A voltage divider follows the equation of Vout=Vin*(R2/(R1+R2)). So in your case 4.7kohm/(4.7+4.7) = 1/2. You can adjust your R1 and R2 values however you want in order to get the ratio you want. For ease I would recommend you keep R2 as 4.7 KOhm and just change the R1 value.
Now as far as your schematic and having 2 sets of R1s and R2s, this is because you essentially have 2 inputs. The connector you are using has 3 pins, ground, left, and right. There are different variances to drum kits, but the one I used in the past would put one instrument on left and a different on right. You need 1 set of R1/R2 per input signal which results in you having 2 per plug.
As a side note, I would suggest you consider NOT doing this voltage divider method and instead go with an actual sound board that lets you adjust the levels.
Let's do another thought-experiment:
Imagine that we gradually shorten the connection between the two resistors until it is infinitesimally small. Now you effectively have one resistor with the divider-point somewhere in the middle. One end of this resistor is connected to the supply voltage, let's say 5 Volts. The other end is connected to ground, which we'll call 0 Volts because we'll use it as the reference point for our voltage measurement.
Again, imagine that we gradually move the divider-point up toward the 5V end or down toward the 0V end. At what location along this combined resistor would you expect the measured voltage to drop from 5 Volts to 0 Volts?
Is it clear that the voltage doesn't have a step change at one point, but is proportional to the fraction of the resistor below the divider-point? Voltage drop is linear along the length of the combined resistance.
Now imagine that we gradually restore the length of that connection we stretched in the first visualization. And imagine that the connection itself has no resistance - zero. (It's not quite zero, but it's so close that we can ignore it). Is it clear that the voltage will be no different at either end of the connection as you stretch it out again?
The total voltage drops linearly over the total resistance, and the divider point "samples" that voltage at a specific part of the total.
The current doesn't "care" or "know" whether is flowing through many individual resistors, or one continuous one; the voltage drops continuously along each segment of the resistance path. The "zero resistance" connections just don't matter.
Assuming you're talking about translating logic levels- if you're not too concerned about speed or power consumption, the resistors will work.
However, assume something like 30pF of input and stray capacitance, and resistors that equal (say) 10K in parallel and 40K in series (2 20K resistors). The resistors will add a delay of around 300 nsec and will consume 90uA for each that is high.
If you're talking about the power supply for an IO port, you should use a regulator. The regulator will draw less current and will provide a stable "stiff" voltage (preventing port pins from interacting with each other, for example).