You are on page 1of 3

❖ Graph 1

➢ Trend, Steepness, Manipulation

Risk of CVD (au)


1] Describe the relation between smoking level
& CVD risk in males and females (3)
1-In Both males and females, as smoking level
,
,the risk of CVD (+ve correlation).
2- Linear increase in both.
3-Risk in males is always higher than females at
all smoking levels.
4-The rise in males is steeper than in females. Smoking level (au)
5-Manipulation: Subtract, ratio , % (choose the
easiest). ex. At smoking level 2 au, the risk in
males is higher than females by 4 au. ( never
forget the unit!!!) not just copy figures !
(ie fixed the 2 au on the X axis ,and subtracted 8 - 4 on the Y axis .
-wrong: At smoking level 2 au, the risk in males is 8 au while that in females is 4 au
-wrong.. the risk in males by - - - - XX (because you didn’t mention females) !!
2] Give evidence from the graph that there are other factors/ genetic factors linked
to increased risk of CVD
-Even non smokers have some risk of CVD in both males & females (at smoking level
zero!)
-Risk in males is always higher than females at the same smoking level.
3] What evidence shows that smoking is correlated with risk of CVD.
- In both males & females , as the smoking level , the risk of CVD lineary.
❖ Graph 2
In this graph, points of trend and manipulation are
Risk
the same as graph 1, however, since both lines are
parallel, they have same steepness !! (point 4 )

Smoking level
Error Bars (variability)

- Show the degree of spread of the results of the repetitions around the mean.

(I) Narrow error bars Reliability! wide error bars

- Low variability, so -High variability, so

→ - High reliability The mean can be the → - Low reliability


same, but in (A), the
mean 10 results of repetitions mean 10
are more similar/
close/clustered to the
mean, less variable
so more reliable!!

(A) (B)

Error bars can be obtained by:


a) Range: simply difference between highest & lowest results of the repetitions.
(However, one anomaly would give a wide error bar based on the “Range”.,,,so
better use SD→
b) Standard deviation: S.D. shows the average degree of spread of the data around
the mean, considering the sample size ( The number of repetitions.)
It is calculated by a certain equation. This is more objective, as only one or two
anomalies within large number of repetitions still narrow error bar. (see graphs)

S.D. of the same


Range experiment &
same results
(II) Overlap of error bars
- When assessing whether there is /is not a “ SIGNIFICANT difference/correlation”
between groups ex. blood pressure with and without using a certain drug.

No overlap
overlap

   

- No overlap of error bars - Overlap of error bars


- No common values bet. gp  & gp  - Common values bet. gp  & gp 
- The highest values of gp  are
- similar to the lowest values of gp
- There is a significant difference - There is a non significant difference
between the results of the 2 groups. between the results of the 2 groups.
- You can make a valid conclusion→ - May be due to chance.
This drug may be effective. -

Confirm by statistical tests in A2 !!

You might also like