Newbie: 1 Control Group with no treatment + 1 Treatment Group

#1
Hi,

I would never ever have thought there would be a forum to discuss and ask for help on statistics but this merely highlights why I love the internet so much.

I recently did a small trial at work.

I had 17 patients with a hip fracture who required surgery.

I took blood glucose levels pre operation, 2 hour post operation and 24h post op. There is a normal range between 4.5 - 6.0.

2 Groups:

Control Group: I measured there levels and treated them with nothing.

Treatment Group: I gave them a special drink pre operatively.

My hypothesis is that the Control group will have higher blood glucose after their operation and that the treatment group will have lower blood glucose levels.

ie I want to prove that by giving this drink before the operation I can show that the glucose levels taken after the operation show they are lower ie not hyperglycaemic.

So I have 2 tables with 3 columns each. One table is Control and the other is Treatment, each with 3 sets of blood glucose levels. There is an exact point of >6.0 where I consider the glucose to be HIGH ie hyperglycaemic.

How do I start to analyse my data? I appreciate any help whatsoever.

Should I post figures here?

Regards
Darren
 

Dragan

Super Moderator
#2
Hi,

I would never ever have thought there would be a forum to discuss and ask for help on statistics but this merely highlights why I love the internet so much.

I recently did a small trial at work.

I had 17 patients with a hip fracture who required surgery.

I took blood glucose levels pre operation, 2 hour post operation and 24h post op. There is a normal range between 4.5 - 6.0.

2 Groups:

Control Group: I measured there levels and treated them with nothing.

Treatment Group: I gave them a special drink pre operatively.

My hypothesis is that the Control group will have higher blood glucose after their operation and that the treatment group will have lower blood glucose levels.

ie I want to prove that by giving this drink before the operation I can show that the glucose levels taken after the operation show they are lower ie not hyperglycaemic.

So I have 2 tables with 3 columns each. One table is Control and the other is Treatment, each with 3 sets of blood glucose levels. There is an exact point of >6.0 where I consider the glucose to be HIGH ie hyperglycaemic.

How do I start to analyse my data? I appreciate any help whatsoever.

Should I post figures here?

Regards
Darren


What you are describing is a Within Subjects Factorial Design: A x (B x S) where A=2 (experiment or control) B = 3 (treatment conditions or measures) and S is your subjects factor. This particular type of design is sometimes referred to a two-factor desgin with repeated measures.

I would suggest you look at Keppel and Zedeck Chapter 17 pages 294-305. This book is easily accessible to read and it will show the computations.

Another reference would be Keppel and Wickens.

(Just Google the names.)


Minor point: Remember, you can't really prove your assertion. Rather your statistical analysis can only offer evidence to support your hypothesis.