Endocrine Disrupting Chemicals and Neurodegeneration

[mprec001]

I am currently working to download the data files and convert them to a usable file to upload. The board does not allow for SAS files, so I will convert them to excel and upload in the upcoming days.

[mprec001]

Attached is the combined dataset I will be using for my class project and analysis. This dataset is an aggregate of the various datasets in NHANES that captures all the variables I will be using. As I run different analyses, I will refer to this main dataset and clean it as necessary.

This dataset includes:

Taste and Smell Questionnaire data (CSQ)
Demographic data (DMQ)
Heavy Metal Blood Levels (PBCD)
Smoking Questionnaire Data (SMQ)
Blood Chemistry Profile (Bio_Pro) - To account for cholesterol, triglyceride, and blood glucose levels

[mprec001]

Since many of my variables are continous, I had to perform discretization.

I decided to approach this first by making my variables binary, i.e., 1 if it is above a certain threshold, or 2 if it is below a certain threshold.

Variables that were already discrete were left the same.

[mprec001]

My initial run using Banjo with my data included 237 subjects with 10 variables. It did not appear that my surrogate health indicator for neurodegeneration did not appear to be influenced by cadmium levels. (refer to top graph attached).

For the next run, I intend to include other surrogate health indicators for neurodegeneration, such as other adverse taste/smell disorder questions from NHANES, as well as other blood markers for different heavy metals. I will include these variables separately on their own, as well as combined.

[cwyoo]

My initial run using Banjo with my data included 237 subjects with 10 variables. It did not appear that my surrogate health indicator for neurodegeneration did not appear to be influenced by cadmium levels. (refer to top graph attached).

For the next run, I intend to include other surrogate health indicators for neurodegeneration, such as other adverse taste/smell disorder questions from NHANES, as well as other blood markers for different heavy metals. I will include these variables separately on their own, as well as combined.

Report here descriptive statistics for each variable, e.g., sample mean, median, mode, variance, and s.d. for each variable. Use reference coding for discritization and report the descriptive statistics for each variables that were discritized.

[mprec001]

My initial run using Banjo with my data included 237 subjects with 10 variables. It did not appear that my surrogate health indicator for neurodegeneration did not appear to be influenced by cadmium levels. (refer to top graph attached).

For the next run, I intend to include other surrogate health indicators for neurodegeneration, such as other adverse taste/smell disorder questions from NHANES, as well as other blood markers for different heavy metals. I will include these variables separately on their own, as well as combined.

Report here descriptive statistics for each variable, e.g., sample mean, median, mode, variance, and s.d. for each variable. Use reference coding for discritization and report the descriptive statistics for each variables that were discritized.

I will posting the information.

[mprec001]

My initial run using Banjo with my data included 237 subjects with 10 variables. It did not appear that my surrogate health indicator for neurodegeneration did not appear to be influenced by cadmium levels. (refer to top graph attached).

For the next run, I intend to include other surrogate health indicators for neurodegeneration, such as other adverse taste/smell disorder questions from NHANES, as well as other blood markers for different heavy metals. I will include these variables separately on their own, as well as combined.

Report here descriptive statistics for each variable, e.g., sample mean, median, mode, variance, and s.d. for each variable. Use reference coding for discritization and report the descriptive statistics for each variables that were discritized.

Do you want me to post the descriptive statistics for each variables pre-discritization and post-discritization?

[cwyoo]

My initial run using Banjo with my data included 237 subjects with 10 variables. It did not appear that my surrogate health indicator for neurodegeneration did not appear to be influenced by cadmium levels. (refer to top graph attached).

For the next run, I intend to include other surrogate health indicators for neurodegeneration, such as other adverse taste/smell disorder questions from NHANES, as well as other blood markers for different heavy metals. I will include these variables separately on their own, as well as combined.

Report here descriptive statistics for each variable, e.g., sample mean, median, mode, variance, and s.d. for each variable. Use reference coding for discritization and report the descriptive statistics for each variables that were discritized.

Do you want me to post the descriptive statistics for each variables pre-discritization and post-discritization?

Yes, however, there is no need for categorical variables (unless their number of outcome exceeds what Banjo can handle).

[mprec001]

Descriptive Statistics for each variable used in analysis 1 that were discretized:

AGE:
mean: 61
median: 61
mode: 80
SD: 12.59876238
variance: 158.7288136
Discretization: Age >=60 coded as 1, if not, 2

Urine Cadmium Levels:
mean: 0.645822785
median: 0.442
mode: 0.04
SD: 0.692201476
variance: 0.479142884
Discretization: Levels >=0.5 coded as 1, if not, 2

Serum Tryglycerides:
mean: 129.5400844
median: 114
mode: 55
SD: 67.54699774
variance: 4562.596903
Discretization: Levels >=200 coded as 1, if not, 2

Serum Cholesterol:
mean: 192.7679325
median: 194
mode: 166
SD: 40.2008852
variance: 1616.111171
Discretization: Levels >=240 coded as 1, if not, 2

Serum Glucose:
mean: 114.0042194
median: 104
mode: 93
SD: 35.84872854
variance: 1285.131338
Discretization: Levels >=126, coded as 1, if not, 2

Systolic Blood Pressure:
mean: 128.4388186
median: 126
mode: 134
SD: 19.02834225
variance: 362.0778088
discretization: BP >=140, coded as 1, if not, 2

The following variables were categorical in nature, so no discretization was necessary:

Gender, Education, Presence of Phantosmia (Phantom Odor), Smoking Status.

[cwyoo]

Use reference coding, i.e., 0, 1, 2, ... Also report descriptive statistics for the catagorized variables (including the graphs, e.g., histogram or stem and leaf, etc.) as well.

Descriptive Statistics for each variable used in analysis 1 that were discretized:

AGE:
mean: 61
median: 61
mode: 80
SD: 12.59876238
variance: 158.7288136
Discretization: Age >=60 coded as 1, if not, 2

Urine Cadmium Levels:
mean: 0.645822785
median: 0.442
mode: 0.04
SD: 0.692201476
variance: 0.479142884
Discretization: Levels >=0.5 coded as 1, if not, 2

Serum Tryglycerides:
mean: 129.5400844
median: 114
mode: 55
SD: 67.54699774
variance: 4562.596903
Discretization: Levels >=200 coded as 1, if not, 2

Serum Cholesterol:
mean: 192.7679325
median: 194
mode: 166
SD: 40.2008852
variance: 1616.111171
Discretization: Levels >=240 coded as 1, if not, 2

Serum Glucose:
mean: 114.0042194
median: 104
mode: 93
SD: 35.84872854
variance: 1285.131338
Discretization: Levels >=126, coded as 1, if not, 2

Systolic Blood Pressure:
mean: 128.4388186
median: 126
mode: 134
SD: 19.02834225
variance: 362.0778088
discretization: BP >=140, coded as 1, if not, 2

The following variables were categorical in nature, so no discretization was necessary:

Gender, Education, Presence of Phantosmia (Phantom Odor), Smoking Status.