If you mean listening in the air while crushing the bag under water - the main reason is due to the different acoustic impedances of air and water. Transmission of the sound of the bag popping through the water probably plays a secondary role.
Acoustic impedance is defined as
$I =\rho c$
where $\rho $ is the density of the medium, and $c$ the sound speed in the medium.
The acoustic impedance of water is larger primarily because it is so much denser than air.
The acoustic impedance of water is much larger (about 3400 times) that of air. The transmission coefficient of sound from water to air (and vice versa) is roughly the ratio of the acoustic impedance of air to water, i.e 1/3400.
The actual value of the intensity transmission coefficient from medium 1 to 2 with acoustic impedances $I_1$ and $I_2$ is
$T =4\dfrac{\dfrac{I_1}{I_2}}{(1 + \dfrac{I_1}{I_2})^2}$
Which works out to 0.0012 for medium 1 = water, medium 2 = air
So hardly any underwater sound transmits to the air above.
I am basing this almost completely on the comments above and my own experience of making a "tin can phone" as a kid.
The string was pulled as tight as we could get it, on the basis that a loose string, i.e. no tension, would not carry the sound waves very far.
But the tension in the string also makes the base of the cup vibrate, increasing your chances of getting your shouting into your "phone" being carried down the wire and setting up matching vibrations on the base at the far end.
So, in theory at least, you shout into the phone, the base of the cup vibrates in a certain pattern, the tense string carries that pattern down to the base of the other cup, and it vibrates in the same pattern.
In reality, as far as I remember, I had to shout soooo loud that the other guy could hear me anyway, phone or no phone.
Tin Can Phone: Wikipedia
Sound waves are created as the air vibrates in response to a person's speech or other sounds. A second person's ear collects these sound waves and converts them into nerve impulses which their brain interprets as sound. In normal speech these waves travel through the air, but with a tin can telephone the waves are transmitted through an additional medium of cups and string.
When the string is pulled taut and someone speaks into one of the cans, its bottom acts as a diaphragm, converting the sound waves into longitudinal mechanical vibrations which vary the tension of the string. These variations in tension set up longitudinal waves in the string which travel to the second can, causing its bottom to vibrate in a similar manner as the first can, thus recreating the sound heard by the second person.
The signal can be directed around corners with the aid of a third can positioned on the apex of the corner. The string is threaded through the base of the third can so as to avoid coming into contact with the object around which the signal is to be directed.
The 600 feet Tin Can Phone A video that could really use some editing, far too long, just skip to the last minute.
Best Answer
Many plastics make noise when crushed, and crackle again when opened up, e.g. when removing a candy from its plastic wrapper (to the frequent irritation of theatre patrons). As JMac commented, similar noises occur with many types of paper.
In general this happens with thin materials that do not easily stretch or compress, but that can bend without breaking and form creases and deformations that can store elastic energy. As a sheet is crushed it first bends and then suddenly buckles into a different configuration, making a cracking noise and often a permanent crease. As you continue to crush the sheet, this repeats many times on smaller and smaller scales, producing a stream of crackles. Once crushed, the many creases allow many possible configurations that the sheet can suddenly switch between if you try to open up (or otherwise further reshape) the sheet, producing clicks for every transformation. Materials such as soft cloth fold and deform easily without much resistance or sudden transformations, so the little noise they produce is primarily from rubbing friction.
To learn more, have a look at “Acoustic emission from crumpling paper” or “On the Noise from a Crumpled Candy Wrapper”. (The latter is a simple discussion of this paper.)