FINAL REPORT 18JAN93

File: 123096_sep96_decls1_0008.txt
Page: 0008
Total Pages: 55

Subject: FINAL REPORT   18JAN93                                          

Unit: VAR. BUMED  

Parent Organization: BUMED       

Box  ID: BX303801

Folder Title: VARIOUS BUMED DOCUMENTS FOLDER 8                                                                

Document Number:          2

Folder Seq  #:         12







             of false positive replies. A "YES" was coded for a reply of "moderate" or

             "severel'. Furthermore, inconsistent replies to different questions about the

             same symptom were treated as negative responses for that symptom. For

             example, if a positive response was given for question #5, which asked whether

             a person had "three or more loose (unformed) bowel movements in a 24 hour

             period," and a negative response was given for question #22, which asked about

             the duration of diarrhea, then the subject was assigned a negative value for

             diarrhea.

                  For the respiratory symptoms (wheeze, cough, chest pain, fever with

             cough, runny nose, sore throat, and cold), logistic analysis was used to

             compare prevalence rates among exposure groups after adjusting for smoking

             status, previous respiratory disease (asthma, hay fever, emphysema, or

             bronchitis), and flu inoculation. First, for each respiratory symptom, the

             two-way interactions with each group (smoking status by group, previous

             disease by group and flu inoculation by group) were tested simultaneously

             using the likelihood ratio test. If this test was not significant, the

             interactions were removed from the model, and a likelihood ratio test was used

             on the resulting main effects model to test for differences among groups. If

             significant differences were found, the Wald test was used to compare Group I

             to Group III, and Group 11 to Group III.

                  For the non-respiratory symptoms, logistic analysis was used to compare

             prevalence rates among exposure groups. First, a likelihood ratio test was

             used to test for differences among groups. Again, if significant differences

             were found, the Wald test was used to compare Group I to III, and Group II to

             111. All testing was performed using a-.05. Wald-type confidence intervals

             were also computed.


                                                  7

Document 55 f:/Week-37/BX303801/VARIOUS BUMED DOCUMENTS FOLDER 8/final report 18jan93:1217961126382
Control Fields 17
File Room = sep96_declassified
File Cabinet = Week-37
Box ID = BX303801
Unit = VAR. BUMED
Parent Organization = BUMED
Folder Title = VARIOUS BUMED DOCUMENTS FOLDER 8
Folder Seq # = 12
Subject = FINAL REPORT 18JAN93
Document Seq # = 2
Document Date =
Scan Date =
Queued for Declassification = 01-JAN-1980
Short Term Referral = 01-JAN-1980
Long Term Referral = 01-JAN-1980
Permanent Referral = 01-JAN-1980
Non-Health Related Document = 01-JAN-1980
Declassified = 17-DEC-1996