# Little help on this given data (categorical)

#### josh_cantuba

##### New Member
Hello there!

I am a fresh grad from a Statistics course. And now, I am planning to do consultancy and help others with their research. I've done a lot of statistical analyses during and after my college years. But now, I encountered a client with a pretty complicated data.

Here's an outline of their given data:
Variables:
1. Age (Numeric)
2. Gender (M or F)
3. Left or Right Handed
4. Usage of handheld Devices (Less than 5 years or not)
5. Usage of thumb on devices (Left, Right, or Both)
6. Discomfort in any of the following areas:
a. Neck
b. Shoulder
c. Upper Arm
d. Lower Arm
e. Elbow
f. Forearm
g. Wrist
h. Hand
(The usual option was: left, right, or both but I am planning to convert the data given to either yes or no depending on if they feel pain or not in each part)
7. Any past injuries on the body parts stated in the body parts stated in #6 same options (Left, etc...)
8. Time spent on the ff activities
Emailing, texting, and
instant messaging
b. Scheduling (calendar,
appointments)
c. Internet browsing
d. Making phone calls and
talking on the phone
e. Listening to music,watching
videos, and taking pictures
f. Gaming: using mobile
phone, or hand-held video
game
g. Others
(options for each are less than 2 hrs, 2-3, 3-4, and more than 4, planning to convert the data into binary data: less than 3 hours and opposite otherwise)

They are planning to analyze if there are any relationships between usage of devices and body discomfort/injuries.
What I am planning to do is:

Use logistic regression with each data in variables #6 and #7 as dependent variables and setting variables #1,2,3,4, and 5 as independent variables so I can know how the said independent variables contribute to the odds of experiencing pain on each areas. However, this will lead to 16 distinct logistic regression models which is a lot.

I usually handle quantitative data, so this is a challenge to me. Hope you guys know a more compact way of analyzing the data and would also like to know if my idea is on the right track.

Thank you!

#### GretaGarbo

##### Human
However, this will lead to 16 distinct logistic regression models which is a lot.
Is that so much? I don't think so. It is good to work isn't it?
(But someone will talk about family wise error rates, or about jelly beans.)

#### josh_cantuba

##### New Member
Is that so much? I don't think so. It is good to work isn't it?
(But someone will talk about family wise error rates, or about jelly beans.)
Will have to do it then. Well yeah it is really fun to work haha do you happen to know any additional analysis i can do given the data? Thank you so much for your time

#### hlsmith

##### Not a robit
You can run multiple logistic regression, but as mentioned you will need to adjust your alpha level in order to address false discovery rate. Also, with this many hypotheses, you may need to also make sure you have ample sample size.