I have recently done my spiral dive recovery lesson in a glider ASK21. The instructor taught me to level the wings first, then ease out of the dive by pulling the stick back. He said that pulling the stick back would only tighten the spiral and not help with reducing the speed. Why is that? Should the nose not pitch up and reduce the speed?
The effects of rudder and elevator get mixed when the aircraft flies at a roll angle. See it this way: When you pull the elevator at a roll angle of 60°, it will mostly (86.6%) work like a rudder would for a level airplane.
Why those 86.6%? When looking at the aircraft in an earth-fixed reference system, the additional tail force from an elevator deflection will produce a side load component proportional to the sine of the roll angle (which is 86.6% at 60°). The vertical component which would help to get the nose up is only 50% at a roll angle of 60° because it is proportional to the cosine of the roll angle.
Coordinated flight at higher roll angles is almost like flying a loop sideways. You will only lift your glider's nose in the reference system of the glider and tighten the loop. The reference system that counts more here is the one in which gravity acts, and you should lift the glider's nose up in this system.
Not knowing this subtlety ended the life of several early aviators until it was found out in 1912 that the rudder is what lifts the nose in a spiral dive. Of course, it is even more efficient to first roll wings level and make the elevator work properly again.
The horizontal-loop thing works in reverse, too: Rüdiger Kunz, the designer of the Standard Austria, was always fond of telling the story that happened on occasion of the first flight: At some point, the pilot started to fly loops. After the third, he asked Herr Kunz over the radio: "Notice anything?" - "Yes, you're climbing" was his answer. Thereupon the pilot replied: "That's how you can circle in thermals, too!"
In a steep spiral with a steep roll attitude, that is true as most of the lift vector is directed into a horizontal component. Therefore aft stick pressure is simply going to force the aircraft into a tighter turn. In addition this is going to impose much higher structural loads at high air speeds which may be beyond the design limits of the glider and put you at risk of a structural failure. Your best bet from a steep spiral is to roll wings level, then pull out of the dive.
The other posts answer the question, but I think they fail to make an important point, or at least they fail to emphasize it. You are practicing the Spiral Dive Recovery, which is a "recovery" procedure. You are trying to extricate yourself from something like an unusual attitude. You are attempting to recover the aircraft, and not trying to stay in the turn!
If you are recovering from a spiral dive, then you level your wings first so that when you pull back on the stick all the lift generated is towards getting you out of the dive. This is the "recovery" of the exercise. This is solely about stopping your descent.
If you remain in the turn while pulling back on the stick then some of the lift will go into tightening the turn, instead of pulling out of the dive.
It might be possible that your intention is to be in a safe spiral dive, and in that case it might be appropriate to tighten the turn. For example, you might want to get the nose around quicker. This would not be a typical recovery procedure though.
I was doing a popup delivery in the single seat, light-attack jet A7-E Corsair when I successfully "recovered" from a turn in a steep dive. At 100 feet AGL, 750 MPH, with the target 3 miles away the pilot pulls the nose up to between 15-30 degrees to reach an apex in the climb of approximately 3,000 feet. At the apex the aircraft is rolled on its back and the target identified, pulling the nose onto it while inverted. In a 7-15 degree dive the run is made and bombs released with the egress fast and low. This is a very dynamic maneuver close to the ground, with lots of tasks to complete in a short period of time.
I found myself in the dive while still in a hard turn trying to get my nose on the target to release my bombs. The HUD indicated a 15 degree dive, and a safety article I read months earlier flashed through my head. Thank you Grandpa Pettibone! "If you find yourself in a 15 degree dive in a popup maneuver level your wings immediately and pull up." If I hadn't leveled my wings, and tried to stay in the turn while recovering I wouldn't be writing this today. As I came over the cleared target area, pulling in buffet, and finally had the flight path directed above the ground, I was looking up at the tree tops that surrounded it. This is a successful recovery procedure.
It may help you to understand the situation if you consider this-- imagine that you trimmed the glider for a moderately fast glide-- like perhaps 60 or 70 knots. Now without exerting any forward or aft pressure on the stick, slowly roll the glider into a steep bank-- perhaps 45 or 50 degrees. You'll find the airspeed rapidly increases as the glider enters a spiral dive. This is partly due to the fact that for any given angle-of-attack, the steady-state airspeed is proportional to the square root of the load factor, and banking increases the load factor. This is also partly due to the fact that an aircraft-- especially a relatively slow-flying aircraft like a glider-- tends to trim to a lower angle-of-attack in turning flight than in linear flight, due to aerodynamic effects generated by the curvature of the flight path.
Once the glider has entered the spiral dive, you'll find that simply rolling the wings back to level is sufficient to make the flight path start curving up out of the dive toward the trajectory associated with the original trim speed-- even if you exert no aft pressure on the control stick at all. The reason for this is that as long as the glider retains some of the excess speed associated with the spiral dive, the wing is generating too much lift to allow the flight path to simply continue in a straight line on a diving trajectory.
Conversely, if you've ever practiced taking your hands completely off the control stick and controlling the aircraft via rudder inputs only, so that you are not making any pitch inputs at all, you've noticed that the rudder has a roll effect, and you've also noticed that the nose tends to rise as you roll towards wings-level, and tends to descend as you roll toward a steeper bank angle.
So it's clear that simply decreasing the bank angle, without pulling the stick aft at all, is sufficient to allow the aircraft to eventually recover from a spiral dive. On the other hand, as other answers have noted, if the bank angle gets steep enough, then pulling the stick aft without decreasing the bank angle will not bring the nose up toward the horizon.