From: Methods used to estimate the size of the owned cat and dog population: a systematic review
Data collection method used to determine pet population estimate | Studies | Advantages | Disadvantages |
---|---|---|---|
Mail out survey using a commercial list of contacts | Reduces bias towards wealthier participants associated with telephone surveys. | Selection bias introduced as households that are not on the commercial list are excluded. May introduce measurement bias as the participant will be aware what the study is about. | |
Overestimation of population may be introduced as a period prevalence is measured in these studies. | |||
Door to door survey | Butler and Bingham [19] | Reduces non-response. | Costly and time consuming, probably only feasible in a small study area. |
Selection bias may have been introduced as only houses that were within 500 meters of a road were included and only roads that were passable by vehicle were used. Also true random selection was not used. | |||
Ortega-Pacheco et al. [51] | Costly and time consuming, probably only feasible in a small study area. | ||
Selection bias may have been introduced in this study as only households with a telephone could be included, this may have led to households with a higher SEC being over represented. Random selection was not used in the door to door surveys. | |||
Random-digit dialled telephone survey | Cost effective and logistically allows a large number of participants to be recruited in a short period of time. | Large numbers of non-domestic based numbers may be included leading to greater non-response. | |
Selection bias may have been introduced in this study as only households with a telephone could be included; this may have led to households with a higher social economic class (SEC) being over-represented. | |||
Randomised telephone survey using a list of numbers | Slater et al. [5] | Cost effective and logistically allows a large number of participants to be recruited in a short period of time. Reduces number of non-household based numbers associated with random digit dial surveys. | Selection bias may have been introduced in this study as only households with a telephone could be included, and if the telephone number was not listed it could not be included. |
An explanation of the study was given at the start of the interview, which may lead to measurement bias as households with pets might be more likely to complete the questionnaire. | |||
Murray et al. [3] | Selection bias may have been introduced in this study as only households with a telephone could be included, and if the telephone number was not listed it could not be included. |