You can treat your five age groups as fixed effects. That is, you use a dummy for each age group except for your reference (ideally, the group with the highest number of observations in it). In your code, assuming that the oldest age group 37-59 is a reference, it would look like:

model

https://methods.sagepub.com/dataset/howtoguide/age-in-ess-2016-r

Using a dummy variable for each age group will maximize the explained variance, but before you do that, you might look at whether you can accomplish the same thing with a non-linear variable.

To do this, you should first make sure that your age group variable is coded as interval rather than ordinal. To do that, you should score each category as the mid-point of the age group it represents (e.g., 6 to 12 = 9). After you run a regression with this recoded variable, then add a "quadratic" term which in this case is the square of your age group variable. You can then graph the quadratic equation you get.