I like the readability of:
if !params[:foo].blank?
  stuff
end
Or just:
stuff if !params[:foo].blank?
But I didn't know if I was giving myself any performance gain in writing it this way. Finally I took the time to Benchmark it using the following:
iterations = 1_000_000
    Benchmark.bm do |bm|
      # using if &&
      bm.report do
        iterations.times do
          if params[:event_type] && params[:event_type] != ''
            with['event_type'] = params[:event_type].to_i
          end
        end
      end

      # using !blank?
      bm.report do
        iterations.times do
          if !params[:event_type].blank?
            with['event_type'] = params[:event_type].to_i
          end
        end
      end
      # using unless blank?
      bm.report do
        iterations.times do
          unless params[:event_type].blank?
            with['event_type'] = params[:event_type].to_i
          end
        end
      end
    end
And the results are:
user     system      total        real
   3.190000   0.010000   3.200000 (  3.215461)
   3.460000   0.010000   3.470000 (  3.481160)
   3.490000   0.010000   3.500000 (  3.509005)
The least ruby-like method of writing this turns out to be the fastest...