So với các nghiên cứu quan sát khác thì "Sai số chọn" trong nghiên cứu thuần tập là:

Giá trị suy luận căn nguyên (khả năng chứng minh mối quan hệ nhân quả) của các thiết kế nghiên cứu được sắp xếp như sau:

* Nghiên cứu bệnh chứng (c): Đây là thiết kế yếu nhất trong việc suy luận căn nguyên vì nó bắt đầu với bệnh và hồi cứu để tìm yếu tố nguy cơ. Điều này dễ dẫn đến sai lệch do nhớ lại và khó xác định trình tự thời gian.
* Nghiên cứu thuần tập hồi cứu (b): Nghiên cứu này mạnh hơn bệnh chứng vì nó xác định кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого кого

Phương pháp nghiên cứu bệnh chứng là một phương pháp nghiên cứu hồi cứu (retrospective), trong đó người ta so sánh một nhóm người mắc bệnh (nhóm bệnh) với một nhóm người không mắc bệnh (nhóm chứng) để tìm hiểu xem có những yếu tố nào trong quá khứ có thể liên quan đến sự phát triển của bệnh. Một trong những nhược điểm lớn nhất của phương pháp này là sự phụ thuộc vào hồ sơ và trí nhớ của đối tượng nghiên cứu, điều này có thể dẫn đến sai lệch (bias) do thông tin không đầy đủ hoặc không chính xác. Vì hồ sơ bệnh án có thể không đầy đủ hoặc bị mất, và người tham gia có thể không nhớ chính xác các sự kiện hoặc yếu tố phơi nhiễm trong quá khứ. Do đó, đáp án "Tài liệu, hồ sơ cần thiết không hòan chỉnh" là nhược điểm lớn nhất của phương pháp này. Các đáp án khác có thể đúng trong một số trường hợp nhất định, nhưng không phải là nhược điểm cốt lõi của phương pháp nghiên cứu bệnh chứng như việc thiếu tài liệu hoặc hồ sơ hoàn chỉnh.

Thiết kế nghiên cứu thực nghiệm (c) cho phép kiểm soát các biến số và thiết lập mối quan hệ nhân quả rõ ràng nhất, do đó có giá trị suy luận căn nguyên cao nhất. Thiết kế trường hợp (b) có thể gợi ý mối liên hệ nhưng không chứng minh được nhân quả. Thiết kế tương quan (a) chỉ cho thấy mối liên hệ thống kê giữa các biến, không thể suy luận về quan hệ nhân quả. Vậy, trình tự giảm dần về "Giá trị suy luận căn nguyên" là: Thực nghiệm (c), Trường hợp (b), Tương quan (a).

Giá trị suy luận căn nguyên (causal inference) thể hiện khả năng xác định mối quan hệ nhân quả giữa các yếu tố. Trong các thiết kế nghiên cứu:

* Trường hợp (Case study): Thường mô tả chi tiết một hoặc một vài trường hợp cụ thể, ít có khả năng suy luận nhân quả do thiếu nhóm so sánh và kiểm soát các yếu tố gây nhiễu.
* Thuần tập tương lai (Prospective cohort study): Theo dõi một nhóm người theo thời gian để xem xét sự phát triển của bệnh hoặc kết quả dựa trên các yếu tố nguy cơ đã được xác định trước. Có khả năng suy luận nhân quả tốt hơn nghiên cứu trường hợp vì có nhóm so sánh (những người không tiếp xúc với yếu tố nguy cơ) và có thể xác định mối liên hệ theo thời gian.
* Thực nghiệm (Experimental study): Là thiết kế mạnh nhất để suy luận nhân quả vì nhà nghiên cứu chủ động can thiệp (thường là ngẫu nhiên) vào yếu tố nghiên cứu và so sánh kết quả giữa nhóm can thiệp và nhóm chứng. Điều này cho phép kiểm soát các yếu tố gây nhiễu và xác định mối quan hệ nhân quả một cách rõ ràng nhất.

Vì vậy, giá trị suy luận căn nguyên tăng dần theo thứ tự: Trường hợp, Thuần tập tương lai, Thực nghiệm, tương ứng với a, c, b.

Giá trị suy luận căn nguyên (causal inference) đề cập đến khả năng xác định mối quan hệ nhân quả giữa một yếu tố (nguyên nhân) và một kết quả (hậu quả). Các thiết kế nghiên cứu khác nhau có khả năng suy luận căn nguyên khác nhau.

* Thực nghiệm: Thiết kế thực nghiệm, đặc biệt là các thử nghiệm ngẫu nhiên có đối chứng (RCTs), được coi là tiêu chuẩn vàng để suy luận căn nguyên. Trong thiết kế này, nhà nghiên cứu chủ động can thiệp và kiểm soát các yếu tố, từ đó có thể xác định mối quan hệ nhân quả một cách mạnh mẽ nhất.
* Thuần tập tương lai: Nghiên cứu thuần tập tương lai theo dõi một nhóm người theo thời gian để xem liệu một yếu tố nguy cơ có liên quan đến kết quả hay không. Mặc dù có thể cho thấy mối liên hệ, nhưng khó loại trừ các yếu tố gây nhiễu khác, làm giảm khả năng suy luận căn nguyên.
* Thuần tập hồi cứu: Tương tự như thuần tập tương lai, nhưng dữ liệu được thu thập từ quá khứ, làm tăng nguy cơ sai lệch và khó khăn trong việc xác định trình tự thời gian giữa nguyên nhân và kết quả.
* Bệnh chứng: Nghiên cứu bệnh chứng so sánh những người có bệnh với những người không có bệnh để xem liệu có sự khác biệt về yếu tố nguy cơ hay không. Thiết kế này dễ bị sai lệch chọn mẫu và khó xác định trình tự thời gian, do đó khả năng suy luận căn nguyên thấp nhất.

Do đó, thiết kế thực nghiệm có giá trị suy luận căn nguyên cao nhất vì nó cho phép kiểm soát và can thiệp, giúp thiết lập mối quan hệ nhân quả mạnh mẽ hơn.