|
Overview
Specific
Aims
Importance
Background
Prelim.
Studies
Methods
Study
Design
Protocol
Data
Collection
Analysis
Data
Introduction
Basic
Data Sets
Variables
Defs.
Codebook
Results
Working
Papers
Publications
References
|
|
Variable Defintions- Some variables were created in the study
and are of special interest for analyses. A brief description of the variable,
the variable name, the SAS code, and the variable definition is given
for each variable.
| |
Light seasons are defined as the following intervals: (centered
at the equinoxes, maximizies light variation)
- Season 1: Nov. 6 - Feb. 4
- Season 2: Feb. 5 - May 6
- Season 3: May 7 - August 5
- Season 4: August 6 - Nov 5
|
| |
Common seasons are defined as the following intervals: (as defined
via equinoxes)
- Season 1 (winter): Dec. 21 - March 20
- Season 2 (spring): March 21- June 20
- Season 3 (summer): June 21 - Sept. 20
- Season 4 (fall): Sept. 21 - Dec 20
|
| |
Data on dietary intake, and exercise were collected each
quarter for study participants. A protocol was set to collect
these measures in a window of time surrounding the measures
of Cholesterol. The window of time corresponded to the times
from 4 weeks prior to the lipid measure, to 2 weeks after
the lipid measure. Thus the time window for dietary and physical
activity data is:
-28 days to + 14 days,
inclusive.
This time window is used when determining which 24-Hr measures
of intake (or MET activity) are to be used to estimate intake
for a corresponding lipid measure. In the 24-hour data set,
a variable corresponding to the quarter of the measure is
included for the linking. Only 24 hour records with a value
for the QUARTER variable are within the window for the Quarterly
data.
|
|
| Physical Acitivity Intensity |
ha2_L ha2_M ha2_V |
SAS Code |
| |
Physical activity variability was summarized in an article by Matthews
CE, Hebert JR, Freedson PS, Stanek EJ, Ockene IS, Merriam PA. (2000),
“Comparing physical activity assessment methods in the Seasonal
Variation of Blood Cholesterol Levels Study,? Medicine &
Science in Sports & Exercise 2000;32:976-984.?Additional
analyses used intensity measures, with physical activity coded as
light, moderate and vigorous.?To construct these intensity measures,
data from the 24-hour questionnaire were used.?Basic measures of
activity in this data set were
HA2A light sport hrs*: ha2a
HA2B moderate sport hrs*: ha2b
HA2C hard sport hrs*: ha2c
HA2D ?very hard sport hrs*: ha2d
Where each measure was reported in hours.?Intensity categories
(in MET0hrs/d) from these data were created as follows:
ha2_l = (ha2a * 1.5) ; ? ? ?* Light
intensity activity;
ha2_m = (ha2b*4.0) ; ? ?*Moderate
intensity activity;
ha2_v = sum((ha2c*6.0), (ha2d*8.0)) ; ?*Vigorous intensity
activity;
Some problems were encountered during the process of extracting
one year physical activities data from the overall crosssectional
data set. For details please see the "Details
of Creating Annual Physical Activities Data" page.
|
| |
- Original answers to social desirability questions (variables
DH1 - DH33) were coded 1, 2, and some were reversly coded, in
this dataset they were recoded to 0, 1 coding.
- Some questions had missing answers, and part (not all) of the
missing answers were imputed in this dataset, and the imputation
was done Gender specific.
- Protocol used in the imputation process: Only subjects with
80% or more completed questions are qualified for missing data
imputation.
- DHA1 - DHA33 = imputed Marlow Crowne variables
- See an example for the data imputation
process.
- Index variables were constructed to summarize the 33 questions:
- DH_SD="social desire*scale:* dh_sd"
- DHMISS="# missing DH*questions:* dhmiss"
- DHIMPUTE="imputed DH*questions:* dhimpute"
|
|
Other Links 