Substate estimates for these years are no longer available due to methodological concerns with combining 2020 data with data from 2018 and 2019. We apologize for any inconvenience or confusion this may cause.Because of the COVID-19 pandemic, most respondents answered the survey via the web in Quarter 4 of 2020, even though all responses in Quarter 1 were from in-person interviews. It is known that people may respond to the survey differently while taking it online, thus introducing what is called a mode effect.When the state estimates were released, it was assumed that the mode effect was similar for different groups of people. However, later analyses have shown that this assumption should not be made. Because of these analyses, along with concerns about the rapid societal changes in 2020, it was determined that averages across the three years could be misleading.For more detail on this decision, see the 2018-2020substate data page.