MVP - Presenter comparision: view injected in constructor vs takeView(View v) / dropView()

Question

I've seen two approaches to MVP pattern in Android. Both are used in Android Architecture Blueprints:

public interface Contract {
    interface View {
        void showData(String data);
    }

    interface StartVersionPresenter {
        void start();
    }

    interface DropViewVersionPresenter {
        void takeView(View view);
        void dropView();
    }
}

1) The presenter in which the view is injected via the constructor:

public class StartVersionPresenter implements Contract.StartVersionPresenter {

    private final Contract.View view;
    private final Repository repository;

    public StartVersionPresenter(Contract.View view, Repository repository) {
        this.view = view;
        this.repository = repository;
    }

    @Override
    public void start() {
        loadData();
    }

    private void loadData() {
        repository.getData(new DataCallback() {
            @Override
            public void onDataLoaded(String someData) {
                view.showData(someData);
            }
        });
    }
}

start() is called in onResume() Fragment's method.

2) The presenter in which the view is injected via the method (takeView in my example):

public class DropViewVersionPresenter implements Contract.DropViewVersionPresenter {
    private final Repository repository;
    private Contract.View view;

    public DropViewVersionPresenter(Repository repository) {
        this.repository = repository;
    }


    @Override
    public void takeView(Contract.View view) {
        this.view = view;
        loadData();
    }

    @Override
    public void dropView() {
        view = null;
    }

    private void loadData() {
        repository.getData(new DataCallback() {
            @Override
            public void onDataLoaded(String someData) {
                if (view != null)
                    view.showData(someData);
            }
        });
    }
}

takeView(Contract.View view) is called in Fragment's onResume() method. dropView() is called in Fragment's onDestroy() method.

In both cases the presenter is created in Fragment's onCreate() method.

Why does the second approach is used in more cases rather then the first one? I can see that the first one is simpler because we don't have to check that the view is not null if we want to call the method on it.