Hypoteseprving, kontinuerlige data

1
Hypoteseprøving, kontinuerlige data

• Thore Egeland

2
Innhold

• Kontinuerlige data (målevariable)versus
  diskrete data
• Parvisversus to-utvalg

3
SUBJECT PREMENST POSTMENS 1 5260.0
3910.0 2 5470.0 4220.0 3
5640.0 3885.0 4 6180.0 5160.0
5 6390.0 5645.0 6 6515.0 4680.0
7 6805.0 5265.0 8 7515.0
5975.0 9 7515.0 6790.0 10
8230.0 6900.0 11 8770.0 7335.0
Example Table 9.3 Altman (Dietary intake kJ)
4

mean (premenst-postmens) 6753.636-5433.182
1320.46 sem (premenst-postmens)
110.58
5
Hypotheses and P-value

• H0 Pre - and post menstrual dietary intake
  equalA Pre - and post menstrual dietary intake
  differ
• P-value probability of the observed value
  or something more extreme
  assuming null hypothesis
• observed value 11.94p-value0.000 (SPSS).
  Figure.
• We reject H0 since P-value is below 0.05

6
Two-sample test

• H0 energy expenditure equal in lean and
  obeseA energy expenditure differs in lean and
  obese
• Recall that the alternative A iswhat we would
  like to show
• There are two samples or two groups lean and
  obese.

7
ID GROUP ENERGY 1 0
6.13 2 0 7.05 .... 12
0 10.15 13 0 10.88
14 1 8.79 15 1 9.19
.... 21 1 11.85 22
1 12.79 Number of cases read 22 Number
of cases listed 22
Example Altman p. 193
8
Read second line (Equal variances not assumed)
9
Conclusion

• The mean for the lean and obese was 8.1 and 10.3,
  respectively (mean difference -2.2)
• The difference was significant. The p-value based
  on a two-sample t-test was 0.001.

10
Another example, PEF data

• H0 pefmean sitting equal in male and femaleA
  pefmean sitting differs between male and
  female

11
Conclusion?
12
Assumptions

• 1. Independence All observations are
  independent. This is typically achieved by taking
  random samples.
• 2. Normal distribution
• 3. Equal variances

13
Discussion of assumptions

• 2. Normal distribution Plots (Histograms, QQ) or
  tests

14
(No Transcript)
15
(No Transcript)
16
SPSS
Click Analyze - Compare Means -
Independent-Samples T Test pefsitm is transferred
to test-variable gender is transferred to
Grouping variableClick Define Groups and
enter 1 and 2 for the groups
17
What if assumptions are invalid?

• Use non-parametric methods Paired data
  WilcoxonTwo-sample Mann-WhitneyOR
• Transform data

18
Skewed distribution skewed0.400.96 .
Ln transformed distributionskewed
lnskewd0.40 ln(0.40)-0.910.96
ln(0.96)-0.04 .
We base analysis on lnskewd rather than
skewedSPSS transform- compute
19
Summarizing

• P-values
• Confidence intervals (later)
• Assumptions
• Important to distinguish betweenPairedTwo-sample
  s

20
(No Transcript)