Back to main page

2000
Week 36
Week 37
Week 38
Week 39
Week 40
Week 41
Week 42
Week 43
Week 44
Week 45
Week 46
Week 47
Week 48
Week 49
Week 50
2001
Week 2
Week 3
Week 4
Week 5
Week 6
Week 7
Week 8
Week 9
Week 10
Week 11
Week 12
Week 13
Weeks 14-20

Back to main page

The activities of this week can be briefly summarised: Report writing and data analysis. Jørn had been given a print of the report written thus far and gave thorough and valuable comments. His main criticism was that the report contained too much discussion and too many considerations in the “Method” chapter. His expressed wishes for a shorter, more concise and straight-to-the-point approach. Discussions and considerations should be left to be included in the “Discussion” chapter at the end of the report. Another comment, which we had discussing among ourselves, was the question of which tense to use. Jørn recommended past tense was used until the discussion chapter.

Had a fruitful meeting with Henrik Spliid, where we settled the final questions relating to our data analysis. One fundamental point, which had been a little unclear, was how data should be tested for normal distribution. It was our impression that the practice at the Centre generally is to test for normality on the raw data. According to Spliid, this is “objectively wrong”, as he put it. It is the residuals computed by the descriptive model (in our case a GLM), which must be tested for normality. The reason being that the raw data may be strangely distributed – not necessarily because they are not normally distributed but rather because the raw data are affected by a cocktail of different factors in different combinations. The statistical model takes this cocktail of factors into consideration, and yields normally distributed residuals insofar the model fits the data well.

Spent the weekend writing our report to Spliid – the report, which marks the determination of our course with him. It has been a most positive experience working with Spliid; he has been very patient, kind and helpful throughout the entire process.