G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences

In the citing paper

“The estimate was that we needed at least 30 participants per group to account for a 20% dropout rate.”

In the citing paper

“A statistical power analysis was performed using GPower 3.1 [72,73] to estimate sample size for a comparison of differences between two dependent means (matched pairs).”

In the citing paper

“According to G*Power v3.1.97(Faul et al., 2007), a minimum sample size of n = 1145 was needed to detect a small effect size (f 2 = 0.02) for our eight-predictor Comprehensive School Survey (CSS) difference-in-difference (DiD) regression models, based on α = 0.05 and, conservatively, power = 0.95.”

In the citing paper

“At a power equal to 0.80 (two-tailed α = 0.05) and effect size for enjoyment equal to 1.0, a minimum sample size of eight individuals was determined [26] which is similar to a previous study [12].”

In the citing paper

“According to G*Power's calculation, a conservative WilcoxonMann-Whitney two-tailed test requires 90 participants per condition to detect medium-sized effects (Cohen's d = 0.5) at the significance level of 𝛼 = 0.05 with power (1𝛽 ) = 0.90.”

In the citing paper

“The results established a need for nine participants per group (f = 0.25 and actual power = 80.70%) to detect differences with an assumed Type I error of 0.05 and a Type II error rate of 0.20 (statistical power = 80%).”

In the citing paper

“The sample size was calculated using G*Power ver. 3.1.9.7 (Heinrich-Heine-Universität Düsseldorf) with a significance level of α=.05 and statistical power of .80. Assuming an effect size of Cohen’s d=0.85, the minimum required sample size was 22 participants per group [23].”

In the citing paper

“For a two-tail test of the difference between two independent means (two groups), with a 1:1 allocation, a power of 0.80, a significance level of 0.05, and an effect size of 0.80, the minimum required sample size for each group was determined to be 26 individuals.”

In the citing paper

“The sample size was estimated using G-Power 3.1 software with eight predictors, a power of 0.95, and an effect size of 0.15. The smallest number of samples that needed to be analyzed with sufficient power was 160 [74].”

In the citing paper

“Taking into account a power of 0.90, 7 predictors (i.e., the maximum number of predictors included in multiple linear regression), and α = 0.05, a sample of 73 subjects for each group is adequate to detect a minimum effect size of f2 = 0.15, which is considered a medium effect [59].”

In the citing paper

“To detect a medium effect size with three predictors, and .05alpha level, and a .8/80%power which is often considered to be sufficient for a medium effect size (Cohen, 1988), this necessitated 77 participants.”

In the citing paper

“A prior power analysis conducted in G Power at alpha 0.05 specified a minimum target of 120 participants for a two-group comparison with baseline adjustment, with planned over-recruitment to accommodate attrition and incomplete documentation, using conventional small to medium effect size expectations for educational interventions (Cohen 1992;Faul et al. 2009Faul et al. , 2007)).”

In the citing paper

“The sample size was calculated using the G-Power software version 3.1 [64] for a moderate effect size set at 0.5, an expected power of 0.8 and an alpha error probability of 0.05.”

In the citing paper

“Key parameters included α = 0.05, power = 0.80, three groups, and two measurements, with effect sizes denoted by r = 0.39 and f = 0.10. These calculations yielded a required sample size of 297 individuals.”

In the citing paper

“considering an effect size of 0.10, 13 independent variables associated with sleep quality based on the relevant published studies [26–28], a significant level (P) of 0.05, and 95% power, the sample size was calculated to be 287.”

In the citing paper

“To calculate the sample size, we used the G*Power 3.1 software (30), taking into account an alpha of 0.05 and a statistical power of 0.80 on a two-tailed t-test. We need a sample size of 90 participants (45 participants per condition) to reach an effect size of 0.60 for meaning in life.”

In the citing paper

“(using the G∗Power3.1 power analysis program, an a-priori power calculation at 0.8, medium effect size f = 0.25 recommends a sample size of 28; Faul et al., 2007).”

In the citing paper

“To reach an effect of f =0.25 and 80% Power, a two-tailed alpha of 0.05, two groups and three points of measurement where entered and resulted in a total sample size of 154.”

In the citing paper

“Given a maximum of 15 bodily interventions and 15 control interventions per person and assuming a participation rate in daily interventions of 80% (which is a rather conservative estimate, as compared to the previously observed 96% participation rate in an earlier study [51]), we expect on average a total of 24 completed smartphone-triggered interventions per participant. Assuming a correlation of”

In the citing paper

“With an alpha of 0.05, a power of 0.9, and an effect size of f = 0.25, a total sample size of 46 participants would be needed (i.e., 23 per group).”

In the citing paper

“Anticipated sample size was calculated based on recent research evaluating the efficacy of telemedical-based treatment of posttraumatic stress symptoms [20, 21, 38].”

In the citing paper

“To detect an effect size of 0.1, a power calculation with 80% power and an α-error probability of 0.05 resulted in a total sample size of 967 participants.”

In the citing paper

“Sample size has been calculated for a 3 (between) × 4 (within) ANOVA using G*Power software (release 3.1.3) [50]. Setting alpha to 0.05, power to 0.80, and possible dropout, a total sample of 90 is considered enough to detect a small effect size (f = 0.20).”

In the citing paper

“Post hoc sample size calculations on the primary outcome in this study (AAQ-ABI), indicates a total sample size of 58 ( ES Partial η2 = .17, Power = .80), is required to establish a significant interaction effect from pre to post intervention (Faul, Erdfelder, Lang, & Buchner, 2007).”

In the citing paper

“The sample size and power were estimated by A priori or Criterion calculation using G*Power software (Version 3.1.9.7, RRID:SCR_013726) (Faul et al. 2007).”

In the citing paper

“Results of the power analysis indicated a minimum of 22 subjects with an actual power of 0.96.”

In the citing paper

“A minimum sample size of 52 nursing students was calculated based on a significance level of 0.05, an effect size (d) of 0.8, and 80% power [33].”

In the citing paper

“An observed (post-hoc) power analysis using G*Power (Faul et al., 2007) with 𝛼 = .05, observed effect size of d = .369, n = 34 for sexual minorities and n = 157 for heterosexual individuals, indicated that observed power was .49 to detect between-group differences in YBOCS change.”

In the citing paper

“An a priori analysis was conducted using G*power analysis (version 3.1; Faul et al., 2007), which indicated that a sample size of 114 was needed to detect the model's significance at α = .05,power (1 -β error probability) = 0.80, effect size f 2 = 0.15 (medium), with the number of predictors = 9.Therefore, this study's sample size of 292 people was sufficient.”

In the citing paper

“A power analysis using G*Power [51] indicated that a sample of 93 OCD patients would be sufficient to detect within- factors interaction with power > 95 % to test a small-medium effect size (based on our pilot study [49]) with a Type I error α < .05.”

In the citing paper

“A sensitivity analysis (Faul et al., 2007) showed that with alpha set to 0.05 and power set to 0.80, the OCI-R subsample of 83 participants allows detecting the effect of OCD symptom severity on alerting benefits of size η 2 ≥ 0.08, and the effect of OCD symptom severity on the reduction in Simon/Flanker costs on trials that start with a warning cue of size η 2 ≥ 0.025. Smaller effects could be”

In the citing paper

“However, a sensitivity analysis (Faul et al., 2007) showed that with alpha set to 0.05 and power set to 0.80, the matched sample of 80 participants allows detecting the effect of OCD on alerting benefits of size η 2 ≥ 0.09, and the effect of OCD on reductions in Simon/Flanker costs on trials that start with a warning cue of size η 2 ≥ 0.024. Smaller effects could be missed.”

In the citing paper

“We planned to collect data from at least 261 participants.We computed the required sample size using the software G*Power [34] for an ANOVA with main efects and interactions; specifying the default efect size of 0.25, a signifcance threshold of = 0.05 = 0.0045 (i.e., due to testing multiple hy-11 potheses; see Section 4.2.6), a desired power of 0.8, 24 groups, and the respective degrees of freedom”

In the citing paper

“Wedetermined our target sample size (N=432) with an a priori power analysis using the statistical software G*Power [38, 39].”

In the citing paper

“Power analysis with G*Power (v3.1; 109) based on traditional NHST, with alpha=.05, power=.80, 3 groups, and 8 measurements (the 4 tasks described above, with 2 variables from each task in each model) indicates that our N=110 can reliably detect within-group and interaction effects of d=0.20, and between-group effects of d=0.46 or larger. Thus, the study is sufficiently powered to address its prima”

In the citing paper

“According to G*Power analysis [67], at least 172 participants were needed to detect a medium effect size (f = 0.25) for a between-subjects design with power 0.90 (α = 0.05).”

In the citing paper

“A G*Power analysis [67] showed that at least 256 participants were needed to detect a medium effect size (ηp2 = 0.03) for a between-subject design with a power of 0.80 (α = 0.05).”

In the citing paper

“Using an exact two-sided test, a sample of 21 participants (the number for whom we obtained personality data) would be sensitive to effects of ρ = 0.57, assuming 80% power, an alpha level of 0.05, and a null hypothesis correlation of 0.”

In the citing paper

“An a priori power analysis, calculated using G*Power (Faul et al., 2007), was used.A MANCOVA model was used, with between-and within-group factors, to detect a medium-to-large effect, f = 0.30.This effect size was determined based on the calculated effect size in an exercise study examining sCort outcomes with similar experimental design (Hill et al., 2008).Power was preset at 0.8.”

In the citing paper

“Utilising G*Power version 3.1.9.7 [61], with an assumed power of 80%, a type I error probability of 0.05, a two-tailed test, and effect size of r = 0.5 (SPL effect sizes mean), a total sample size of 128 participants is needed to detect a true between-group difference. Including a 20% attrition rate, a total sample size of 160 participants is required.”

In the citing paper

“The sample size for this study was computed using G*Power 3, a statistical power analysis program for social and medical sciences (33). We set the sample size computation based on 10 predictors, an effect size of 0.15, an alpha error probability of 0.05, and a power of 0.95. The computed target sample size was 172 participants.”

In the citing paper

“For the analysis, the software G*Power (version 3.1.6 [54]) was used, and a required sample size of 128 participants was obtained (64 per group).”

In the citing paper

“To determine the required sample size, a power estimation was carried out with G*Power [38]: ANOVA with repeated measures and within-between interactions, 2 groups and 3 measurement times, small expected effect 0.1, desired power of 0.80 (1-β), p = 0.01 (4 outcome variables), correlation among repeated measures 0.7 results in a total sample size of N = 142.”

In the citing paper

“an a priori power analysis with G*Power [181] using repeated measures analyses of variance (ANOVA, within-between-interaction), an assumed power of 0.80, an alpha-level of 0.05, and an expected drop-out rate of 20% [98], revealed a required sample size of 152 participants.”

In the citing paper

“Power analyses were conducted using G-Power [71] to determine the most appropriate sample size for the pilot randomized controlled trial.”

In the citing paper

“Using the G*Power 3 [19] software and previous meta-analyses of the effect size of mindfulness interventions on anxiety [20], we estimated that we would need a sample size of 290 cases for the study.”

In the citing paper

“A post hoc power analysis was conducted using G*Power3 (Faul et al., 2007); with 10 predictors, 87 participants, and a critical α of .05, the study had sufficient power (exceeding .99) to detect a large effect size (f 2 = .60).”

In the citing paper

“The sample size was calculated using G*Power [34] and considering our main statistical analyses (repeated measure ANOVAs with a medium effect size of 0.25, power of 0.80, two groups, and four time measurements), with an addition of an estimated 30% drop-out rate.”

In the citing paper

“A power analysis with the software G*Power [29] resulted in a target sample size of N = 103 to detect a medium-sized effect (f = 0.28) with α = 0.05 and a power of 0.80 for an ANCOVA.”

In the citing paper

“The sample size was determined using G * Power software,version 3.1.9.4; Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Ger-many) [27] and derived from the insertion of the α (two-tailed) (0.05), effect size d (0.80), and power β (0.80) and ultimately calculated as 52 subjects for the paired t-test.”

In the citing paper

“A minimum of 316 university students was deemed necessary to have enough statistical power according to the G-power software (multiple regression; R2 deviation from zero) [52], based on a 5% risk of error, 80% power, f2 = 2.5% and 10 factors to be entered in the multivariable analysis.”

In the citing paper

“An a priori power analysis was performed using G*Power3 [22] to examine the difference between the means of two independent groups on the CERQ scales, with a medium effect size (d = 0.50) and an alpha of 0.05. The results showed that a total sample of 140 was required to achieve a power of 0.80.”

In the citing paper

“Based on previous RCT studies of ACT interventions in HCPs [24, 29], we estimated a small to medium effect size (f = 0.22). The estimated sample size necessary to detect a small to medium effect size at an alpha rate of 0.05 (two-tailed) using the G* Power 3 programme [37], was 104 participants.”

In the citing paper

“Different effect sizes found in the literature based on the transdiagnostic perspective of EDs were considered to estimate the study power in this study. These calculations were performed with the software program G*Power 3.1 [59] and published in the study protocol [58].”

In the citing paper

“To detect a medium to large effect size of d = 0.80, with an alpha level of α = 0.05 and a power of 0.80, G*Power [66] calculated a sample size of N = 70.”

In the citing paper

“The proposed sample size of 50 would leave 22 women per group (assuming 12% attrition) to provide outcome data and yield 80% power for a smallest detectable effect size of .87 (G*Power 3.1.9.4) [24].”

In the citing paper

“The required sample size was computed for an analysis of variance (ANOVA) with 2 groups (between variable; ACT + TAU vs. TAU) × 2 times (within variable; Time 0/Time 1).”

In the citing paper

“Therefore, to capture an effect size 0.5 as statistically significant in a two-tailed test at alpha = 0.05 and a power of (1-Beta) = 0.80, a minimum of 64 participants per Arm will be required, as computed by G*Power 3.0.10 (Faul et al., 2007).”

In the citing paper

“A G*Power analysis (Faul et al., 2007) was conducted with power selected at 0.8, alpha at 0.05 and with an a priori moderate effect size of d = 0.5 (ηp2=0.06) selected (based on the criteria of Cohen (1988)).”

In the citing paper

“Statistical power analysis using G*Power [43] indicates that with an α-level at 0.05 and, a β-level at 0.10, the total sample size in a design like this should be 206 to detect differences in the main and interaction effects of a moderate to medium effect size, f = 0.25, with a one-two-way analysis of variance (ANOVA).”

In the citing paper

“Power analyses with G*Power (Faul et al. 2007) for the detection of small effects (Cohen’s f = 0.12) for the interaction between time (pre, mid-5, mid-15, post) and treatment condition (TAU-M versus TAU-PMR versus TAU) (ANOVA repeated measures, within-between-interaction, α = 0.05, power = 0.80, number of groups = 3, number of measurements = 4, pre-post correlation r = 0.05, nonsphericity correc”

In the citing paper

“To estimate a statistically sufficient sample size for the study, a priori power analysis using G‐Power (Faul et al., 2007) was conducted.”

In the citing paper

“Based on these effect sizes, a power analysis was conducted using G* Power (Faul, Erdfelder, Lang, & Buchner, 2007).”

In the citing paper

“A power analysis in the software program G*Power (Faul et al., 2007) indicated a detectable effect size of r = 0.15 in a sample of 66 participants (power = 0.80, Cronbach ' s α = 0.05).”

In the citing paper

“Based on the effect size observed between NIRS devices during isometric exercise at 50% MVC, a power analysis for a dependant sample t -test was conducted using G*Power to determine a sufficient sample size using an alpha of 0.05, a power of 0.80, a large effect size ( dz ¼ 2.9 ), and two tails. 30”

In the citing paper

“Sample size was calculated (G*Power 3.1, Christian-Albrechts-Universitat, Kiel, Germany) using pain during arm elevation (no load) as the primary outcome, α = 0.05, and power of 80% for between-group analysis [24].”

In the citing paper

“This sample size was chosen on the basis of prior work on the bias blind spot (Mandel et al., 2022) with sensitivity power analyses using G*Power (Faul et al., 2007) revealing that this sample provided at least 80% power to detect a minimum effect size of = .113 for the conducted mixed ANOVAs and d = 0.28 for follow-up independent samples t-tests.”

In the citing paper

“Power analysis was performed using G∗Power (59, 60) with parameters taken from previous reports or similar experiments.”

In the citing paper

“For the tests that we used (Wilcoxon Signed-rank test), we found that the needed sample size was always below the sample size we used for our primary studies, viz., 1000.”

In the citing paper

“due to the complete linkage disequilibrium of L28P with APOE*4 (C112R)”

In the citing paper

“In 1999, two groups independently identified a novel and rare missense mutation in the APOE gene [9, 10] in which the leucine residue is replaced by proline at codon 28 (L28P; rs769452).”

In the citing paper

“A priori power calculations were conducted using G∗Power 3.142 and indicated that a total sample size of 55 would be appropriate for our main behavioural analyses (e.g., linear regressions with 6 predictors) in each tandem group (effect size f2 = 0.15, power = 0.80, α = 0.05).”

In the citing paper

“Power calculation for the sample set was performed on G*Power (v.3.1.9.7) [28]. We performed a “compromise power analysis” for one-way ANOVA (F-tests) for a total sample size of 60 spread equally across 3 groups. The calculations were performed assuming a large effect size (f = 0.4) and the ratio of type II to type I error (β/α) equals 1. The power of the statistical test (1-β) was thus determined”

In the citing paper

“According to Cohen’s sample size recommendation [30] and using G*Power [31] and the 10 times rule of thumb, a minimum of 98 participants were required for this study.”

In the citing paper

“In order to detect such an effect with power 1-β = 0.80, future researchers should use a sample size of at least 202 per group (Faul, Erdfelder, Lang, & Buchner, 2007).”

In the citing paper

“A G*Power analysis, a commonly used software for power analysis in the social and behavioral sciences [17,30], showed that 787 participants were needed to detect a small effect size (Cohen's 𝑓 = .10)with 80% power and 𝛼 = .05.”

In the citing paper

“Our initial sample size was calculated using the G*Power tool (Faul et al., 2007). Based in other effects described in the literature, the effect size considered is 0.8 (the mean difference is 0.8 standard deviations). In these conditions, for power of 0.8 the estimated sample size is 15.”

In the citing paper

“Based on a power analysis using G*Power [61], given α = 0.05, power of 0.80 and effect size of 0.24 (as per a meta-analysis of dual n-back training studies [14]) for a repeated measures, within-between interaction, a minimum of 139 individuals would be required for analysis.”

In the citing paper

“Following study approval from the University of Hertfordshire Life and Medical Sciences Ethics Committee (LMS/SF/UH/00011), and power calculation assessment for sample size (G*power3, Dusseldorf [34]; using α = 0.05; 1 − β = 0.80; based on observed data [9,14])”

In the citing paper

“Finally, the statistical power of each study included in the review was assessed by comparing the sample sizes in these studies against that recommended by a power calculation conducted with G*Power (Faul et al., 2007).”

In the citing paper

“An initial power analysis in G*Power 3.0 (ref.51) targeted on the detection of medium effects (d = 0.5, α = 0.05 and 1 − β = 0.80) suggested a sample size of n = 34.”

In the citing paper

“A power analysis using G*Power software45 indicated that a sample size of 18 would achieve 90% power for a repeated measures design, given a large effect size (partial η2=0.14) and α=0.05.”

In the citing paper

“According to the power analysis conducted with G*Power (Faul et al., 2007), between subjects studies should have at least 64 participants in each group and within subjects studies should have at least 34 participants, to have 80% power to detect a moderate effect.”

In the citing paper

“Sample size was calculated a priori with the statistical software G*Power [49] considering previous studies on complications of pain treatments [30,50].”

In the citing paper

“The sample size of 123 was powered (with power = 0.8, alpha = 0.05, calculated with GPower [40]) to detect the effects of Cohen’s d = 0.45.”

In the citing paper

“Power analysis conducted using G*Power 3 software (Faul, Erdfelder, Lang, & Buchner, 2007) indicated that a sample size of N = 98 with both pretest and posttest data would yield power of .80 to detect a medium effect size (Cohen’s f = .25).”

In the citing paper

“For the main sample, prior to data collection, a power analysis concerning sample size for split-half reliability (bivariate correlation, two tailed) was performed by means of G-power [36], resulting in a suggested sample size of N = 138.”

In the citing paper

“To provide 80% power (5% level of significance) and assuming a medium effect size of 0.25 and a loss to follow-up of 25% after 12 months, the target sample size was 104 participants [9].”