To understand how to correct the problem, it often helps to think carefully about what, exactly, might be happening. But in summary, it may help to:
- take a shoulder-width stance,
- externally rotate your feet—slightly, but not excessively,
- allow the knees to go over the toes,
- squat to the ankles (i.e. sit down, not ‘back’),
- maintain neutral anterior pelvic tilt,
- maintain a knee angle that is the same as the hip angle, and
- extend the hip and knees simultaneously, and at the same rate.
(Now, if you want to understand why, please bear with me.)
Single-joint or ‘isolation’ exercises are generally controlled by a small number of muscles, whose involvement is dependent only upon bone orientation and joint angle. The consequent expression of movement, therefore, closely describes how, and to what degree, the relevant agonists, antagonists, and synergists are being activated.
Multi-joint or ‘compound’ exercises, by contrast, are controlled by a large number of muscle groups, whose involvement is dependent upon interactions between the joints involved, and upon the degree of freedom afforded by the exercise. The lesser the degree of freedom, the more variant the possibilities of muscle activation. This is why ‘free-weight’ exercises are generally superior for developing functional strength: the greater degree of freedom afforded by these exercises forces the lifter to activate the muscles in a certain way.
To illustrate this point, consider three common variations of the chest press: machine, barbell, and dumbbell. Each of these exercises involves the same primary joint actions—humeral flexion and horizontal adduction, and elbow extension. Yet the potential variation in muscle activation between them is enormous. The machine chest press, being a closed chain exercise, can be performed with almost any permutation of muscle activation. The lifter could theoretically have no activation of the pectoralis major or deltoids whatsoever, and perform the movement exclusively with the triceps brachii! This is possible because the final position of the load is a function of any one of the joint angles involved. Similarly, the barbell bench press could be performed without any activation of the pectoralis major, and the dumbbell press could be performed without any activation of the triceps brachii.
So too is enormous variation possible with the squat. The squat involves leg ‘triple extension’—that is, extension of the hip, knees, and ankles—and therefore engages some combination of the hip, knee, and ankle extensors. Involvement of the ankle extensors (the triceps surae) is dependent upon the position of the centre of mass relative to the calcaneal tuberosity. Again, theoretically, the triceps surae could be completely uninvolved if the centre of mass passed perfectly through the point of contact between the heel and the floor—that is, if our weight were as far backward as it could be without our falling over.
However, since the kinetic chain is open at the top—that is, since the barbell and body are free to migrate forwards and backwards—the load must be supported by both the knee extensors (quadriceps femoris) and one- and one-/two-joint hip extensors (the gluteals and hamstrings, respectively) in some capacity that results in a net upward vector. It is the relationship between the three joint angles that determines the degree of muscle activation required of each group; or alternatively, it is our muscle action that determines the posture that we tend to adopt in performing the squat. And contrary to the statements of some commentators, since the moments at the joints must be balanced, we cannot over-emphasise one muscle group without also altering our position and/or emphasis of the others.
Of course, it should be mentioned that our physical geometry influences our technique—both that which is optimal and that which is possible. But of the criteria that we can modify, those that alter muscle activation are the depth and uprightness of our squat. The depth of our squat is, in turn, limited by the width of our stance and the flexibility of our ankles. Since the hamstrings are bi-articular, their activation depends on the difference between the hip and knee angles. The more positive the difference, the less the hamstrings are activated, and the more, therefore, that the gluteals must be activated in order to balance the moment created by the quadriceps. Maintenance of an anterior pelvic tilt through the activation of the erector spinae increases the effective angle of the hip.
The key to maximising activation of the gluteals, therefore, is to adopt a stance that allows us to squat to full depth. The feet should generally be spaced at shoulder width, and slightly externally rotated. The knees should be held outward (which further engages the gluteals) such the the knees track in the direction of the feet. And the erector spinae should be activated powerfully to counteract the force generated by the hip extensors. The squat should be performed such that the knee and hip angles are equal. This is achieved by squatting towards the ankles—not ‘sitting back’—and allowing the knees to track forward freely. If the knees are not permitted to move freely past the line of the toes, both the depth of the squat and the angle of knee bend will thereby be limited. Finally, the hip and knee should be extended at the same rate, thus discouraging a two-phase lift in which the tail is lifted before the back is extended.
I hope that is helpful. Please feel free to ask for clarification for any of these points.