PART B: Please use STATA for this portion (30 points)

(10 points) A. In a hypothetical study using the 2017 BRFSS, where the study purpose is to estimate the effect of feeling in good health on regular physical activity in the non-institutionalized US adult (18+) population, please complete the following analysis tasks. In this study the outcome is physical activity and the exposure is self-reported general health status. You can use the code provided below to create the variables. Physical activity:

. generate exercise=1 if EXERANY2==1

(145,436 missing values generated)

. replace exercise=0 if EXERANY2==2

(107,444 real changes made)

. label define exerciselbl 0 "No exercise in past 30 days" 1 "some exercise in past 30 days"

. label val exercise exerciselbl

. tab exercise EXERANY2, missing

. tab exercise EXERANY2, missing

| EXERCISE IN PAST 30 DAYS

exercise | 1 2 7 9 . | Total

----------------------+-------------------------------------------------------+----------

No exercise in past 3 | 0 114,783 0 0 0 | 114,783

Some exercise in past | 301,030 0 0 0 0 | 301,030

. | 0 0 588 462 33,153 | 34,203

----------------------+-------------------------------------------------------+----------

Total | 301,030 114,783 588 462 33,153 | 450,016

General health:

. generate goodgenhlth=1 if genhlth<=3

(87,679 missing values generated)

.replace goodgenhlth=0 if genhlth>=4 & genhlth<=5

(86,530 real changes made)

. label define goodhealthlbl 0 "fair to poor general health" 1 "excellent, very good, good health"

. label val goodgenhlth goodhealthlbl

. tab genhlth goodgenhlth, missing

GENERAL | goodgenhlth

HEALTH | fair to p excellent . | Total

-----------+---------------------------------+----------

1 | 0 74,356 0 | 74,356

2 | 0 145,479 0 | 145,479

3 | 0 142,502 0 | 142,502

4 | 62,213 0 0 | 62,213

5 | 24,317 0 0 | 24,317

7 | 0 0 707 | 707

9 | 0 0 439 | 439

. | 0 0 3 | 3

-----------+---------------------------------+----------

Total | 86,530 362,337 1,149 | 450,016

1. Accounting for weighting and complex survey design, what is the prevalence of physical activity among those in good general health? Recall:*svyset _PSU [pweight= _LLCPWT ], strata( _STSTR ) vce(linearized) singleunit(certainty)

2. Accounting for weighting and complex survey design, what is the prevalence of physical activity among those not in good general health?

3. Without using hypothesis testing describe the difference in exercise for those in fair to poor general health compared to those in good to excellent general health.

4. To conduct a hypothesis test on the independence of the exposure and the outcome: What are your null and alternative hypotheses?

5. What statistical test did you use to test this hypothesis?

6. Interpret your results and compare to your preliminary conclusions.