Lecture 11_Measurement Error Flashcards

1
Q

Measurement Error

A

failure of the recorded responses to reflect the true characteristics of the respondents

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Measurement error variance

A

Variance across respondents- social desirability- men diff answers than women-but ind person makes same errors
(y_i=μ+ε_i)
Variance across trials -diff answer each time for same survey
(y_it=μ+ε_it)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Measurement error bias

A

When response deviation is non zero over conceptual replications

(y_i=μ+ε_i)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Ways to Study Measurement error (5)

A
Repeated Measures
Lab investigations -non exp and exp
Interpenetration/random ass. To diff methods/modes
record checks; external sources
collecting correlates of error
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Repeated measures TYPES (4)

A

reinterview survey with subsample of survey sample
single int with repeated measures
combo of survey and admin rec’s
multiple rec’s collected by panel survey that refer to diff time pds

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Lab investigations - Non experimental

A

Expanded interviews (Think aloud, Comprehension probes, Debriefing)
Methods to target source of error (vignettes, response timing)
Confidence measuresBehavior coding

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Issues with lab investigations

A

Attention level of Respondents
Motivation in response process
Type of participants
Reality

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

random ass. To diff methods/modes

A

Split ballot; split sample within surveys
Difference between T groups used to est biasMost focus on bias not Measurement error variance
Inference is to the sampling frame

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Assumptions of lab investigations

A

Reality

Heterogeneity of errors represented

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Interpenetration definition

A

Randomly assign subsamples to different production units (interviewers)
Many diff treatments are applied and compared- more common to focus on variance
All survey attributes are replicated for diff int. Estimation of variance due to production units (interviewers)
Interviewer variance

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Interpenetration model

A

y_ij=χ_i+M_ij+ε_ij

χ_i:true
M_ij:effect of jth method

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Issues with interpenetration/Interviewer variance models

A

Changes survey conditions
Logistical issues -(Attrition, Refusal conversion ass. Of interviewers)
interpreted to mean that reduction in assigned cases (m) leads to lower int effects

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Record checks; external sources Assumptions

A

Most interested in bias NOT measurement error variance

Record is assumed true value

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Record checks; external sources TYPES

A

Reverse record check
Forward record check
*Full design record check

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Reverse record check

A

Sample people with known characteristic covered in record

Measures failure to report on something E.g., abortions, crime incidents

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Record check model

A

y_i=R_i+ε_ι

R: record value for reporter

17
Q

Issues with Reverse record check

A

Cant measure overreporting (only a subset of events sampled)

18
Q

Forward record check

A
Samples reporters from a frame; then compares survey answers to recorded value
Measures overreports (when survey report not found in record)E.g., medical events
19
Q

Issues with forward record check

A

If no report is given; no record search conducted

20
Q

*Full design record check

A

Requires full population covered by one or more record systems; full access
All events captured
Asking about variables related to the events that are available for population of interest
Measure both under and over reporting E.g., official divorce records

21
Q

Issues with record checks

A

Assumes all records are complete

Linking errors

22
Q

Using Benchmarks

A

Universe data available to compute the same statistic as the sample
e.g., voting

23
Q

correlated response variance

A

interviewer variance

answers from one respondent are correlated with answers interviewed by the same interviewer

24
Q

simple response variance; A.K.A. reliability in psychometrics

A

respondent vary in their response deviations over conceptual trials anyway- even without interviewer effects

25
Q

Index of consistency

A

proportion of the element variance of y that’s due to the variance of the response deviations

Var(ε)/Var(y)

26
Q

Measure response variance

A

Y_i1=χ_i+ε_i1
Y_i2=χ_i+ε_i2

Variance in errors

E=(ε_i1-ε_i2)

27
Q

psychometric “reliability”

A

p_y=1-index of consistency

28
Q

Assumptions to measure Simple Response Variance

A

No change in true values, over X trials (threatened by long time between trials)
No memory effect of first trial on second (threatened by short time between trials)
Survey conditions the same between trials
(threatened by using supervisors to conduct int’s)

29
Q

Parallel measures

A

survey items that measure the same underlying construct with different response deviations
Both measures are asked of each respondent as indicators of same underlying attribute
Correlation should be high

30
Q

Generalizing parallel measures/multiple indicators

A

if multiple constructs measured and different methods are used for each- then separate out variance into
True values
Method effects
random measurement error variance

31
Q

Challenges of using multiple indicators

A

Context effects- similar questions
estimation issues with discrete variables
“just-identified” models

32
Q

Paradata in surveys

A

data collected as a byproduct of field activities

  • response latency
  • keystroke data
  • probng behavior
  • edit failure records