# Help with gathering statistic information.

#### billclintonnn

##### New Member
Here is what I have done thus far. I can identify what each of the required information is, but I don't think I'm getting it exact. Here is my problem:

A heavy equipment manufacturer introduces new models of their backhoes and forklifts, and sells several of these machines. A few months later they discover a problem. An unacceptably high number of their customers needed to seek repairs of the machines’ hydraulic systems within the first few weeks of operation. The company’s engineers go to work on this problem, and soon think that they have found a solution. They believe that an additive poured into the hydraulic oil may greatly extend the number of hours these machines can be used before repairs become necessary. A few tests conducted under laboratory conditions indicate that this solution shows promise, but “real world” customer experience is needed before they can be sure. Impressed by these preliminary results, the company’s management gives the research team the green light to test their theory using up to 20 newly manufactured machines that will be sold during the next few weeks.

CLEARLY identify the following:
Subjects (Backhoes and forklifts)
Levels
Treatments
Placebo (if any)
Blinding (single or double) (I think it's single)
Design (completely random, block design, or matched pairs)
Blocks (if any) ???
Response variable (The additive being poured into the oil)
Diagram ???

My teacher goes way too fast and it's hard for me to understand by just reading the book. I know that the factors, levels and treatments are similar in a way. Can anyone help me identify what is which or point me in the right direction? I appreciate it!

#### noetsi

##### No cake for spunky
There is no standard usage, wording, in statistics so the terms you note could mean various things. My guess is that factors are the variable (here an additive) you are considering although (as the discussion you noted points out) other factors could include other environmental effects such as weather or maintenance. Ideally you would control for those.

Levels commonly means the possible levels of your predictor. Here you have two - you either have an additive or you don't. I assume the treatment is what you actually do to test your hypothesis commonly called an intervention. Here the treatment would be adding the additive. A placebo is commonly used to see if people react to simply having an intervention. That is unlikely to occur with a machine - the additive is unlikely to work because the machine thinks it is going to work So you probably don't need a placebo. An example of how you could do this if you were so inclined would be to add something you knew would have no effect and see if people concluded that the process was improving. But that would not help the vehicle (it might harm it) and is not needed so you would not do it.

There is an exception to the above and it goes to blinding. If there is no ready measure to rate performance - if it was largely subjective- then people might rate performance better because the additive was added even if no real change in performance occurs. In that case you might want people not to know what was being done when they did the rating. But it should be possible to generate an objective measure of performance for your experiment and it would be difficult to create a situation where people did not know if the additive was being added -and potentially harmful to add an additive that did not work. So blinding makes little sense here, any more than a placebo, and I see nothing in your comments that suggest it.

You have not provided enough details to comment on the design. The ideal way to do such a design would be to randomly assign treatment and non-treatment to vehicles. This is the gold standard of designs. I am not sure what you would block or match for since factors other than the additive that might change performance have not been suggested. But if you theorized other causes than the oil as causing the problem you would need to control for them by something like matching or blocks.

I think the response is the performance of the hydraulics. The oil is just one theorized problem.

Note a good way to get an instructor to slow down is to ask them a question that shows you put some thought into it

#### billclintonnn

##### New Member
Thank you noetsi much appreciated. However, this is legit what the problem is I didn't leave anything out. On my recent test it was easier to identify all the terms, but here it's really confusing to me. I'll dig more into it, but any other help would be welcomed.

#### noetsi

##### No cake for spunky
I believe you. What additional information do you need.

#### billclintonnn

##### New Member
I just want to make sure this is right. I know I'm pretty much repeating what I said and you said, but stay with me.

Subjects - Back hoes and forklifts. This is what the problem is revolving around.
Factors - This is where I'm confused. Isn't it just the Additive being placed into the machine?