1. Do YOU Pay Attention?
A Mantel-Haenszel Analysis of an Observational Study
PH700A- Introduction to SAS
McKayla Smith, Alnino Guarino,
Ankita Kalraiya
2. Hypothesis/Rationale
There is an ASSOCIATION between sitting in the front of
the classroom and paying attention with sex being a
confounder. Confounder?
3. Background
● Being current students at SDSU, we wish to investigate
and uncover differences in gender and whether or not
sitting in the front of the class impacts one’s
ability/desire to pay attention.
● There is LIMITED current literature addressing topic.
5. Methods- Population, Variables, Possible Confounder
Study population: college students → graduate level
Public Health 601, 602, 603, 604, 605, 621
Independent Variables: level of attention given
(actively paying attention, not paying attention), sitting
position (first 3 rows of class, rear)
Possible confounder: sex (male, female)
6. Methods- Mantel-Haenszel Chi-Square Analysis
● SAS 9.4
● Crude odds ratio
● Stratum-specific odds ratio by sex
● Breslow-Day test of homogeneity to test equality of the
stratum specific odds ratios
● Adjusted odds ratio
● Confidence intervals
7. SAS Log
1 proc import datafile="D:PH700A Final Project.xlsx"
DBMS=excel OUT=project REPLACE;
2 run;
NOTE: WORK.PROJECT data set was successfully created.
NOTE: The data set WORK.PROJECT has 237 observations and 3
variables.
NOTE: PROCEDURE IMPORT used (Total process time):
real time 0.76 seconds
cpu time 0.17 seconds
3 proc format;
4 value positionfmt 1="front" 2="back";
NOTE: Format POSITIONFMT has been output.
5 value sexfmt 1="female" 2="male";
NOTE: Format SEXFMT has been output.
6 value attfmt 1="yes" 2="no";
NOTE: Format ATTFMT has been output.
7 run;
NOTE: PROCEDURE FORMAT used (Total process time):
real time 0.09 seconds
cpu time 0.04 seconds
8 proc freq data=project;
9 tables position sex attention;
10 format position positionfmt. sex sexfmt. attention attfmt.;
11 tables position*attention/measures;
12 tables sex*position*attention/CMH;
13 run;
NOTE: Writing HTML Body file: sashtml.htm
NOTE: There were 237 observations read from the data set
WORK.PROJECT.
NOTE: PROCEDURE FREQ used (Total process time):
real time 1.50 seconds
cpu time 0.36 seconds
14 proc sort;
15 by sex;
16 run;
NOTE: There were 237 observations read from the data set
WORK.PROJECT.
NOTE: The data set WORK.PROJECT has 237 observations and
3 variables.
NOTE: PROCEDURE SORT used (Total process time):
real time 0.03 seconds
cpu time 0.03 seconds
17 proc freq data=project;
18 tables position*attention/measures;
19 by sex;
20 run;
NOTE: There were 237 observations read from the data set
WORK.PROJECT.
NOTE: PROCEDURE FREQ used (Total process time):
real time 0.07 seconds
cpu time 0.01 seconds
8. Results- Descriptive Statistics
Position
● Front: 116- 48.95%
● Back: 121- 51.05%
Those who Pay
Attention
● Yes: 177- 74.68%
● No: 60- 25.32%
Sex
● Female: 138- 58.23%
● Male: 99- 41.77%
Total sample size =
237
9. Results- Crude Odds Ratio
Crude OR: 2.1516
The odds of sitting
in the front and
paying attention are
2.1516 times as
likely as sitting in
the front and not
paying attention. CI
(1.1723, 3.9488)
10. Results- Stratum-Specific OR (Female)
Female OR: 1.8985
Odds of being female in
the front and paying
attention are 1.8985
times as likely as being
a female sitting in the
front and not paying
attention. CI (0.8555,
4.2135)
Sex= 1 (Female)
11. Results- Stratum-Specific OR (Male)
Male OR: 2.5107
Odds of being male in
the front of class and
paying attention are
2.5107 times as likely
as being a male in the
front of class and not
paying attention. CI
(0.9744, 6.4696)
Sex= 2 (Male)
12. Results- Breslow Day Test for Homogeneity
● p-value (0.6578) → fail to
reject null that the stratum
specific odds ratios are
equal
● Adjusted odds ratio (2.1372)
not more than 10% different
from crude odds ratio → we
report crude odds ratio
(2.1516)
13. Conclusion/Major Findings
Crude odds ratios = adjusted odds ratios…
● Sex is NOT a confounder in whether sitting position is
independent of paying attention.
● REJECT original hypothesis.