Daily undulating periodization (DUP) applied to resistance training is characterized by non-linear and non-sequential training of strength, hypertrophy and power. Rather than following a linear pattern of organised blocks of training, each with a specific focus, DUP undulates the focus of training daily throughout a given microcycle. The athlete is then required to either train hypertrophy, strength and power in a controlled and fixed sequence within the week (DUP) or in an order the athlete and coach feel appropriate on a given day, namely flexible DUP (FDUP). Colquhoun et al. (2017) compared DUP and FDUP over a 9-week training period. 25 resistance trained men were randomly assigned to the DUP or FDUP group. All participants were required to have a minimum of 6 months training experience. No significant differences in training intensity or volume were found between groups. Both groups improved on all training variables (1RM bench press, squat and deadlift). No differences in questionnaire assessed motivation to train, session RPE or training satisfaction were evident between groups. Authors concluded that both training methods offered a similar training effect.
Looking closely at the data generated from this study the baseline average strength of the FDUP group was lower on all three lifts (squat, bench press and deadlift), although p-values were above 0.05 in all group x group comparisons. P-values were not reported specifically but were summarised as a table legend indicating this value being below 0.001 or below 0.05. Referring to basic principles of frequentist statistical interpretation, the p-value indicates the closeness of fit of the actual data to a data model based on the hypotheses being assessed and the statistical test being conducted. The authors have correctly interpreted the lack of closeness of the data and model as cause to reject the alternative hypothesis (difference between groups). Subsequently, the authors have accepted the null hypothesis (no difference between groups). Another interpretation of this analysis could conclude that the levels of noise in the data (e.g. the coefficients of variation) were too high for significance to be reached. In other words, the statistical test ‘could not see the wood for the trees’. An awareness of the degree of closeness of the actual and modelled data (actual p-values), coupled with professional and critical appraisal of the full range of evidence could have resulted in a more critical and valid concluding remark. This is important as based on this individual study, practitioners may assume that both DUP and FDUP are equally effective. However, given a more critical analysis of the data an alternative and more reasonable conclusion could have been reached.
I would recommend tighter controls, practically resolved through more rigorous and specific selection criteria. The authors stated that all participants had more than 6-months training experience. More specific detail is needed to ascertain the nature of the groups training experience (e.g. homogenous? or heterogenous?). I would speculate that if the group members’ amount of experience was highly variable this would have affected the initial strength levels which were evidently dissimilar, indicating differing degrees of training adaptation and accommodation to training stressors. If one group were effectively ‘hard gainers’ due to being further along in their training history, they would experience a greater level of exercise stressor accommodation and result in a slower rate of performance/strength improvement. Going back to my original statement the average baseline 1RMs were lower in the FDUP group. Coupled with the large amount of noise and variance in the data (COV for FDUP bench press was 20.98 at baseline and 18.38 post training; COV for DUP bench press was 17.63 at baseline and 16.71 post training), this variance may explain the non-significance in the statistical test reported by the authors.
Subsequently, my own interpretation of this study leads to a call for further research utilising tighter controls with a more homogenous group (e.g. well trained and/or more experienced athletes). Alternatively, a more selective group assignment with an available pool of participants may enable tighter levels of variance in the data resulting in clearer distinctions between two very different mean values (although not “significant”).
Table 1: Coefficients of variation for 1RM tests performed at baseline and post 9-week training period
Colquhoun, R. J., Gai, C. M., Walters, J., Brannon, A. R., Kilpatrick, M. W., D’Agostino, D. P., & Campbell, W. I. (2017). Comparison of Powerlifting Performance in Trained Men Using Traditional and Flexible Daily Undulating Periodization. The Journal of Strength & Conditioning Research, 31(2), 283-291. doi:10.1519/jsc.0000000000001500