Original article written by Dylan Suarez, Kevin Carroll, Jake Slaton and colleagues.
Background
The ability for athletes to produce force underpins virtually all actions in sport [1]. Further to this, our ability to produce force rapidly is arguably more important, especially for sports where common or critical movement patterns are constrained by time [2]. Such examples of this include: maximal velocity sprinting (105 ms), curvilinear sprints (150 ms), acceleration (190 ms), and fast stretch-shortening cycle plyometrics (< 250 ms) [3]. When assessing maximal force production capability and the rate at which we can produce that force, multi-joint isometric testing has now become commonplace, with many practitioners and researchers utilising the isometric mid-thigh pull [IMTP] [4,5] and isometric squat [6,7]. Whilst the assessment of peak force is both expected and common during these tests, they are often used for the assessment of rapid force generation capabilities as well, via metrics such as: rate of force development (RFD) and impulse, over pre-defined time periods on the force-time curve. Given that most sporting actions are both underpinned by: a) force production, and b) time constraints to perform the task, monitoring rapid force production capabilities seems like a useful exercise for practitioners.
Despite this argument, previous research has outlined that metrics such as RFD often exhibit very high variability – i.e., coefficient of variation (CV) values ~50% [4,7], bringing in to question whether such data can actually be used in day-to-day practice. The present study argues that such differences in reliability may be due to factors such as: a) the position adopted during testing, b) method of analysis, and c) how familiar athletes are with the test. In addition, the authors argue that the majority of studies have been set up to assess for “peak force”, whilst actually concurrently measuring multiple early force-time variables such as RFD and impulse, as well. This seems like a fair point, given that previous research has indicated that it may be challenging to obtain true maximal force and RFD measurements, within any given trial – as one focuses on pure force production, whilst the other focuses on how fast that force can be produced [8].
Thus, the primary aim of the present study was to determine the difference between early force-time characteristics of a new shortened, 1-second IMTP procedure (SHORT) vs. a traditional, 3-second procedure (TRAD). In addition, the authors employed a test-retest design, so within and between-session reliability was also examined. Finally, given recent literature has highlighted the importance of providing a ‘biological basis’ for using a test (i.e., is it associated with independent measures of performance for a given sport or athlete) [9], the authors also determined the relationships between early force-time characteristics and vertical jump ability, for both protocols.
What the authors did
Participants (n = 14 males with an average of 8.5 years resistance training experience) volunteered for this study and were asked to attend 3 sessions. During the first, they were provided with an explanation of procedures and a chance to familiarise themselves with test protocols. The following day (session 2), consisted of a hydration test (to determine whether poorly hydrated individuals might have their performance impacted), a standardised warm-up, both SHORT and TRAD protocols, and loaded and unloaded squat jumps. Session 3 took place one week later and was a copy of session 2.
To ensure that both IMTP protocols were provided with a comparable set of instructions (and noting that the SHORT protocol was trying to determine whether it was a better reflection of early / rapid force-time characteristics), participants were instructed to “pull as fast and hard as possible”. The selected metrics for both protocols included: a) peak force, b) force at 90 and 200 ms, c) RFD at 90 and 200 ms, and d) impulse at 90 and 200 ms. These time frames were chosen as they provide a representation of ‘early’ and ‘late’ rapid force generation capabilities, which have been suggested to be attributed to slightly different mechanisms [10]. Participants also conducted squat jumps, starting from a 90º degree knee flexion position, with either: a) a near weightless pipe, or b) a 20kg barbell across the rear of the shoulders, to represent unloaded and loaded squat jump performance, respectively. Two trials of each test protocol were conducted and an average used for all subsequent data analysis.
What the authors found
Reliability
Whilst the authors presented both within and between-session reliability data, Table 1 below has been adapted to provide both sets of reliability data in one. In order to be deemed “acceptably reliable”, metrics needed to report an ICC > 0.8 and a CV value < 10%. In addition, previous research has suggested that between-session reliability is a more meaningful indicator of reliability (compare to within-session), as it enables practitioners to have more confidence in the true variance of a given test or metric [11]. Thus, it seems that regardless of protocol (SHORT or TRAD), RFD at both 90 and 200 ms, may be challenging to use in practice, owing to the associated noise. Important to acknowledge though, previous literature has outlined the importance of monitoring the component parts for any given ratio data [9], and RFD is also a ratio (i.e., change in force / change in time). However, it seems rare for studies to report the component parts of RFD; something which may help to explain which variable (if only one) is responsible for the associated noise we often see in this metric.
| Within-session | Between-session | |||
| Metric | CV (95% CI) | ICC (95% CI) | CV (95% CI) | ICC (95% CI) |
| SHORT | ||||
| PF | 3.3 (2.6, 4.5) | 0.97 (0.94, 0.99) | 5.1 (3.7, 8.4) | 0.95 (0.85, 0.99) |
| PF @ 90 | 5.4 (4.3, 7.4) | 0.94 (0.88, 0.97) | 5.6 (3.7, 8.4) | 0.95 (0.86, 0.99) |
| PF @ 200 | 4.4 (3.5, 6.0) | 0.95 (0.90, 0.98) | 5.9 (4.2, 9.7) | 0.97 (0.89, 0.99) |
| RFD @ 90 | 13.4 (11.0, 18.0) | 0.90 (0.80, 0.96) | 20.3 (15.0, 34.0) | 0.76 (0.38, 0.92) |
| RFD @ 200 | 8.4 (6.6, 12.0) | 0.93 (0.84, 0.97) | 14.0 (10.0, 23.0) | 0.84 (0.55, 0.95) |
| Imp @ 90 | 4.8 (3.8, 6.6) | 0.95 (0.89, 0.98) | 10.2 (7.3, 17.0) | 0.92 (0.75, 0.97) |
| Imp @ 200 | 4.6 (3.6, 6.3) | 0.95 (0.90, 0.98) | 5.3 (3.8, 8.7) | 0.97 (0.91, 0.99) |
| TRAD | ||||
| PF | 2.3 (1.8, 3.2) | 0.98 (0.97, 0.99) | 4.1 (2.9, 6.8) | 0.96 (0.89, 0.99) |
| PF @ 90 | 5.5 (4.3, 7.5) | 0.92 (0.84, 0.96) | 4.0 (2.9, 6.6) | 0.91 (0.73, 0.97) |
| PF @ 200 | 4.7 (3.7, 6.4) | 0.95 (0.88, 0.98) | 4.6 (3.3, 7.6) | 0.94 (0.83, 0.98) |
| RFD @ 90 | 23.0 (18.0, 31.0) | 0.71 (0.46, 0.86) | 22.1 (16.0, 36.0) | 0.65 (0.18, 0.88) |
| RFD @ 200 | 10.6 (8.3, 15.0) | 0.91 (0.80, 0.96) | 12.1 (8.7, 20.0) | 0.87 (0.63, 0.96) |
| Imp @ 90 | 4.3 (3.4, 5.9) | 0.95 (0.90, 0.98) | 6.2 (4.4, 10.0) | 0.87 (0.62, 0.96) |
| Imp @ 200 | 4.5 (3.5, 6.2) | 0.94 (0.87, 0.97) | 4.3 (3.1, 7.1) | 0.92 (0.77, 0.98) |
Between-Protocol Differences
When assessing differences between the SHORT and TRAD protocols, the authors cleverly presented this information as a mean difference, with accompanying Hedges g effect size data, inclusive of 95% confidence intervals. Comparisons between like-for-like metrics can be seen in Table 2. To summarise, peak force was the only metric where the TRAD protocol produced higher values. In contrast, the SHORT protocol produced significantly larger values at 90 and 200 ms, for peak force, RFD and impulse. Given the primary aim of this study was to determine whether a SHORT IMTP protocol was able to favour early rapid force-time characteristics, this preliminary evidence provides favourable results.
| Protocol / metric | Mean differences | Hedges g (95% CI) |
| PF (N) | -106.97* | -0.07 (-0.21, 0.07) |
| PF @ 90 (N) | 350.92 | 0.50 (0.31, 0.70) |
| PF @ 200 (N) | 453.51 | 0.49 (0.32, 0.66) |
| RFD @ 90 (N/s-1) | 3532.12 | 0.79 (0.39, 1.17) |
| RFD @ 200 (N/s-1) | 2102.41 | 0.54 (0.29, 0.78) |
| Imp @ 90 (N·s) | 14.66 | 0.38 (0.16, 0.60) |
| Imp @ 200 (N·s) | 63.44 | 0.49 (0.32, 0.67) |
| * signifies peak force was higher in the TRAD protocol | ||
Correlations with Jump Performance
Of all the data collected and relationships ran, the only metric to show meaningful (significant) association with squat jump performance was RFD @ 90 (unloaded squat jump: r = 0.56; loaded squat jump: r = 0.66). No other meaningful correlations were present for either protocol. Although limited, it’s also worth highlighting that practitioners should be mindful of any relationships reported between RFD (and any independent variable) in the present study, as it showed unacceptable reliability. Simply put, given this data exhibits a lot of ‘noise’ there is likely an argument to say that no further analysis should be undertaken for RFD at either time point.
Limitations
With the inclusion of a familiarisation session and test-retest design, the present study appears to have been designed and conducted very well; kudos to the authors. However, if I was being picky, I would say that some additional test protocols would have been useful for the associative analysis, beyond the squat jump, as previous research has outlined a strong association between IMTP peak force and a variety of different independent markers of performance, such as: shot putt distance, 25-m split time during track cycling, 5 and 20-m sprint time, and 505 COD time (Comfort et al. 2019).
What this means for coaches
This study provides preliminary evidence that a shortened IMTP protocol provides comparable data for peak force to a TRAD 3-second protocol, whilst concurrently exhibiting more favourable results for early force-time data using the metrics of force, RFD and impulse, at specific time points (i.e., 90 and 200 ms). Importantly, this enables practitioners to reduce the amount of time they spend testing, when utilising the IMTP protocol. The authors also suggested that the SHORT protocol be used regularly to determine changes in rapid force-time data, whilst the TRAD protocol is still integrated into testing sessions on occasions (e.g., every 3-months), to determine chronic strength adaptations, given this method is likely to exhibit the higher peak force value of the two methods.