Household survey data, such as the American Community Survey (ACS) and the Survey of Income and Program Participation (SIPP), often show a larger proportion of SNAP households with income above the program’s gross income thresholds than does SNAP Quality Control (QC) data.  Although the QC data, which are drawn from program operations data, might be considered more reliable, estimates from household survey data have been taken as evidence that the loosening of eligibility rules has allowed a growing number of  “less needy” households to draw benefits.  In particular, many States have adopted a Broad-based Categorical Eligibility (BBCE) policy under which a higher gross income standard is applied and the asset test has either been waived or made less restrictive. Yet policies like BBCE, though likely altering the income distribution among SNAP recipients, do not explain the divergent results on income obtained from ACS and QC data measured over the same time period. An alternative explanation is that a number of measurement issues affecting the household survey data account for much of this discrepancy. In this study, we characterize in greater detail SNAP access by household income as reported in the ACS. A fuller analysis of household income is desirable because of the inexact nature of estimating eligibility with survey data.  One of the primary measurement issues has to do with the fact that the ACS reports annual income whereas SNAP eligibility is based on monthly income. Our approach focuses on the implications of employing a more refined definition of the SNAP household that more closely resembles the size of benefit units found in administrative records. We also use the linked administrative data to explore the “intensity” of participation of individuals whose annual household income puts them above the (annualized) eligibility threshold. We look at three dimensions of intensity: the number of months in the program, the amount of benefits received and the time since last receipt. We find that higher-income SNAP households had on average a lower intensity of participation and were less likely to be current (as of the ACS interview month) participants. This last result indicates that a change in household circumstances not captured by the survey may explain why some higher-income households are identified in the survey as SNAP participants. We also examine other characteristics of higher-income SNAP participants, such as employment duration, in order to understand the way in which income dynamics may affect ACS estimates of participation and eligibility. Lastly, the analysis explores whether higher-income SNAP participants are more likely to have had their incomes imputed and the implications thereof.