Empirical Correlates of Survey Participation
- Patterns of being at home
- Demographic characteristics of refusers
o Analytic problem in describing characteristics of respondents and / or non respondents
§ Adjustment for probabilities of selection of sample person
§ Many surveys use a respondent selection technique that chooses at random one person among those eligible in sample household
§ Even if sample household selected with equal probability, this procedure yields unequal probabilities of selection for persons by household size.
Person living alone has higher chance of selection than one living with 5 other people
§ Many demographic characteristics not evenly distributed over household size – women and elderly tend to live in smaller households
§ Without compensation for unequal probabilities of selection, many sample surveys over represent persons who tend to live in small households.
Efforts by Survey Methodologists to Reduce Refusals
- Response rates are survey statistics that have their own sampling and non sampling error variances.
- Efforts to increase response rates
o Sole goal of increasing proportion of sample measured
o Without formal concern about nature of people added to sample or about response errors they commit
- Contact with sample household before survey request.
o Frequency of reading letter
o Affiliation of letter writer
o Other contents of letter
- Descriptions of survey at time of introduction
- Attempts to increase benefits of cooperation
- Discussion of attempts to reduce non response from refusals
o Mass media publicity
o Proxy respondents à interesting interplay of response error and non response error
o Uses of multiple modes of data collection
o Interviewer training
- Most attempts to reduce non response do so in an indiscriminant way
Sociological Concepts Relevant to Survey Non Response
- Why refusals occur
- Exchange relationship
Psychological Attributes of Non Respondents
- Literature on altruism and compliance
- Psychology of persuasion
No comments:
Post a Comment