aids and Behavior · doi:10.1007/s10461-011-0118-5 · Cited by 2
Pilot Cognitive Functioning and Training Outcomes
King, R. E., Retzlaff, P. D., Barto, E. L., Ree, M. J., Teachout, M. S., & Carretta, T. R. (2012) · Defense Technical Information Center (DTIC) · doi:10.21236/ada566316 · Cited by 2
Dimensional Analyses of Taxonic Data
Grove, W. (2007) · Psychological Reports · doi:10.2466/pr0.101.2.617-635 · Cited by 2
“This method avoids loss of information resulting from dichotomizing the pi (Cohen, 1983).”
MethodFitzpatrick, D., & Burns, J. (2018) · Campbell Systematic Reviews: Better Evidence for a Better World · doi:10.1002/cl2.194 · Cited by 1
Tracking West Nile virus dynamics using viral loads from trapped mosquitoes
Alahakoon, P., Marchinton, I., Koch, R. T., Fauver, J., & Hay, J. A. (2026) · bioRxiv · doi:10.1101/2025.07.02.662782 · Cited by 1
“RT-qPCR measures the amount of viral material in the pool as cycle threshold (Ct) values, which are inversely correlated with the quantity of viral RNA, though pool status is typically treated as positive or negative resulting in statistical information loss [17, 18].”
Reinterpreting I2 Thresholds: Toward Context‐Specific Heterogeneity Assessment in Evidence Synthesis
Martí-Carvajal, A. J., & Streiner, D. L. (2026) · Cochrane evidence synthesis and methods · doi:10.1002/cesm.70091 · Cited by 1
Douma, E., van den Houdt, S. C. M., Kop, W. J., & Kupper, N. (2026) · Journal of Psychosomatic Research · doi:10.1016/j.jpsychores.2026.112906 · Cited by 1
Beyond behavioral intention: HEXACO-driven market segmentation and trust mechanisms in V2G adoption
Lien, C.-S., Liang, J.-K., Wang, J., & Lu, C.-C. (2026) · Research in Transportation Business & Management · doi:10.1016/j.rtbm.2026.101652 · Cited by 1
Salvador, R., Cordeiro, F., Castuera, R. J., Rebelo‐Gonçalves, R., & Monteiro, D. (2026) · Healthcare · doi:10.3390/healthcare14111445 · Cited by 1
“This methodological approach avoids the loss of variance and the oversimplification associated with dichotomization, allowing for a comprehensive mapping of the perceptual–evaluative gap [44–46].”
Towards a quantifiable measure of orthographic congruence between two languages
Yan, D., Mairano, P., & Casalis, S. (2026) · Behavior Research Methods · doi:10.3758/s13428-026-02981-7 · Cited by 1
“…biased effect sizes, spurious effects, sharp discontinuities in associations, sharp changes in individual differences, inaccurate and invalid A-values, and lack of comparability across studies (Cohen, 1983; Cohen et al., 2003; Harrell, 2015; MacCal-lum et al., 2002; Maxwell & Delaney, 1993) .””
Beyond discretization: why cognitive science should embrace continuity
Pietro, I. D., & Viviani, G. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1777565 · Cited by 2
“Methodologically, discretization reduces statistical power and masks the effect’s functional form (e.g., Cohen, 1983; MacCallum et al., 2002; Young, 2016).”
Hold me tight: Towards the detection of the flow experience using a commercial video game controller
Michailidis, L., & Barcias, J. L. (2026) · Computers in Human Behavior Reports · doi:10.1016/j.chbr.2025.100900 · Cited by 1
Sobrinho, G. M. R., Ferreira, H. G., Mcduff, D. R., & Filgueiras, A. (2026) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph23030362 · Cited by 1
“In this context, the use of continuous symptom measures allows for a more accurate representation of psychological functioning than diagnostic thresholds derived from cut-off points, which are known to reduce statistical power and obscure meaningful variability within and between groups [37,38].”
Blom, E., & van Witteloostuijn, M. (2026) · Journal of Experimental Child Psychology · doi:10.1016/j.jecp.2026.106486 · Cited by 1
Scheiner, R., Burner-Fritsch, I., Fegg, M., Feddersen, B., & Bausewein, C. (2026) · Palliative Care and Social Practice · doi:10.1177/26323524251406617 · Cited by 1
“Another point that weakens the significance of the mean value comparisons is the artificial dichotomisation 87,88 of the setting and the use of a categorised 89 variable for professional experience.”
An exploration of the association between sub bands of gamma and individual symptoms of melancholia
Sharpley, C. F., Evans, I. D., Bitsika, V., & Odierna, G. L. (2026) · Behavioural Brain Research · doi:10.1016/j.bbr.2026.116060 · Cited by 1
Epstein, R., Godoy, S., Wang, N., & Newland, A. (2025) · Scientific Reports · doi:10.1038/s41598-025-30555-8 · Cited by 2
“Correlations for such variables tend to be small 142 .”
Dženopoljac, V., Ognjanović, J., Dženopoljac, A., Kraus, S., Cheng, C.-F., & Puumalainen, K. (2025) · Journal of Knowledge Management · doi:10.1108/jkm-04-2025-0538 · Cited by 1
Malik, A. A., Buang, A. B., Amazu, C. W., Nazir, S., Rusli, R., Abbas, A. N., Leva, M. C., Asgher, U., & Demichela, M. (2025) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2025.2583472
“Literature reports that the median split approach as a cutoff when dichotomizing a continuous psychological variable tends to be effective when the variable follows a normal distribution (Cohen, 1983), a condition met in our case.”
Afsharian, A., Loh, M., & Dollard, M. (2025) · Safety and Health at Work · doi:10.1016/j.shaw.2025.10.005 · Cited by 1
“Acknowledging that dichotomization will lose a certain level of statistical power and information of a construct [39], and there is potential impact of different time intervals [40], we conducted additional statistical testing with a continuous time structural equation modeling (CTSEM) [41].”
Using the Sociocognitive Model of Moral Psychology to Understand Auditor Actions
Thompson, K., & Stefaniak, C. M. (2025) · International Journal of Auditing · doi:10.1111/ijau.70003 · Cited by 1
Considering Approaches to Pervasiveness in the Context of Personality Psychology
Lajoie, D. (2025) · Personality Science · doi:10.1177/27000710251392207 · Cited by 1
“Now, it is well-established that binning is generally a bad idea (e.g., Cohen, 1983).”
Hülshoff, A., Nonte, S., & Reintjes, C. (2025) · Journal of Education for Teaching International Research and Pedagogy · doi:10.1080/02607476.2025.2502825 · Cited by 1
“Furthermore, the post-hoc dichotomisation of variables may result in relevant information loss (Cohen 1983).”
Integrating Advances in Personality Science to Re-Examine the Trait Theory of Voice Disorders.
Welch, B., Helou, L., & Wright, A. G. C. (2025) · Journal of Speech, Language and Hearing Research · doi:10.1044/2025_jslhr-25-00034 · Cited by 1
Keefe, R. (2025) · Schizophrenia bulletin · doi:10.1093/schbul/sbae191 · Cited by 1
“11 Further, such cutoffs need to be pre-specified and are always arbitrary.”
Evaluating the performance of personality-based profiling in predicting physical activity
Katahira, K., Takano, K., Oba, T., & Kimura, K. (2024) · BMC Psychology · doi:10.1186/s40359-024-02268-6 · Cited by 2
“Several studies in psychology and medicine have indicated that loss of information due to dichotomization or categorization [14, 43, 44].”
The threshold for teratogenic effects on child intelligence of prenatal exposure to phenylalanine
Widaman, K., & Helm, J. L. (2024) · Intelligence · doi:10.1016/j.intell.2024.101868 · Cited by 1
Mentoring up: How mentorship training influences mentee intent to change behavior
Orsini, J., Hyun, S., & Sancheznieto, F. (2024) · Journal of Clinical and Translational Science · doi:10.1017/cts.2024.589 · Cited by 1
“However, much of the focus of the statistical community [24,30,31] is on issues with dichotomization.”
BackgroundGiordano, K. A., & Wasserberger, K. W. (2024) · Sports Biomechanics · doi:10.1080/14763141.2024.2377219 · Cited by 1
“Pitfalls of discretising a continuous variable to create categorical subgroups are well documented in other fields (Cohen, 1983; MacCallum et al., 2002; Preacher et al., 2005; Rucker et al., 2015) but are often either overlooked in sports biomechanics research or unknown to sports biomechanics…”
Tiwari, S., & Ambinakudige, S. (2024) · Journal of Hunger & Environmental Nutrition · doi:10.1080/19320248.2024.2355938 · Cited by 1
Optimal dichotomization of bimodal Gaussian mixtures
Jhan, Y.-N., Li, W., Ruan, S.-H., Sie, J., & Lian, I.-B. (2024) · Statistical Papers · doi:10.1007/s00362-023-01521-1 · Cited by 1
“8 2 (Cohen 1983).”
Sharpley, C. F., Evans, I. D., Bitsika, V., Arnold, W. M., Jesulola, E., & Agnew, L. L. (2023) · Symmetry · doi:10.3390/sym15122201 · Cited by 3
“…were used as the indicators of FAA; CRP was derived from serum samples; and depression was measured both as a dichotomous and a continuous variable to extend the comprehensiveness of statistical outcomes mentioned above when simple “depressed” versus “non-depressed” categorization is used [34].”
MethodA mobile digital device proficiency performance test for cognitive clinical research
Andrade, A. C., Bido, D. D. S., Barros, A., Boot, W., & Bertolucci, P. (2023) · arXiv.org · doi:10.48550/arxiv.2310.01774 · Cited by 1
“…influence, but may cause information loss, reduced statistical power and add potential bias, so after this study step it was considered not ideal (Cohen, 1983) as it could lead to potential misrepresentation of the relationships among variables due to the elimination of variability within the…”
BackgroundThe Relationship Between Schizotypal Traits and Affect-driven Volition in Healthy Adults
Lam, J. Y. T., Ng, M. H. F., Valimaki, M. A., & Yee, B. K. (2023) · Journal of Psychopathology and Behavioral Assessment · doi:10.1007/s10862-023-10067-3 · Cited by 1
Validation of a single question to measure internal health locus of control in Swedish primary care
Eriksson, M. C. M., Lindblad, U., Daka, B., & Lundgren, J. (2023) · Scandinavian Journal of Psychology · doi:10.1111/sjop.12923 · Cited by 1
Sampige, R., Tang, D., Frankel, L., & Master, A. (2023) · Merrill-Palmer quarterly · doi:10.1353/mpq.2023.a923223 · Cited by 1
“Increasing the sample size in each group saved statistical power for further statistical analyses when comparing income-level groups (Cohen, 1983).”
BackgroundAboud, Y. A. A., Hussin, M. F. A. @., & Hamdan, M. N. (2023) · International Journal of Academic Research in Economics and Management Sciences · doi:10.6007/ijarems/v12-i1/16081 · Cited by 1
Paige, D. D., Smith, G. S., & Rupley, W. H. (2022) · Literacy Research and Instruction · doi:10.1080/19388071.2022.2138648 · Cited by 1
“Cohen (1983) recommends that when the desire is to create two categorical groups from a normally distributed, categorical variable, dividing at the median tends to produce the best results.”
BackgroundCarpenter, R. K., & Stinson, J. D. (2022) · Partner Abuse · doi:10.1891/pa-2021-0029 · Cited by 1
Jan, P., Erik, V. E., Douglas, G., Peter, C. G., Cynthia, D., Stuart, J., Charles, P., James, J., & Matthias, E. (2021) · doi:10.17816/dd70821-58176 · Cited by 1
“Однако при группировании данных теряется информация, следовательно, может снизиться статистическая мощность (statistical power) [87], особенно, если авторы прибегают к дихотомии [82, 85, 88].”
Chew on this! Oral stereognosis predicts visual word recognition in typical adults
Cummine, J., Huynh, T. K. T., Cullum, A., Ostevik, A. V., & Hodgetts, W. E. (2021) · Current Psychology · doi:10.1007/s12144-021-01647-1
“We chose this approach over a dichotomization of the data based on themean, which has been reported to be equivalent to removing 38% of the data, and substantially reducing power (Cohen, 1983).”
MethodWiedermann, W., Herman, K., Reinke, W., & von Eye, A. (2021) · Development and Psychopathology · doi:10.1017/s0954579421000018 · Cited by 1
“Some authors argue that dichotomization should be avoided due to the tremendous loss of information affecting the validity of subsequent statistical models (Cohen, 1983; DeCoster, Iselin, & Gallucci, 2009; MacCallum, Zhang, Preacher, & Rucker, 2002; Maxwell & Delaney, 1993; Rucker, McShane, & Preacher, 2015), others a…”
BackgroundDavies, S. (2020) · doi:10.17863/cam.47561 · Cited by 1
“As mentioned in Chapter 1, it is theorised that up to one third of information about pattern variation is lost when dichotimising continuous outcomes in this way (Cohen, 1983).”
BackgroundInfluentialA practical generalized propensity-score estimator for quantile continuous treatment effects
Alejo, J., Galvao, A., & Montes-Rojas, G. (2020) · The Stata Journal · doi:10.1177/1536867x20930997 · Cited by 1
“See, for example, Cox (1957); Cohen (1983); van Belle (2008); and Fedorov, Mannino, and Zhang (2009) for more comprehensive discussions on problems associated with discretizing continuous variables.”
Devlin, E. J., Whitford, H. S., Denson, L. A., & Potter, A. E. (2020) · Journal of Psychosomatic Research · doi:10.1016/j.jpsychores.2020.109979 · Cited by 2
Shin, H., & Youn, N. (2020) · Psychology of Aesthetics Creativity and the Arts · doi:10.1037/aca0000303 · Cited by 1
Generality of likelihood ratio decisions
Glanzer, M., Hilford, A., Kim, K., & Maloney, L. T. (2019) · Cognition · doi:10.1016/j.cognition.2019.03.023 · Cited by 3
Análise das Propriedades Psicométricas do Questionário de Monitoramento Parental
Cassoni, C., Elias, L. C. D. S., Marturano, E. M., & Fontaine, A. (2019) · Revista Avaliação Psicológica · doi:10.15689/ap.2019.1803.16277.04 · Cited by 1
“Na avaliação da magnitude das diferenças, considerou-se o efeito como pequeno quando η2p≤0,05; médio, quando η2p variou de 0,051 a 0,25; elevado para η2p entre 0,26 e 0,50; muito elevado quando η2p>0,50 (Cohen, 1992; Marôco, 2014).”
BackgroundChen, T.-Y., & Sie, Y.-S. (2019) · doi:10.7939/r3-sb0h-jj73 · Cited by 1
Hammond, M. S. (2017) · Journal of The First-Year Experience & Students in Transition · doi:10.65005/154230717822029449 · Cited by 1
Statistical Issues in the Evaluation of Clustering of Metabolic Syndrome in Spousal Pairs
K, S. M. C. (2015) · Journal of Biometrics & Biostatistics · doi:10.4172/2155-6180.1000233 · Cited by 2
Generalized linear models with coarsened covariates: a practical Bayesian approach.
Johnson, T. R., & Wiest, M. (2014) · Psychological methods · doi:10.1037/a0034274 · Cited by 1
“…covariate, and the failure to detect important relationships between the response variable and the uncoarsened covariate such as nonlinearity (e.g., Cohen, 1983; Irwin & McClelland, 2003; MacCallum, Zhang, Preacher, & Rucker, 2002; Maxwell & Delaney, 1993; Vargha, Rudas, Delaney, & Maxwell, 1996).”
BackgroundPerformance implications of entrepreneurial stylistic fit: evidence from Australia
Najmaei, A. (2014) · International Journal of Business Competition and Growth · doi:10.1504/ijbcg.2014.060315 · Cited by 1
DongMinKim, & 양, 정. (2012) · Korea Journal of Counseling · doi:10.15703/kjc.13.6.201212.2523 · Cited by 1
Competition and Personality in a Restaurant Entry Game
Bergstrom, T., Sonstelie, J. C., & Parendo, S. (2011) · doi:10.1515/bejte-2015-0049 · Cited by 1
“Cohen (1983) showed that forced dichotomization of a continuous variable results in a loss of statistical power equivalent to throwing away 38–60% of the data. McCallum et al. (2002) conclude that dichotomization of quantitativemeasures in psychological tests is “rarely defensible” and often yields misleading results,…”
BackgroundInfluentialEvaluation bias against work in progress: evidence from a Spanish marketing conference
Calderón-Martínez, A., Ruiz-Conde, E., Parreño-Selva, J., & Ruiz, F. M. (2010) · Higher Education · doi:10.1007/s10734-009-9306-x · Cited by 1
“In this way we avoid two important disadvantages (Hittner 2004) of the Sharma et al. (1981) procedure: (1) reduction in the statistical power as a consequence of codifying a continuous variable into categories ( Cohen 1983 ); and (2) unclear comparability among the resulting groups, created on criteria such as the mea…”
MethodTutz, G. (2007) · doi:10.1002/9781118445112.stat06939 · Cited by 1
Evidence for unconscious priming using a Bayesian single-subject approach
Sánchez-Fuenzalida, N., van Gaal, S., van den Hurk, Z., Stein, T., & Fahrenfort, J. J. (2026) · Neuroscience of Consciousness · doi:10.1093/nc/niag052
“Note that dichotomizing continuous measures inevitably entails some information loss ( Cohen 1983), a critique that has been raised over the median-split technique in priming literature ( Zerweck et al. 2021, Meyen et al. 2022, Stein et al. 2024, Hernández-Gutiérrez et al. 2025).”
Fluctuation of Flow and Affect in Everyday Life: A Second Look at the Paradox of Work
Engeser, S., & Baumann, N. (2014) · Journal of Happiness Studies · doi:10.1007/s10902-014-9586-4
“Continuous data also have the methodological advantage of taking all information into consideration and not reducing variance by splitting the data into ‘‘non-flow’’ and ‘‘flow’’ (Cohen 1983; Cohen et al. 2003).”
BackgroundPathways to Civic Engagement with Big Social Issues: An Integrated Approach
Skarmeas, D., Leonidou, C. N., Saridakis, C., & Musarra, G. (2019) · Journal of Business Ethics · doi:10.1007/s10551-019-04276-8 · Cited by 37
“Nonetheless, the unfortunate practice of reducing continuous variables to categories results in less than full explanatory power of measures, due to discarded information related to people’s social preferences and unnecessary sacrifice of statistical power (see Cohen, 1983).”
BackgroundComparing Nouns and Verbs in a Lexical Task
Cordier, F., Croizet, J., & Rigalleau, F. (2012) · Journal of Psycholinguistic Research · doi:10.1007/s10936-012-9202-x
Student Participation in Academic Advising: Propensity, Behavior, Attribution and Satisfaction
Braun, J., & Zolfagharian, M. (2016) · Research in Higher Education · doi:10.1007/s11162-016-9414-2
Li, F., Chao, M. C., Chen, Y., & Zhang, S. (2017) · Asia Pacific Journal of Management · doi:10.1007/s10490-017-9529-9 · Cited by 25
Log-linear models to evaluate direction of effect in binary variables
Wiedermann, W., & von Eye, A. (2017) · Statistical Papers · doi:10.1007/s00362-017-0936-2 · Cited by 19
Duan, W., Fei, Y., & Tang, X. (2019) · Child Indicators Research · doi:10.1007/s12187-019-09637-3 · Cited by 14
Two advanced methods for adjusting the main coefficient in logistic regression
Yang, Y.-W., & Fu, C.-Y. (2011) · Computational statistics (Zeitschrift) · doi:10.1007/s00180-011-0294-9
The association in two-way ordinal contingency tables through global odds ratios
Camminatiello, I., D'Ambra, A., & D'Ambra, L. (2021) · Metron · doi:10.1007/s40300-021-00224-7
Karcıoğlu, A. M., KAYA, S. K., Atlı, H. Z., Acar, E., & Gunes, A. K. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1955731
Loss of power by degree of heaping in covariables
Becher, M., Sauerbrei, W., & Becher, H. (2026) · BMC Medical Research Methodology · doi:10.1186/s12874-026-03043-4
“Although several critical issues are known for decades [16, 18], dichotomization of continuous variables is still very popular in many areas of research.”
Ankit, J. H. A., & Singh, V. (2026) · Brain Structure and Function · doi:10.1007/s00429-026-03201-w
“While dichotomization is known to reduce statistical power and discard individual-level variability (Cohen 1983), this approach was necessary due to significantly higher depression and anxiety scores in women and to avoid confounding sex and comorbidity status.”
Menéndez, F. M. M., Vittes, J. F. V. Y. R., Tapia, N. G. S., Espinoza, J. A. C., Barzola-Capcha, E. B., Infantes, M. S., Canchan-Casas, M. R., González-Prida, V., & Zacarías-Vallejos, S. R. (2026) · Journal of risk and financial management · doi:10.3390/jrfm19100779
Gu, H.-L., Hu, Y.-H., Zhong, Q., Cheng, Y.-F., & Fang, J. (2026) · Applied Psychology: Health and Well-Being · doi:10.1111/aphw.70242
Garneau, J., Perrault, A., Borgetto, F., Smith, D., Cross, N., Gong, K., Maltezos, A., Tarelli, L., Savard, J., Schwartz, S., et al. (2026) · Sleep Medicine · doi:10.1016/j.sleep.2026.109336
Shingirayi, B. S., & Matos, P. V. (2026) · Nonprofit Management & Leadership · doi:10.1002/nml.70071
Fairchild, A. J., Gupta‐Kagan, J., & Barclay, A. (2026) · Children and Youth Services Review · doi:10.1016/j.childyouth.2026.109120
Wan, J.-M., Zhang, X.-K., Jin, Y.-M., Wang, J.-X., & Lei, Y. (2026) · Biological Psychology · doi:10.1016/j.biopsycho.2026.109404
Mendez, I., Singh, R., Jackson, C., Hinerman, A. S., Rodriquez, E., Martinez, E., Camargo, J., & Pérez-Stable, E. (2026) · Journal of Racial and Ethnic Health Disparities · doi:10.1007/s40615-026-03100-0
Mialhe, F., & Rebustini, F. (2026) · Evaluation & the Health Professions · doi:10.1177/01632787261479346
Barragán-Pazmiño, B., & Pazmiño-Maji, R. (2026) · Multidisciplinary Science Journal · doi:10.31893/multiscience.2027228
“Finally, a methodological limitation of this study is the dichotomization of ordinal and continuous variables, which entails a loss of information and reduced statistical power (Cohen, 1983).”
Tampubolon, G. (2026) · Frontiers in Public Health · doi:10.3389/fpubh.2026.1879570
“Already in 1983 Cohen wrote, “since methods are available for making use of all the original scale information there is no reason to sustain [dichotomizing practice]” (34).”
Virginia parents’ knowledge of juvenile sex offense registration laws
Freelin, B. N., & Mikytuck, A. M. (2026) · Journal of Crime and Justice · doi:10.1080/0735648x.2026.2712955
Kim, M. H., Kim, K., & Park, J. (2026) · European Journal of Innovation Management · doi:10.1108/ejim-11-2025-1607
Polishing the meta-problem of consciousness: empirical evidence for non-English speakers
Celka, K., Wyrwa, M., & Raś, M. (2026) · Philosophical Psychology · doi:10.1080/09515089.2026.2710862
Garg, R., Glasmacher, K. A., Augert, A., & Townsend, J. P. (2026) · Frontiers in Genetics · doi:10.3389/fgene.2026.1776578
“Indeed, applying a binary threshold to a naturally continuous parameter is often arbitrary and problematic (Cohen, 1983; Giannoni et al., 2014).”
Dimakopoulou, E., Ntomali, S., Kaloupsis, S., Karydaki, M., & Bampouras, T. M. (2026) · BMC Sports Science, Medicine and Rehabilitation · doi:10.1186/s13102-026-01875-7
“Therefore, and although some loss of power is inevitable due to the classification of continuous scores into categories, it is not as pronounced as dichotomisation (reported to lose ~ 40% power; [4], while it enables the analysis of state anxiety in relation to the orientation dominance.”
Waki, R., Sato, K., Inoue, J., Kodama, K., Yamada, M., & Higuchi, T. (2026) · Frontiers in Virtual Reality · doi:10.3389/frvir.2026.1855334
From significance to substance: making moderation meaningful in marketing research
To, L. (2026) · Italian Journal of Marketing · doi:10.1007/s43039-026-00140-6
“Cohen ( 1983) showed that dichotomizing continuous variables substantially reduces statistical power, with mean-split dichotomization lowering effective sample size by about 30–40%, and double dichotomization reducing it by up to two-thirds.”
Metadata
- Authors
- Cohen, J.
- Affiliations
- New York University (US)
- Corresponding author
- Cohen, Jacob
- Venue
- Applied Psychological Measurement
- Publisher
- SAGE Publications
- Published
- 1983-06
- Volume
- 7
- Issue
- 3
- Pages
- 249-253
- ISSN
- 0146-6216 (ISSN-L)
- 1552-3497
- OpenAlex
- W1987838173
- Cited by
- 1,564
- References
- 0
- Citations by year
- 2026: 67 · 2025: 59 · 2024: 59 · 2023: 50 · 2022: 63 · 2021: 66 · 2020: 74 · 2019: 68 · 2018: 91 · 2017: 78 · 2016: 66 · 2015: 79 · 2014: 86 · 2013: 53 · 2012: 60
- FWCI
- 3.58
- Type
- journal article
- Language
- en
- Open access
- green
- purl.umn.edu
- Locations
- Applied Psychological Measurement (publishedVersion)
- University of Minnesota Digital Conservancy (University of Minnesota) (submittedVersion)
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Applications
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Keywords
- dichotomization, variable reduction, bivariate normal distribution, correlation, variance accounted for, statistical power, measurement loss, scale reduction
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review