Week 14 Community History Workshop: One Handbook, Many Lenses

Part 1 — The Big Picture

A workplace handbook is a small record of how an organization treats the people in it. This week you learned that intersectionality, Crenshaw's 1989 term, means overlapping categories produce experiences that neither category alone describes, that Loving v. Virginia (1967) unanimously struck down laws prohibiting interracial marriage, that Section 4(e) of the Voting Rights Act of 1965 protected citizens schooled in a language other than English, and that colorism, language, religion, and class can each meet race and gender in one person. In this workshop you code an illustrative ten-rule handbook for every category each rule touches, then count and compare. You set two documented laws side by side, each answering a place where two categories met, catch a chatbot reading overlap as addition, and reflect on what your own position lets you see.

Time budget: Plan for about 90 minutes.

Part 2 — The Task

Guiding question: Which categories does each rule in one workplace handbook touch, how many rules touch two or more at once, and what would a check of one category at a time miss?

Workshop data: An illustrative employee handbook of ten workplace rules

Source: A ten-rule employee handbook from one warehouse, written for this workshop as an illustrative example; it is printed in full in Part 4, describes no real company, worker, or person, and names nobody.

• Rows: 10 workplace rules, one row each in the data table
• Columns: rule number, what the rule says, and two blank columns for your category codes and your one-category check
• Scope: one handbook at one illustrative warehouse

Part 3 — Materials & Procedure

Tools and materials

• A notebook or any free note-taking app for reading a record and for oral-history notes (identified by role and decade, never by name); a free spreadsheet for the lending table's rates and gaps and for any year-stamped figure; any free timeline tool, or a hand-drawn timeline, for the law-and-case timeline; and one approved chatbot used as a sparring partner whose claims about real people, laws, and dates the student checks against an issuing-body or primary source before trusting them
• a free spreadsheet
• any free timeline tool, or a hand-drawn timeline
• a notebook or any free note-taking app
• one approved chatbot

Ways of working

• read-the-record
• oral-history
• compare-and-connect

Procedure

1. Read the ten rules in the data table in Part 4. They are an illustrative handbook, built for this workshop, from one warehouse; no real company, worker, or person is described, and people appear only by role. Tool: a notebook or any free note-taking app. Check: You can say in one sentence what the handbook covers and that it is illustrative.
2. Give each rule every code it touches: R for race or ethnicity (including skin tone), G for gender, L for language or national origin, F for faith or religion, and C for class or place. Write the codes in the codes column, and in the last column write what a check of one category alone would see or miss for that rule. Tool: a notebook or any free note-taking app. Check: Every one of the ten rows carries at least one code and a one-category note.
3. Open a blank sheet. Type Rule, Codes, and Number of codes in cells A1 to C1, enter your ten rows in rows 2 to 11 with the number of codes for each rule in column C, then count the rules with two or more codes with =COUNTIF(C2:C11,">=2"). Tool: a free spreadsheet. Check: Your count of rules with two or more codes plus your count of rules with one code equals 10.
4. Turn each count into a percent of all ten rules: divide the count by 10 and multiply by 100. Then find the gap, in percentage points, between the share of rules with two or more codes and the share with one code, and write each step in the count table. Tool: a free spreadsheet. Check: The two percents add up to 100, and the gap is written in percentage points.
5. Draw a short timeline from 1964 to 1989 and mark 1964, 1965, 1967, 1968, and 1989. Beside each year, write the law, case, or idea the course dates to it. Tool: any free timeline tool, or a hand-drawn timeline. Check: Five years are marked, each with its dated law, case, or idea.
6. Complete the comparison table: set Section 4(e) of the Voting Rights Act of 1965 beside Loving v. Virginia (1967), using only the dated facts the course gives, and name the two categories that met in the people each one answered. Tool: a notebook or any free note-taking app. Check: Every row of the comparison table is filled for both.
7. Paste your coding table into one approved chatbot, then paste the Part 6 prompt. Check every claim it makes about a real person, law, or date against the chapter's dated record before trusting it, or take the equivalent route instead. Tool: one approved chatbot. Check: You have quoted at least three chatbot claims and corrected at least three by hand.

Fallback route

When to use it: Use this route if no spreadsheet or timeline tool is available.

Code and count the rules by hand from the data table: write every code beside each rule, mark each rule that has two or more codes, count the marks, divide each count by 10, and multiply by 100. Draw the timeline by hand on paper, complete every scaffold table, analysis question, and the critique step, and submit the same items for equal credit.

This route earns the same credit as the primary route.

Part 4 — The Scaffold

The data come first: an illustrative ten-rule handbook from one warehouse, with two columns for your work.

Student scaffold table 1

Rule	What the rule says	Your codes (R, G, L, F, C)	What a one-category check would see or miss
1	No languages other than English may be spoken in the break room, even on breaks.
2	Hair must be worn in one of the styles pictured in the appendix; braids and locs are not pictured.
3	No hats or head coverings may be worn on the floor except the safety helmet the company issues.
4	A promotion requires a written recommendation from two current supervisors.
5	Night-shift staff may not swap shifts to cover child care; anyone who misses a shift is written up.
6	The commuter benefit goes only to staff whose home address is in a list of approved ZIP codes.
7	Name tags hold one first name of no more than eight letters.
8	Paid holiday leave covers only the holidays printed on the company calendar.
9	Front-desk staff must wear the uniform skirt.
10	Supervisor openings are posted only to staff who hold a four-year college degree.

Count the rules by how many categories each one touches, and turn each count into a percent of the ten rules.

Student scaffold table 2

Rules that touch	Count	Percent of all rules	Rule numbers
Two or more categories
One category
All rules	10	100

Compare two documented laws, each answering a place where two categories met, using only the dated facts the course gives.

Student scaffold table 3

Question	Section 4(e) of the Voting Rights Act of 1965	Loving v. Virginia (1967)
Whom it answered
Which two categories met in those people
What it did
How many years apart the two are, with the subtraction

Part 5 — Analysis Questions

1. Using your sheet or a hand count, how many rules touch two or more categories and how many touch one, what percent of all ten rules is each, and what is the gap between the two percents, in percentage points? Show each count, each division, and the subtraction.
2. Compare rule 2 with rule 3. Which categories does each one touch, and why does the course treat a dress rule like rule 2 as a borderline case to discuss rather than one to settle?
3. In 3–4 sentences, explain why a check by gender alone and a check by class alone could each pass rule 5, yet the rule can still land hardest where gender and class meet. Use intersectionality as overlap, not addition.
4. Using your comparison table, set Section 4(e) beside Loving v. Virginia (1967). In 3–4 sentences, say what the two share, one difference, and how many years apart they are, with the subtraction shown.
5. Positionality memo — choose one and answer it in 4–6 sentences. (a) What about your own background (the jobs you have held, the languages you speak, the faith or customs your family keeps) shaped which rules seemed fair to you and which stood out? (b) What can an illustrative handbook not tell you about the workers its rules touch, and which dated record or consenting interview would you need to check it?

Part 6 — AI-Critique Moment

Choose one equal-credit critique route. Route A: ask an approved AI tool to critique your completed work, then check its response against your own evidence. Route B is the no-AI equivalent printed below.

Approved tools: Claude, Gemini, ChatGPT, Grok, Microsoft Copilot.

Prompt to paste:

I've pasted my completed coding table above. It codes ten rules from an illustrative employee handbook at one warehouse, built for a class activity; it describes no real company, worker, or person. In about 200 words, please do six things. First, say which groups of workers these rules are about and what those workers are like. Second, explain intersectionality and say whether a worker who is a woman and an immigrant faces two problems that add up. Third, say when the Supreme Court decided Loving v. Virginia and what it struck down, and what Section 4(e) of the Voting Rights Act of 1965 protected. Fourth, say whether the break-room language rule has good and bad sides, staying neutral, and whether the country should pass an English-only law. Fifth, give one quotation from a civil rights leader about workplace rules and one statistic on how many workers face rules like these. Sixth, explain whether the census category Multiracial is a race in the biological sense and why the Multiracial count rose from 2010 to 2020.

Test the response for these likely failures:

• Chatbots often say the language rule is about "immigrant workers" as if they were one group with one language and one experience.
• Asked what the workers are like, a chatbot may say women of one group naturally prefer certain hairstyles or that some groups are naturally suited to night shifts.
• Asked for a quotation and a statistic, a chatbot will often invent a line for a civil rights leader and a percentage of workers facing such rules.
• Listing the census categories, a chatbot may describe American Indians as people of the past who are now absorbed into the Multiracial count.
• A chatbot may say laws against interracial marriage had two sides because states had their reasons.
• Asked whether the country should pass an English-only law, a chatbot may give a verdict as if the question were settled.
• Chatbots sometimes place Loving v. Virginia in 1954 or 1964, or say Section 4(e) protected non-citizens.
• Writing about interracial marriage, a chatbot may use an outdated label for people of more than one race as if it were current.
• A chatbot may call language, ethnicity, and nationality one and the same thing when it reads the break-room rule.
• Chatbots often call Multiracial a biological race and explain the 2010 to 2020 rise as population growth.
• A "balanced" chatbot summary may say the hair rule is only about a professional look, though a few workers felt singled out.

In 3–5 sentences, quote at least three things the chatbot wrote, say what it got right, and correct at least three of its errors by hand, naming for each the dated record or course definition you checked it against.

Equivalent route (no AI required)

No AI account is required for this workshop. This route assesses the same objectives (Objective 9 and Objective 2) for comparable effort, about 30 of the 90 minutes, under the same scoring criteria and the same credit. Instead of the chatbot, critique this flawed analysis, printed here in full: "These rules are all about immigrant workers, who share one language and one way of life. Women of one group naturally prefer braids, and some groups are naturally suited to night work. As one civil rights leader famously said, 'Every rule is a wall,' and a study shows that 60 percent of workers face rules like these. American Indians are a people of the past and are now counted as Multiracial. Laws against interracial marriage had two sides, since states had their reasons. The country should obviously pass an English-only law. Loving v. Virginia was decided in 1954, and Section 4(e) protected non-citizens. The old word for mixed-race people is still fine to use. Language, ethnicity, and nationality are the same thing. Multiracial is a biological race, and the count rose because more mixed families were born. To stay neutral: the hair rule is about a professional look, though a few workers felt singled out. Intersectionality means adding a worker's identities together." Mark it up: quote each flawed sentence, name what is wrong (many workers read as one uniform group, a stereotype stated as fact, an invented quotation and statistic, Native nations described as gone, a documented injustice framed as both sides, a current policy question presented as settled, a case date and a law misstated, an outdated term used without comment, language, ethnicity, and nationality used as one thing, a census category read as biology, a neutral-sounding summary that quietly takes a side, and intersectionality read as addition), and correct at least three errors by hand, naming the dated record or course definition behind each correction. Submit the marked-up analysis in place of the chatbot report and share link; it earns the same credit.

Part 7 — What to Submit

• The completed data table with your codes and one-category notes
• The completed count table and comparison table
• Your hand-drawn or tool-made timeline with the five years marked
• Written answers to all five analysis questions, including your positionality memo
• The chatbot critique report with its share link, or the equivalent route's marked-up flawed analysis

Deliverable:

A short Community History Workshop report on an illustrative workplace handbook: the completed data table with your codes and one-category notes, the count and comparison tables, a timeline with five years marked, written answers to five analysis questions including a positionality memo, and a critique of a chatbot's reading of the same handbook, or, on the equivalent route, a marked-up correction of the printed flawed analysis. Submit in any accessible file type or as a link; a recorded response may be submitted as audio or video with a transcript, or as text. Choose a written, recorded, or visual form for your answers; every form is graded by the same scoring criteria.

Rubric

The rubric below shows how each submitted part is evaluated; bands are full, partial, or none.

Workshop evaluation criteria

Criterion	Full	Partial	None
Coding the handbook and counting the overlaps	Every rule carries defensible codes, the two counts add to ten, each percent is correct with its division shown, and the gap is given in percentage points.	The codes or counts are attempted but one count or percent is wrong or unshown, the gap is missing or given in percent, or several rules lack a code.	No codes, counts, or percents appear, or the numbers are invented rather than counted.
Reading the rules in the field's terms	Each one-category note says what a single check would see or miss, rule 2 is read as a borderline case where race and gender may meet, and rule 5 is explained as overlap rather than two problems added up.	The notes or readings are attempted, but a note is missing, a reading is thin, or overlap is treated as addition.	No notes or readings of the rules appear.
A documented comparison across groups	The comparison table is complete with dated facts for both, the timeline marks all five years, and the answer names what the two share, one difference, and the gap in years with the subtraction shown.	The table, timeline, or answer is attempted but a row is empty, a fact is undated, or the comparison stops at listing.	No comparison appears.
The AI slip caught and corrected	Through the chatbot report with its share link or the equivalent route's marked-up flawed analysis, equally, at least three quoted errors are each named and corrected by hand against the dated record or a course definition.	Through either route, equally, one or two errors are corrected, or corrections lack the record or definition behind them.	Through either route, equally, no critique appears, or model output is pasted without any checking.
An honest reflection on positionality	The memo answers one of the two options in specific terms, naming what the student's own position let them see or miss, or naming the record or consenting interview that would check the handbook.	The memo answers an option in general terms without a specific example.	No memo appears.

Quality Gate

• Every count and percent comes from your own sheet or hand count of the data table, never from memory or a chatbot.
• Each count divided by 10 and multiplied by 100 equals the percent you report for it, and the gap between two percents is reported in percentage points.
• Every group is named as its members most commonly name themselves, every person is named only by role, any religious practice is described without judgment, and every date matches the chapter's dated record.
• No law, date, quotation, statistic, or detail about any real person or community is invented anywhere in the submission.
• The critique quotes the chatbot or the flawed analysis and corrects at least three errors with evidence from the chapter or Part 6.