When Jefferson High sophomore Aaliyah Nguyen told a room of peers that “coffee makes me sleep worse,” the biology graduate student running the session didn’t argue — she asked for numbers, definitions and a plan to test it.

Nguyen was one of about 60 students who crowded into a lab classroom at Cedar Valley Community College this week for “Ask It Like a Scientist,” a daylong workshop aimed at preparing teams for the regional bioscience fair.

In small groups, students arrived with observations: a brother who “gets sick less” after joining swim team; a friend who “locks in” after energy drinks; a gamer who swears late-night sessions make his hands “shaky.” By the end of the morning, those statements had been rewritten into questions with measurable outcomes, clear variables and realistic methods.

“An observation is a spark,” said workshop lead Priya Deshmukh, a doctoral candidate in physiology. “But judges — and reviewers — don’t grade sparks. They grade testable questions.”

What made a question “strong”

Deshmukh asked each table to pin its draft question to a whiteboard. Then she circled the same four points on nearly every sheet: specificity, measurability, biological grounding and feasibility.

At one table, a group started with: Does caffeine affect sleep? Deshmukh pushed them to narrow it.

“Which kind of sleep change?” she asked. “How much caffeine? Over what time? In whom?”

The group revised it to: Among 10th-grade students, does drinking 200 mg of caffeine after 3 p.m. change total sleep time on school nights compared with no caffeine after 3 p.m.?

Across the room, another group’s first draft — Does gaming affect your body? — was flagged as too broad. The students shifted to a question that tied behavior to physiology: Does 30 minutes of fast-paced video gaming increase heart rate and change reaction time compared with 30 minutes of reading?

Fair coordinator Elena Ruiz said the strongest projects tend to survive the “so what” question.

“If it’s biologically grounded, you can explain a mechanism,” Ruiz said. “If it’s feasible, you can actually collect the data you’re promising.”

Descriptive vs. experimental: Two ways to investigate

The workshop split investigations into two lanes. One is descriptive — documenting patterns without assigning a cause. The other is experimental — changing one factor and measuring what happens.

On a projector, Deshmukh displayed paired examples side-by-side and asked students to identify which was which.

Example set 1 (sleep and caffeine)

• Descriptive investigation: Among students who choose to drink coffee on school days, is there an association between number of cups consumed and self-reported sleepiness during first period?

  • Students noted this could be done with surveys and would show a relationship, not proof of cause.

• Experimental investigation: If participants drink either caffeinated coffee (200 mg caffeine) or decaf at 4 p.m., how does total sleep time that night differ between groups?

  • “Now you’re manipulating something,” Deshmukh told them. “That’s where cause-and-effect becomes possible — if you handle controls.”

Example set 2 (gaming, reaction time and physiology)

• Descriptive investigation: Is weekly time spent on competitive gaming associated with resting heart rate among teens at Jefferson High?

• Experimental investigation: Does 20 minutes of competitive gaming change reaction time and heart rate compared with 20 minutes of a calm puzzle game, when the same person does both on separate days?

Senior Malik Owens said the distinction mattered because his team initially planned to “just ask people” if gaming made them faster.

“We realized that’s basically vibes,” Owens said, laughing. “Now we have a plan to measure reaction time and heart rate before and after.”

Variables and controls, explained at the lab bench

At midday, the room shifted from brainstorming to mock data collection. Deshmukh handed out stopwatches and simple heart-rate monitors.

She introduced the terms students would need to defend their work:

• Independent variable: what the investigator changes.
• Dependent variable: what the investigator measures.
• Controls: steps that keep other factors the same so differences can be attributed to the independent variable.

At Nguyen’s table, the independent variable became caffeine condition (caffeinated vs. decaf; or caffeine-after-3-p.m. vs. none). The dependent variable became total sleep time.

Ruiz pressed students to list controls out loud.

“If you’re testing caffeine after 3 p.m., what else might change sleep?” she asked.

Students called out answers: screen time, exercise, bedtime routines, room temperature, stress, naps.

“Your control is not a magic word,” Ruiz said. “It’s the choices you make so those things don’t drown your signal.”

For the gaming project, the independent variable was activity type (fast-paced gaming vs. reading). Dependent variables were reaction time and heart rate. Controls included using the same room, same time of day, same device brightness, and the same reaction-time test each session.

Operational definitions: Making “sleepy” and “fast” measurable

The most heated debates erupted during the operational-definition exercise — when students had to decide what they meant by everyday words.

“People love to measure ‘energy’ until they have to define it,” Deshmukh said, as one group argued over whether yawning counts.

She required each team to write operational definitions — practical, measurable versions of their terms — and then defend them.

Worked example 1: Coffee and sleep

Nguyen’s team settled on:

• “Caffeine intake” = 200 mg caffeine consumed between 3 p.m. and 4 p.m., verified by providing one standardized beverage.
• “Total sleep time” = minutes asleep between self-reported lights-out and wake time, recorded in a sleep diary and cross-checked with a phone step/sleep app screenshot the next morning.
• “Sleep quality” (optional) = a 1–7 rating reported at 7 a.m., with 1 = very poor and 7 = excellent, using the same question each day.

Deshmukh asked the group what would happen if a participant drank an extra soda at dinner.

“That’s a confound,” Nguyen said, rewriting the protocol to include a rule: participants must report any additional caffeinated drinks after noon.

Worked example 2: Gaming, reaction time and physiology

Owens’ team operationalized their terms as:

• “Fast-paced gaming” = 20 minutes of a specific competitive shooter training mode at a fixed difficulty level, using the same console and controller.
• “Reaction time” = median time (milliseconds) across 20 trials on the same web-based reaction-time test, taken immediately before and within five minutes after the session.
• “Physiological arousal” = heart rate (beats per minute) measured with a finger sensor for 60 seconds at rest, then again within two minutes after the session; plus optional skin temperature recorded with a non-contact thermometer.

When a student suggested measuring “stress,” Deshmukh insisted on a definition that could be collected consistently.

“If you want stress, pick a scale and a time point,” she said. “Otherwise you’re just trading one vague word for another.”

A “boxed” plan: From question to prediction

Near the end of the day, Deshmukh taped a sheet of paper to the wall — a simple visual guide students could copy into their notebooks. It was drawn as a single box with arrows moving left to right:

Question → Variables → Operational definitions → Hypothesis → Prediction

Under it, she wrote a sample using the coffee project:

• Question: Does 200 mg caffeine after 3 p.m. change total sleep time on school nights?
• Variables: Independent = caffeine condition; Dependent = total sleep time.
• Operational definitions: caffeine = standardized drink 3–4 p.m.; sleep time = diary minutes cross-checked by app.
• Hypothesis: Late caffeine reduces total sleep time.
• Prediction: Participants will sleep fewer minutes on caffeine days than on decaf days.

Students gathered around to photograph the sheet.

“It’s like a checklist that turns a thought into a plan,” said junior Sofia Martinez, whose group is studying whether backyard bird feeders change the number of mosquitoes near patios. “It also shows you where your idea falls apart if you can’t define something.”

Feasibility — and what students were told not to do

As teams packed up, Ruiz offered a last reminder that feasibility is not just about time.

“No invasive blood tests. No withholding needed medication. No ‘I’ll just test it on my little brother,’” she said. “If you can’t do it ethically and safely, it’s not a project — it’s a problem.”

Several teams left with pared-down plans: fewer variables, shorter study windows, simpler tools. Deshmukh said that was the point.

“Science is creative,” she said, watching students fold posters and tuck away monitors. “But the creativity has to land in something you can measure.”