Improve testing coverage after PR

# Problem

This [PR](https://github.com/code-corps/code-corps-ember/pull/1463#issuecomment-340201852) resulted in a substantial drop in test coverage (~1%) due to refactoring of imports. This seemingly shouldn't have an impact on testing coverage; however, the current thought is the imports from before this PR were picked up as covered by coveralls when in fact the actual logic was not covered. Here is an [example](https://coveralls.io/builds/13934987/source?filename=app%2Fmixins%2Fonboarding-controller.js).

從 app/mixins/onboarding-controller.js 的 Coveralls 範例開始，將其涵蓋率與 PR 1463 中的變更進行比較，尤其關注 import 重構。找出在沒有測試的情況下被報告為已涵蓋的邏輯，然後為該行為新增涵蓋率，並驗證重構後整體報告不再下降。

由索引模型根據 Issue 內容生成。