ArticlePDF Available

A Power Primer

Authors:

Abstract

One possible reason for the continued neglect of statistical power analysis in research in the behavioral sciences is the inaccessibility of or difficulty with the standard material. A convenient, although not comprehensive, presentation of required sample sizes is provided. Effect-size indexes and conventional values for these are given for operationally defined small, medium, and large effects. The sample sizes necessary for .80 power to detect effects at these levels are tabled for 8 standard statistical tests: (1) the difference between independent means, (2) the significance of a product-moment correlation, (3) the difference between independent rs, (4) the sign test, (5) the difference between independent proportions, (6) chi-square tests for goodness of fit and contingency tables, (7) 1-way analysis of variance (ANOVA), and (8) the significance of a multiple or multiple partial correlation.
TutorialsinQuantitativeMethodsforPsychology
2007,Vol.3(2),p.79.
Apowerprimer
JacobCohen
NewYorkUniversity
Onepossiblereasonforthecontinuedneglectofstatisticalpower
analysisinresearchinthebehavioralsciencesistheinaccessibilityofor
difficultywiththestandardmaterial.Aconvenient,althoughnot
comprehensive,presentationofrequiredsamplesizesisprovided.
Effectsizeindexesandconventionalvaluesforthesearegivenfor
operationallydefinedsmall,medium,andlargeeffects.Thesample
sizesnecessaryfor.80powertodetecteffectsattheselevelsaretabled
for8standardstatisticaltests:(1)thedifferencebetweenindependent
means,(2)thesignificanceofaproductmomentcorrelation,(3)the
differencebetweenindependentrs,(4)thesigntest,(5)thedifference
betweenindependentproportions,(6)chisquaretestsforgoodnessof
fitandcontingencytables,(7)1wayanalysisofvariance(ANOVA),
and(8)thesignificanceofamultipleormultiplepartialcorrelation.
FromPsychologicalBulletin(1992),vol112(1),p.155159.Copyright©1992bytheAmericanPsychological
Association. Reproducedwithpermission. Forinformationonhowtoobtainthefulltexttothisarticle,
pleasevisithttp://www.apa.org/psycarticles.
79
... Additionally, equivalence testing for a O 2 p of .01-which would be a very small effect, according to Cohen [58]-was run for the Further, the influence of the application order was tested in a linear mixed model with the First Application Displayed as a fixed effect and the Environment and Assessment Time as random slopes for the raw SSQ total scores. This analysis is a diversification of the first prerequisite hypothesis, as the between-subjects factor First Application Displayed was added. ...
Article
Full-text available
Although Virtual Reality (VR) holds massive potential, its applicability still faces challenges because some individuals experience cybersickness. This phenomenon includes general discomfort, disorientation, and/or nausea, and it threatens not only a pleasant user experience but also the user’s safety. Thus, predicting a user’s susceptibility without relying on screening questionnaires that focus on past experiences, would enable more pleasant, safer VR experiences, especially for first-time users. Hence, the current study uses the participant’s controller input in a virtual Rod and Frame Test (RFT) as an effortlessly trackable performance measure. The RFT is an established method for measuring an individual’s sense of verticality in visually displaced fields. It has been used in the context of simulator sickness and cybersickness. In line with the literature and the subjective vertical mismatch theory, a lower visual dependency is expected to be correlated positively with cybersickness. To evaluate the potential of the RFT as a screening method for cybersickness, a cybersickness-inducing virtual environment (the City) was deployed. In total, data from 76 participants were eligible for the statistical analysis. The study finds a positive correlation between lower visual dependency and cybersickness, but only for the group that took the RFT after experiencing the City and only for the post-RFT cybersickness ratings. As cybersickness symptoms were VR environment-specific, the predictive validity of the RFT considering the VR-specific attributes is limited. Further, other studies attributed different working mechanisms to explain the connection between visual dependence and cybersickness with conflicting evidence. Although the RFT is not applicable as a cybersickness screening method, the effect sizes suggest that the RFT could serve as an additional objective assessment of the individuals’ current state during VR exposure. Future research should systematically explore interconnections between the various factors that contribute to cybersickness, pursuing the idea of open science for context sensitivity.
... In the literature, an effect size above 0.5 is interpreted as large. 27 All items and the scale's total score had a significant positive and large effect size between the two measurements. The overall scale's correlation coefficient was calculated as 0.952. ...
Article
Full-text available
Objective. Management of type 1 diabetes (T1DM) is quite challenging for both adolescents and their families. In this study, we aimed to translate the 14-item Problem Areas in Diabetes-Teen (PAID-T) scale, which measures variables that influence diabetes distress, to Turkish and investigate the Turkish version’s reliability and validity. Methods. One hundred and ninety-four adolescents with T1DM participated in the study. PAID-T and forms for sociodemographic and diabetes characteristics were used for data collection. The scale’s content validity was checked using the Davis technique. Cronbach’s α was used to analyze the scale’s internal reliability and the test-retest for the scale’s reliability. Exploratory factor analysis (EFA) was utilized to examine the factor structure. The fit of the scale was assessed using confirmatory factor analysis (CFA). Results. Of the participants, 54.6% (n=106) were girls. The content validity index values of the scale items ranged between 0.86 and 1.0. The PAID-T scores of girls and boys were similar. No significant difference was found between PAID-T scores with sociodemographic data and diabetes characteristics (p>0.05). The test-retest correlation coefficient of the scale was found to be 0.952. The three-factor (emotional burden, family and friend distress, and regimen-specific distress) model identified in EFA explained 61.8% of the common variance. Fit analysis was performed using CFA for the three-factor model, which did not show adequate fit (x2/df = 2.402, GFI = 0.822, CFI = 0.815, NFI = 0.727, NNFI = 0.772, RMSEA = 0.118). The Cronbach α value of the scale was 0.864. Conclusion. The Turkish version of the 14-item PAID-T showed moderate validity and strong reliability. Accordingly, it can be used as a reliable measurement tool to assess diabetes stress in adolescents with T1DM.
Article
Purpose This study aims to examine the role of employee experience in influencing employee well-being and turnover intentions within organizations. The mediating role of well-being will also be investigated, along with an exploration of whether these relationships differ across genders, specifically in the Indian corporate context. Design/methodology/approach A descriptive, quantitative study was conducted using structured questionnaires to gather data from 111 employees in the Indian corporate sector. The study used a non-probability judgment sampling method. Data was analyzed through SPSS for descriptive and inferential statistics, and partial least squares was used to explore mediation and model fit. Findings The study found a significant impact of employee experience on well-being, as well as a negative correlation between both employee experience and turnover intention and well-being and turnover intention. Well-being was found to partially mediate the relationship between employee experience and turnover intention. Gender-based analysis revealed no significant differences in the relationships between these variables for men and women. Originality/value This research highlights the universal applicability of employee experience as a predictor of well-being and turnover intention, irrespective of gender. By establishing that gender does not moderate these relationships, this study provides new insights challenging traditional assumptions about gender disparities in workplace outcomes.
Chapter
The concept of imagined communities has been overlooked by researchers in second language acquisition in the Iranian context. Actually, imagined communities can establish a policy framework for the consideration of desire, hope, and creativity in identity construction in English as a foreign language (EFL) settings. The present study, following a mixed-methods design, mainly aimed at: (1) examining the association between the students’ imagined communities and engagement in writing tasks and (2) exploring the students’ perspectives of the role of imagined communities in elevating their engagement to learn English. In doing so, a number of 112 homogeneous English-major students participated in the quantitative phase of the study based on convenient sampling, and a pool of six students were participated in the qualitative phase of the study based on purposive sampling. A number of instruments were used to measure the imagined community and engagement in writing tasks. The results of Pearson product-moment correlation showed that there was a large positive association between imagined communities and engagement in writing tasks. Having measured the intercoder reliability and intercoder agreement, the results emerged from the content analysis of the students’ responses revealed 12 common codes. Finally, the study offers some practical implications for EFL students and teachers.
Article
Full-text available
Science education bears the broader objective of nurturing students today to be scientifically-literate citizens of tomorrow who are able to foresee challenges, invent solutions and make responsible decisions for global issues. As a prelude to the new focus of agency in the Anthropocene, this paper presents an intervention on climate change with upper secondary students in a museum of natural history in England. Instructional strategies such as infusing scenarios and arts into scientific discussions were adopted to induce imagination, future-oriented thinking and emotional responses. Statistical results showed that the intervention significantly enhanced participants’ futures literacy, environmental agency and positive emotions. However, it did not increase their interests in learning science in out-of-school context. Implications of this study will shed light on futurising science and climate education in research and practice.
Article
Full-text available
Early childhood is a pivotal period for developing environmental awareness and sustainable behaviors, During this age period, social learning theory contributes to understanding how young children form behavioral patterns. The aim of this study was to evaluate the effectiveness of the Ecological Footprint Awareness Program based on social cognitive learning theory in 60-72-month-old children. This study was conducted using a cluster randomized controlled pretest-posttest experimental design. It was carried out in four preschools located in a city center between April and June 2023. Two of the preschools were assigned to the intervention group while the other two were assigned to the control group. Data was collected using the Ecological Footprint Awareness Scale for Children (EFAS-C). The Social learning theory based on Ecological Footprint Awareness Program was carried out with the children in the intervention group for a period of six weeks, with one session of 40 min per week. The program covered waste management, water and energy use, food consumption, and transportation, with one topic each week. The data was analyzed using descriptive statistics, paired groups t-test, and independent groups t-test, with a 95% confidence interval and p < .05 significance level. At the end of the study, it was found that the ecological footprint awareness of the children who participated in the training program was higher than those who did not participate in the training program. This randomized controlled trial provides strong evidence for the impact of environmental education programs based on social cognitive learning on young children’s understanding and actions regarding their ecological footprint.
Article
Full-text available
Theory-based physical activity (PA) interventions include PA promotion strategies that can be delivered by exercise professionals, friends, family and peers. Peer-delivery presents a valuable opportunity for community implementation. Few peer-led PA interventions for people living with and beyond cancer (LWBC) report the feasibility of their peer mentor training methods. The purpose of this study was to assess the feasibility and acceptability of a peer mentor training program to deliver a behavioural PA intervention to inactive people LWBC using a mixed methods approach. Peer mentors (active people LWBC [≥90 min/week of PA]) participated in an online training program. Weeks 1 to 4 (Phase I) included knowledge and skill development (1-hour online module and 2-hour live workshop weekly). The Assessment phase (Phase II) explored peer mentor readiness (≥80% on a knowledge quiz and ≥3/5 points [Satisfactory] on a mock role play). Feasibility was assessed using enrollment rates, retention rates, adherence, and semi-structured interviews. Acceptability was measured using a satisfaction questionnaire assessing level of agreement with several statements about training program components. Peer mentors ( N = 14; Mage = 65.4 ± 10.7 years) were diagnosed with primarily prostate (57.1%) or breast (21.4%) cancer. Enrollment and retention rates were 73.7% and 92.9%, respectively. Workshops and online modules had 100% and 87.5% adherence rates, respectively. Majority of peer mentors met readiness criteria for the knowledge quiz (92.3%) and mock role play (84.6%) on their first attempt, with 92.3% delivering the follow-up peer-led PA intervention. Peer mentor satisfaction scores ranged from 3.9 to 4.6 out of 5. Interviews generated themes around overall impressions, feedback on timing, structure, and content of the training program and mock role play, and peer mentor preparedness. Structured training for delivering peer-led PA interventions show promise; however, individualized support may be needed for some people LWBC to strengthen mentorship knowledge and skills.
Article
Check-in/Check-out (CICO) behavioral support has been implemented in Finnish School-Wide Positive Behavior Interventions and Support (SWPBIS) schools to cater to students who require personalized behavior support beyond the universal level. Previous studies have demonstrated the effectiveness of CICO as a behavioral support method. However, further research is needed to investigate its effectiveness in a larger sample and to analyze the timeline for behavioral change. This study focused on 51 elementary school students, assessing their behavior at baseline and during the CICO intervention phase using two data collection methods: the Daily Report Card (DRC) and the School Situation Questionnaire (SSQ). Nonlinear growth modeling was employed to examine the effects of the intervention. The results indicated that CICO yielded significant positive effects on behavior within 1 week of initiating support. After the outcomes stabilized, the behavior change remained stable beyond the first week of the intervention. These effects were detected in both the target behavior measured with DRC and the problem behavior measured with SSQ. These findings suggest that CICO interventions produce rapid and sustained changes in behavior. Further, the effects of CICO were observed in various settings within the school environment, indicating distal outcomes.
Article
Full-text available
The long-term impact of studies of statistical power is investigated using J. Cohen's (1962) pioneering work as an example. We argue that the impact is nil; the power of studies in the same journal that Cohen reviewed (now the Journal of Abnormal Psychology) has not increased over the past 24 years. In 1960 the median power (i.e., the probability that a significant result will be obtained if there is a true effect) was .46 for a medium size effect, whereas in 1984 it was only .37. The decline of power is a result of alpha-adjusted procedures. Low power seems to go unnoticed: only 2 out of 64 experiments mentioned power, and it was never estimated. Nonsignificance was generally interpreted as confirmation of the null hypothesis (if this was the research hypothesis), although the median power was as low as .25 in these cases. We discuss reasons for the ongoing neglect of power. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Full-text available
This is an account of what I have learned (so far) about the application of statistics to psychology and the other sociobiomedical sciences. It includes the principles "less is more" (fewer variables, more highly targeted issues, sharp rounding off), "simple is better" (graphic representation, unit weighting for linear com- posites), and "some things you learn aren't so." I have learned to avoid the many misconceptions that surround Fisherian null hypothesis testing. I have also learned the importance of power analysis and the determination of just how big (rather than how statistically significant) are the effects that we study. Finally, I have learned that there is no royal road to statistical induction, that the informed judgment of the investigator is the crucial element in the interpretation of data, and that things take time.
Article
Several MS/PC—DOS programs are now available to help with statistical power analysis and sample-size choice. This article compares these programs with respect to the statistical methods they cover, their user interface, their ease of use, their graphics capabilities, and their computational accuracy.
Article
The problem of testing statistical hypotheses is an old one. Its origins are usually connected with the name of Thomas Bayes, who gave the well-known theorem on the probabilities a posteriori of the possible “causes” of a given event.* Since then it has been discussed by many writers of whom we shall here mention two only, Bertrand† and Borel,‡ whose differing views serve well to illustrate the point from which we shall approach the subject.