scatterplot

[laurenhosking]

I’ve created a scatterplot that portrays how weight and height differs between genders. However I’m unsure what comments to make on what the plot is show. Also this is the most basic scatterplot would making it more advance help with a conclusion

[Rick_SAS]

If you attach your plot we might be able to help. There are many ways to view these data. You might comment on the centers (clusters), the spread (variance/covariance), or on the slopes (regression).

If you are thinking about the data in a regression context (weight is a dependent variable; height is an independent variable) then you can form an ANCOVA (analysis of covariance plot) and discuss whether the males and females have the same intercept and same slope. You can create such a plot by using PROC SGPLOT, or you can use PROC GLM:

proc glm data=sashelp.class plots=fitplot;
class sex;
model weight = height sex height*sex;
run;

proc sgplot data=sashelp.class;
scatter x=height y=weight / group=sex;
reg x=height y=weight / group=sex nomarkers;
run;

[laurenhosking]

thank you so much. This is my plot. are the above things ok to comment on for this.

[PaigeMiller · Accepted solution]

Here are two conclusions I would make from your plot

1. Height and weight are positively correlated
2. Males generally have higher weights and higher heights than females.

[Rick_SAS]

also

3) The lightest and shortest subjects are female. The heaviest and tallest are male.

4) The average height and weight of the females appear to be less than the average height and weight of the males.

5) There are some outliers. Four men are 180-190 cm tall but are more than 100 kg. A few women are 170 cm but more than 85 kg.

[PaigeMiller]

---

@Rick_SAS wrote:

also

5) There are some outliers. Four men are 180-190 cm tall but are more than 100 kg. A few women are 170 cm but more than 85 kg.

---

whether or not something is an outlier is subjective and depends on your assumptions and possibly some subject matter expertise — given my assumptions (which is subjective) and my knowledge of the subject matter (which is subjective) I would say there are no outliers shown here.

[Ksharp]

You could use ELLIPSE to demonstrate what Rick pointed out .

proc sgplot data=sashelp.heart(obs=500);

scatter x=weight y=height/group=sex;

ellipse  x=weight y=height/group=sex;

run;