Language for subgroup labels, Notes2, and updated Source

[cdsolari]

SUBGROUP LABELS BY METRIC:

• [x] ACS-based metrics (Income, Family Structure and Stability, Access to Preschool, College Readiness, Unemployment) All Black, Non-Hispanic Hispanic Other Races and Ethnicities White, Non-Hispanic

• [x] Financial Security: All Majority Non-White Majority White, Non-Hispanic

• [x] Affordable Housing: (Aaron fill in) Previous year

• [x] Housing Instability and Homelessness: 2018 2014

• [x] Neonatal health: All Black, Non-Hispanic Hispanic Other Races and Ethnicities White, Non-Hispanic

• [ ] Transportation Access (both metrics): [Confirmed] All Majority Non-White Mixed Race and Ethnicity Majority White, Non-Hispanic

• [x] Economic Inclusion: All Black Hispanic Other Races and Ethnicities White, Non-Hispanic

• [ ] Environmental Quality: All Majority non-White Mixed Race and Ethnicity Majority White, Non-Hispanic

High Poverty Not High Poverty

• [x] Exposure to Crime: 2017 2015

• [x] Overly Punitive Policing: All Black Other Races White

• [ ] Effective Public Education: All Black, Non-Hispanic Hispanic White, Non-Hispanic

Low-Income Not Low-Income

• [x] Living Wage 2014 2018

ADDITIONAL NOTES IN SUBGROUPS TAB:

• [x] Financial Security: Notes2: “Majority” means 60% or more of residents in a zip code are composed of that group.

• [x] Economic Inclusion: Notes2: “Black” includes Black Hispanics. “Other Races and Ethnicities” includes those in races not explicitly listed and those of multiple races. Those who identify as other race or multiple races and Hispanic are counted in both the “Hispanic” and “Other Races and Ethnicities.”

• [x] Overly Aggressive Policing & Exposure to Crime (this should go in the Detail tab): Notes: Although these are the best national data source, jurisdictions should use their local data if it is available. The FBI cautions using UCR data to rank or compare locales because there are many factors that cause the nature and type of crime to vary from place to place.

• [x] Overly Punitive Policing (in subgroup tab): Notes2: Ethnicity is not specified in the data. Those of multiple races are only included in "Other Races."

• [x] Transportation Access: Notes2: “Majority” means 60% or more of residents in a census tract are composed of that group.

• [x] Environmental Quality: Notes2: “Majority” means 60% or more of residents in a census tract are composed of that group. "High poverty" refers to census tracts in which 40% or more of people live in families with incomes below the federal poverty line.

• [ ] Effective Public Education: [Confirmed] Notes2: Research expects that annual improvement in English for Hispanic children will exceed those of White, Non-Hispanic children because Hispanic children have a lot of room to grow for improvement. Research also expects that annual improvement in English for students in low-income or economically disadvantaged households will exceed those of non-economically disadvantaged households because students in more advantaged households have limited room to advance. This metric, like all metrics, needs to be considered in context. “Low-income” means students are determined to be eligible for their schools' free and reduced price meals under the National School Lunch Program.

• [x] Living Wage (EARLIER YEAR) Notes2: Because we only have access to 2018 MIT living wage data, we use the CPI-U to deflate the 2018 values to be equivalent to 2014 dollars. The numerator of this figure (average earnings from the median job in an area) is based on 2014 data (QCEW).

Revisions for Source for cases where we add a prior year or change the data source for subgroups:

• [x] ACS-based metrics (Income, Family Structure and Stability, Access to Preschool, College Readiness, Employment) [Aaron to confirm the 5-yr data source is correct] Source2: American Community Survey (ACS) Public Use Microdata Sample (PUMS), 2018 (5-yr)

• [x] Housing Instability and Homelessness: Source2: US Department of Education Local Education Agency data (via Edfacts Common Core of Data), School Year 2018-2019 & 2014-2015

• [x] Transportation Access: Source2: Location Affordability Index data using National Transit Database data (via HUD AFFH data), 2016 (5-year); American Community Survey (via HUD AFFH data), 2016 (5-year) [Aaron to confirm the ACS year is correct]

• [x] Environmental Quality: Source2: National Air Toxics Assessment (NATA) data (via HUD AFFH data), 2014; American Community Survey (ACS), 2014 (5-yr)

• [x] Exposure to Crime: Source2: Federal Bureau of Investigations (FBI) Uniform Crime Statistic (UCR) Crime in the United States data series (via ICPSR), 2017 & 2014; American Community Survey, 2017 (1-yr) & 2014 (1-yr); New York State Division of Criminal Justice Services County Index Crime Rates, 2017 & 2014

• [x] Living Wage Source2: U.S. Bureau of Labor Statistics (BLS) Quarterly Census of Employment and Wages (QCEW) data, 2018 & 2014; Massachusetts Institute of Technology (MIT) Living Wage Calculator for a 1 adult and 2 child household, 2018; BLS Consumer Price Indexes for All Urban Consumers (CPI-U).

[cdsolari]

Aaron, we also need to change the metric name for preschool. The final metric name should be:

• [x] "Metric: Share of 3- to 4-year-olds enrolled in nursery school or preschool"

• [x] We also need to alter the description tab for that metric here:

[cdsolari]

I confirmed that for the Transportation Access metrics, the White is Non-Hispanic. In the mapping tool, the label is explicit:

[cdsolari]

I found out the definition of low-income (or economically disadvantaged) for Effective public education. I'll put the text here so it is easy to see, but I'll also edit the initial language comment text and note it is "[confirmed]". “Low-income” means students are determined to be eligible for their schools' free and reduced price meals under the National School Lunch Program.

[cdsolari]

Hi @awunderground, I know this might have something to do with the merge, but I wanted to flag some high-level things that didn't make it into the final data sheets to check for:

[VivianSihanZHENG]

• [x] filter out empty years
• [x] access to preschool shouldn't have more years
• [x] most recent years should be at the top
• [x] update notes
• [x] update source
• [x] housing instability needs the extra years added to "more years" - 2014 data missing
• [x] some metrics missing extra subgroups (e.g., effective public education)
• [x] add env quality subgroups
• [x] add poverty subgroups
• [x] some metrics have no longitude data (e.g., environmental quality)?
• [x] " becomes "" in new notes and data sources when knitting the pages

[cdsolari]

@awunderground I updated the living wage note2 and source2 based on information from Kevin.

[cdsolari]

@awunderground I reviewed the data sheets example you sent out on Friday. Below are a list of things to fix/consider before we send it to the rest of the group. I also have a list of other things we might propose to the team while they review. We can certainly qualify any of these with notes about how hard a change it would be, etc so that we put it in perspective. But, I'd like to fix some things before the rest of the group does their review:

• [x] Source2 doesn't show for affordable housing in "more years" tab.
• [x] The note for both transportation metrics should have "Majority" in quotes, but it shows up as Majority.' No quote exits for financial security, but it should be in quotes.
• [x] Exposure to crime is missing the note about the data source and not using it to compare jurisdictions (it does show for overly punitive policing. It should be present in the Detail and More Years tabs.
• [x] You decided to add on text to the existing notes rather than have a new line or Notes2. Downside is that it can easily be missed. How hard is this to change?
• [ ] Where applicable (e.g. economic inclusion), note2 is added before the confidence interval line, perhaps making it more challenging to notice a difference.
• [x] The environmental quality notes are set up differently than the others. I like this better, but instead I would want the new note (e.g. note2) to be on a new line.
• [x] Anything we can do about delineations? Thicker bottom line for every "Quality" row? Maybe bold for quality row values? (poor, good, marginal) and can do it for all tabs if that's preferable?

Extra things to consider (maybe leave this for the rest of the group to weigh in on?):

• Maybe in detail tab for PUMS data, say 2018 (1-yr) instead of just 2018 to contrast with the 2018 (5-yr) in the subgroups.
• Should we add a Notes2 for debt in collections why we don't have a mixed race category?
• We still have a lot of NA. Should we explain that?
• Affordable housing metric - to the notes, should we add more for the interpretation. Like those with higher than 100 have more units than households in need, and those with lower than 100 do not have enough affordable housing supply to meet their need. Note that we do this for other metrics. Environmental quality is treating the other subgroup for high poverty as a separate table, but I expected it to be all in the same one. I actually like the break-up, but it does lead to repetition and the notes are a little confusing. Should we keep it this way? This is also true for effective public education.

[awunderground]

@cdsolari There is not source2 for affordable housing. Is it just "U.S. Department of Housing and Urban Development Office of Policy Development and Research (HUD PD&R) Fair Market Rents and Income Limits, 2018 & 2014; American Community Survey, 2018 & 2014"?

[awunderground]

@cdsolari

Where applicable (e.g. economic inclusion), note2 is added before the confidence interval line, perhaps making it more challenging to notice a difference.

Is this important now that there is better spacing?

[cdsolari]

@awunderground Ah, see that I had the place-holder for you for the affordable housing labels, but not a place-holder for the source. The source2 you indicate in your comment makes sense. Thanks!

Thanks for the link to your sheets in progress. Yeah, I think the spacing is much better and it is ok to leave the confidence interval line where it is now that we can more clearly see the notes2 information in a separate line. I think the notes are longer than they have to be by formatting it this way, but I'm ok with that. I think that makes it a little harder for the printable pages situation, but that is probably the least of our worries compared to tables like family structure ;).

I also like the shading you were able to put in to break things up! I assume for when we have the group/year column, the formatting forces the shading to cover that column as well. I think it is more valuable to have the shading than not. Could the group/year column be second to the furthest left column? That furthest left column is a label for the data and so is the group/year column. I assume it isn't, but I figured I'd just ask explicitly to be sure!

[cdsolari]

@awunderground After our meeting with Greg, the next steps (not ordered) are:

• [x] Ask data viz experts about shading and change accordingly
• [x] add a section within "additional notes" about the "NA" values. This can come after the Confidence Intervals section. Include the following text:

"NA" in fields for metric values and data quality values indicates that the data are suppressed due to sample sizes or because that element is not applicable to that community (e.g. no zip code in the county is majority non-white).

• [x] In places where the value of the metric is "NA" make the data quality value be "NA" as well
• [x] For the affordable housing metric, add language to the notes before the line "Housing units are counted as affordable for a given income level regardless of whether they are currently occupied by a household at that income level." That text is:

Values above 100 suggest that there are more affordable housing units than households with those income levels. Values below 100 indicate a shortage of affordable housing for households with those income levels.

• [x] bold the "Source:" and "Notes:" text under each table.

[cdsolari]

@awunderground We have two new edits confirmed:

• [x] for the two metrics where we have two subgroups (environmental quality & effective public education), we have two separate tables. The notes related to the specific subgroup should only be there. For example, the race/ethnicity table for effective public education shouldn't have the notes about "Low-income" and the table about low-income shouldn't have any notes related to race/ethnicity. Only keep the text in notes2 that relates to the specific subgroup. Let me know if you need me to spell this out in a separate issue/comment.
• [x] Greg offered some edits to note2 for effective public education. Note that related to item one above, the first note2 applies to the race/ethnicity table, and the next one relates to the low-income table:

Research suggests that annual improvement in English for Hispanic children will exceed those of White, Non-Hispanic children because Hispanic children, on average, start with lower levels of English language skills and can improve more quickly than children with higher baseline skills.

Research suggests that annual improvement in English for students in low-income or economically disadvantaged households will exceed those of non-economically disadvantaged households because students in more advantaged households, on average, start with lower levels of English language skills and can improve more quickly than children with higher baseline skills. ‘Low-income’ means students are determined to be eligible for their schools’ free and reduced price meals under the National School Lunch Program.

[cdsolari]

@awunderground This looks great! Everything is in there, so that's great! I have a question on the College readiness that I'd like to raise with Greg. We have a poor data quality were the value is 100.0% but the confidence interval is (100.0%, 100.0%). I guess this is possible if the sample was >30 but all of them had a high school degree? I'll share the link with Greg (and cc you) and we'll talk through this one issue. Thanks!

[cdsolari]

@awunderground This relates to the above comment when the lower and upper bound of a confidence interval equal each other. We decided to not show these cases. Please search for cases where they are equal, and replace the value in the table with "NC." In the "Additional Notes" section, after the "NA" statement, add the following: "NC" in fields for confidence intervals or lower/upper bounds means that we are not able to calculate this because the underlying data lack variation.