EPID 301 Lecture Notes - Lecture 4: Frequency Distribution, Pain Scale, Statistical Hypothesis Testing
! "#"$%&
$!
!'(
)*$+measurement errors,(selection error
- "!
!
./
0!!
+"#1
/
!$$
%%+
%!
1!
0/
!
!2!
33
%!&!
'4!
!5
%!&%(
67)
!
*!!
2!
8 +%
find more resources at oneclass.com
find more resources at oneclass.com
%
,!
)*$+$!-9
(
: "!1
%!&!
!%+-.
;
!#
%!&%(
/+<,
!'!2
%#;(&!0!
0%+;
%!$
!2;1
<
1#+
!1
!=
1$
> ?,
*31!
;
@;
3#;(A00
B !!'0
( 2#+#!
( ';($;
( 1/;("!
/!!17
31!7
@( 2
( 3<!
find more resources at oneclass.com
find more resources at oneclass.com
1CD)*!!;(
D
( 0E#C4!
2'$;( 0
F 6
/#+.C4(
+
1+C!9
-#+!/!22!
202(
C2!'.
G !
* ) * .
?/ # # # #
% ) # # #
) ) # #
%"60"$!
) ) # #
H
I0data
C &,!
2!1;,
79!/!2
JJ)*$+
",+KKKL!1,
,4,
MMN)MN+KKKL!<11(
O
$!,
2!P
&,
!;/
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
The diference between a measured value and its true value. It can be random (due to sampling variability) or systematic (aka bias) Error can also be due to laws in who is chosen to be part of the study or other factors related to participation; these are selection errors. Note: diference between measurement errors (misclassiication) vs. selection error (participation: sources of measurement error. Respondents were questioned on sociodemographic and clinical variables, pattern of internet gaming use and several self-report scales. Issues: who get selected / how people responded to their questions. Social desirability bias: lead the results to be socially desirable state (false results may have occurred) Who are more similar/ distant from the interview (who is doing interview, way the questions are posed) Preconceived with the risks of the interviewees (protocols that are very standardized will minimize the measurement errors) No standard protocols = may obtain diferent types of answers from subjects.