The impact of income-support interventions on life course risk factors and health outcomes during childhood: a systematic review in high income countries

Table 6 Biases identified in the studies included in the review

	Information bias	Misclassification	Omitted variable bias	Reporting bias	Selection bias	Other
*Description*	Differences in the collection, recall, recording or handling of information used in a study	Incorrect classification of participants into categories	The statistical model leaves out one or more relevant variables	Selective disclosure or withholding of information by parties involved in the design, conduct, analysis, or dissemination	Individuals or groups in a study differ systematically from the population of interest
Almond et al. [34]		•		•
Baker et al. [38]		•	•
Brownell et al. [23]		•			•^a	•^a
Chung et al. [33]	•	•
Hamad et al. 2015 [29]	•	•			•	•^b
Hamad et al. 2016 [30]	•	•
Hoynes et al. [31]		•	•	•
Komro et al. 2016 [36]		•		•
Komro et al. 2019 [26]					•	•^c
Leyland et al. [22]	•			•
Milligan et al. [25]				•	•
Morris et al. [24]				•
Rosenthal et al. [35]			•	•
Strully et al. [27]		•				•^d
Wehby et al. [37]		•		•
Wicks-Lim et al. [32]	•	•				•^e

a. They could not ascertain how much unmeasured confounding influenced results, causing endogeneity bias. Furthermore, they limited the evaluation to women receiving welfare rather than examining all low-income women receiving the income supplement during pregnancy, thus limiting the generalizability of results
b. Recall bias may be present from women interviewed not in the year of childbirth
c. Observed result of LBW could be an underestimation of the differential effect by race because black women had a double rate of stillbirths
d. Mother's life environment and attitude or unmeasured genetic variation could influence results
e. Some factors that could cause spurious correlation such as neighbourhood gentrification, were impossible to control for

ISSN: 1471-2458