One or Two Way ANOVA?

[squath]

Hi all,

Just had a quick question regarding whether or not I should be using a one or two way ANOVA to see the effect of caffeine on heart rate. My groups are: control, coffee, energy drink. My sample size is 50 for each group and for each individual I tested at the following times: initial, 30 mins., 60 min., 90 min., 120 min. after consumption.

My goal is to see if the experimental groups have a significant effect on HR but I'm unsure how to go about setting this up in Excel. Any ideas would be greatly appreciated.

 

[rogojel]

hi,
I would go for two-way , one factor would be the drink, the other the time. I would be curious to see what you get.

regards

 

[Dragan]

You could also consider Trend Analysis (using contrast coding) because your second factor is quantitative (time) with equal spacing i.e. initial, 30, 60, 90, 120.

 

[mostater]

You may also want to add group * time interaction to the mix to see if the effect of group on HR varies over time. For example, coffee may impact HR much later than an energy drink. I am not certain what Excel can handle (probably a 2-way ANOVA but unsure about interactions).

Also, were subjects randomized to each of the 3 groups? If not, there may be other differences between groups that affect heart rate for which you may need to control (covariates). Having different baseline HR between groups would be a particular concern. If subjects were randomized then baseline HR and covariate differences in general would likely be a non-issue.

 

[hlsmith]

Good points mostater. Also were interventions standardized. I believe there is quite a bit of variabilityin coffee options and energy drink options, along with sizes (fluid ounces). Without randomization as M pointed out, I can think of many issues (e.g., age, hydration, physical activity levels, social economic status, etc.) that could be grouped with choice.