ARM - week 6

Question 1

description versus causal inference

Answer 1

. Description:
§ You ruin your analysis by statistical adjustment (partial association)
§ Results no longer make sense
2. Causal inference:
§ You ruin your analysis by lack of statistical adjustment
§ Results are biased (confounders)
In the PC labs, we used description to explore the data before running (adjusted) regression analysis and draw causal inference.

Question 2

description goal

Answer 2

Goal:
§ To identify patterns in data
§ Obtain factual information

Goal is not to:
§ Explain these patterns
§ Draw causal conclusions

Question 3

prediction

Answer 3

Goal:
§ Predict/forecast the future (or the past, or the present)
§ If you know A, B, C, what can you say about D?

Goal is not to:
§ Draw causal conclusion (‘what would have happened if …’)

Question 4

prediction versus causal inference

Answer 4

1. Prediction
   § You include any (X) variable that contributes to predicting the outcome (Y)
   § You do not care about confouding bias
2. Causal inference
   § You only include the (X) variable(s) of interest and confounding variables to avoid bias
   § You care (a lot!!!) about avoiding bias
   In work group 2, we predicted the WTP for a mammogram of a non-Hispanic (white) woman, who spent all of her life in the US, was 50 years old in 1999, 15 years of education, etc.

Question 5

strengths of qualitative research

Answer 5

• Rich description of processes and experiences
• Knowledge construction and power relations
• Moving targets and phenomena in formation